. ‘r ,0 .3? . :97;va 2...... “Lb” r . 3 u h- A ’ xiii: \- ..a.... . .5“. V i? fr}? F. o. . J. lot»: ,3 n}; . 333:3. “$4 : . » .. 1...(::su.. 211:: it; bk“. ‘4': r. 13.? :1 {5.17:3 1.23:1! ‘ “9,414.32: €0.33? 7!...) 4):.) ft... .. ~M.£\.n. u II .35. f. z. . s. .1 12.21.! 3.3.! 5.: as: .0). 5...“: unfit) ~ :LJ ; L. . .A1x . . z! y .1. . V _ . . . . I fig. .. 7 _ _ A? 3.1.. : ‘ . 3:: a .. smma I- ’l 1.13ka 2 at)"; Michigan State University l This is to certify that the dissertation entitled CONTROL STRATEGIES FOR JOHNE'S DISEASE IN DAIRY CATTLE presented by ROXANNE BEE PILLARS has been accepted towards fulfillment of the requirements for the Ph.D degree in Large Animal Clinical Science (Epidemiology) {X I fl 'MdjoY Fsrof'essor's Signature 9-1 5-08 Date MSU is an Affirmative Action/Equal Opportunity Employer -o-o—--—----.—._.—.—.....a - PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 5108 KzlProleccalPresIClRC/DateDue.indd CONTROL STRATEGIES FOR JOHNE’S DISEASE IN DAIRY CATTLE By Roxanne Bee Pillars A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Large Animal Clinical Sciences 2008 ABSTRACT CONTROL STRATEGIES FOR JOHNE’S DISEASE IN DAIRY CATTLE By Roxanne Bee Pillars A five-year longitudinal study was performed to better understand Johne’s disease (JD) control in dairy cattle in terms of its impact on the environmental reservoir for the disease, the disease burden within the herd, and the cost-effectiveness of control programs. The objectives of this study were to: (1) describe the distribution Mycobacterium avium paratuberculosis (MAP) in the environment of infected dairy farms over time, and observe how that distribution changes as herd prevalence changes; (2) evaluate the effectiveness of management practices in reducing the JD burden within the herd; and (3) determine if the management practices implemented to control JD were cost-effective. Seven dairy herds infected with JD participated in this study. Upon study enrollment, each herd implemented a JD control program designed specifically for that farm, based on a JD risk assessment and the operation’s goals and capabilities. The risk assessment was repeated annually, and the control program modified as necessary. Within herd JD prevalence was monitored annually by fecal culture and/or serum ELISA testing of all adult cows. Every six months, samples of feed, water, and bedding were collected and cultured for MAP, from the pre-weaned calf, weaned calf, lactating cow, and maternity areas, as well as the primary manure storage area and pasture when appropriate. A questionnaire was developed and administered to each producer and/or herd manager yearly, to collect information on the costs incurred as a direct result of the JD control program. Based on the data collected, descriptive statistics were generated. Logistic regression was used to assess the effectiveness of management changes in ‘ preventing infection with MAP, and the net present value (NPV) of the each farm’s JD control program was calculated. Environmental contamination with MAP was consistent over time. When herd prevalence was >2%, MAP was cultured fi'om the lactating cow floor and/or manure storage 75% of the time. When herd prevalence was 52%, MAP was never cultured from any area sampled. Management practices associated with neonatal calf care were found to have the greatest impact on cows subsequently testing positive for JD as adults. Specifically, those factors were: exposure to adult cows other than dam at birth (OR = 1.09, 95% CI: 1.06 — 1.13), and feeding colostrum from one cow to multiple calves (OR = 1.10, 95% CI: 1.09 — 1.12). When designing JD control programs, implementing management practices that minimize the exposure of newborn calves to Mycobacterium avium paratuberculosis being shed by infected adult cows should take priority. The NPV for the JD control program varied greatly across the herds. When calculated across all cows in the herd, the costs of the JD control programs implemented on these herds averaged $30/cow/year with a median of $24/cow/year. The annual losses due to JD averaged $79/cow/year with a median of $66/cow/year. Investing in a JD control program can be cost effective, and doing something to control JD was always a better economical decision than doing nothing. ACKNOWLEDGEMENTS I would like to thank my graduate committee members, Drs. John Kaneene (Center for Comparative Epidemiology), Daniel Grooms (Department of Large Animal Clinical Sciences), Joseph Gardiner (Department of Epidemiology), and Christopher Wolf (Department of Agricultural, Food and Resource Economics), for their guidance and assistance in completing this project. A special, sincere thank you is extended to my major advisors, Drs. John Kaneene and Dan Grooms. Thank you for always having open doors, open minds, and an endless reserve of patience. Another special thank you is sent to Dr. Joseph Woltanski, USDA-APHIS, Veterinary Services, East Lansing, M], for always being willing to get down and dirty while collecting samples, and still being able to write legibly. I would also like to thank Joseph Hattey, Diagnostic Center for Population and Animal Health, Michigan State University; and Tom Martin, Geagley Laboratory, Michigan Department of Agriculture; and their respective assistants; for processing and testing the thousands of fecal and serum samples collected over the course of this study. The data analysis is only as good as the test results on which it is based. This project was funded in a large part by a grant from the USDA/APHIS/VS- J ohne’s Disease Program, with additional support provided by the Michigan Department of Agriculture, and the Center for Comparative Epidemiology, Michigan State University. iv TABLE OF CONTENTS LIST OF TABLES ............................................................................. LIST OF FIGURES ........................................................................... INTRODUCTION ............................................................................. CHAPTER 1 LITERATURE REVIEW ..................................................................... 1.1. Introduction ...................................................................... 1.2. Mycobacterium avium paratuberculosis ..................................... 1.3. Transmission of Johne’s disease .............................................. 1.4. Pathogenesis of Johne’s disease ............................................... Immunology of MAP infection ............................................... Clinical manifestations of Johne’s disease ................................ A. Stage I: “silent” infection ....................................... B. Stage II: subclinical infection ................................. C. Stage III: clinical infection .................................... D. Stage IV: advanced clinical infection ........................ 1.5. Control and prevention of Johne’s disease .................................. Therapeutic treatment of MAP infections ................................. Vaccination for Johne’s disease ............................................ Diagnostic testing to identify MAP infected animals ................... A. Tests that detect MAP .......................................... Bacterial culture Genetic probe B. Tests that detect the immune response to MAP ............ Serological or antibody tests ............................... Tests that detect the cell-mediated immrme response. C. Summary of testing ............................................. Implementation of farm management practices to control Johne’s disease ......................................................................... A. Environmental reservoir of MAP on dairy herds ........... B. Risk factors for Johne’s disease on dairy herds ............ C. Farm practices to control Johne’s disease.................... 1.6. Economics of Johne’s disease ................................................ Production and performance losses due to Johne’s disease ............ Economic costs associated with production losses caused by Johne’s disease ............................................................... Economic costs of J ohne’s disease diagnostic and control programs. 1.7. Zoonotic potential of MAP ................................................... 1.8. Conclusion ....................................................................... CHAPTER 2 LONGITUDINAL STUDY OF THE DISTRIBUTION OF MYCOBA CT ERIUM A VIUM PARA TUBERCULOSIS IN THE ENVIRONMENT OF DAIRY HERDS PARTICIPATING IN THE MICHIGAN JOHNE’S DISEASE CONTROL DEMONSTRATION HERD PROJECT ................................................... 60 2.1. Abstract ........................................................................... 62 2.2. Introduction ...................................................................... 62 2.3. Materials and methods ......................................................... 65 Farms ........................................................................... 65 Determination of herd prevalence .......................................... 66 Environmental sampling .................................................... 66 Bacterial culture .............................................................. 68 Descriptive data analysis .................................................... 69 Statistical data analysis ........................................................ 69 2.4. Results ........................................................................... 71 Herd prevalence .............................................................. 71 Environmental culturing .................................................... 71 Statistical data analysis ........................................................ 76 2.5. Discussion ........................................................................ 79 CHAPTER 3 LONGITUDINAL STUDY TO EVALUATE THE EFFECTIVENESS OF MANAGEMENT PRACTICES IMPLEMENTED TO CONTROL JOHNE’S DISEASE ON INFECTED DAIRY FARMS IN MICHIGAN .......................... 86 3.1. Abstract ........................................................................... 87 3.2. Inn‘oduction ...................................................................... 88 3.3. Materials and methods ......................................................... 94 Herds ............................................................................. 94 Within herd JD prevalence and incidence ................................. 94 Herd risk assessment ......................................................... 94 Individual cow data .......................................................... 95 Implementation and monitoring of JD control program ................ 95 Statistical analysis .............................................................. 96 A. Comparing JD serum ELISA to fecal culture ............... 96 B. Calculating JD prevalence ..................................... 97 C. Effectiveness of JD control program ......................... 98 D. Determining which management practices are effective in the JD control program ..................................... 100 E. Effect of dam’s JD test status on offspring’s JD status. . .. 103 3.4. Results ............................................................................ 103 Descriptive data analysis .................................................... 104 A. Herds ............................................................... 104 B. Cows .............................................................. 105 Statistical data anlaysis ........................................................ 106 A. Comparing serum ELISA test with fecal culture ........... 107 B. Within herd JD prevalence over study period ............... 107 vi C. Effectiveness of JD control program ......................... Relative risk of cows exposed to JD control program testing positive compared to cows not exposed to JD control program ............................................. Incidence of JD ............................................. D. Determining which management practices were most effective in JD control program .............................. E. Regression model fit analysis ................................. F. Effect of dam’s JD test status on offspring’s JD status... 3.5. Discussion ........................................................................ Comparing the serum ELISA test with fecal culture .................... Within herd JD prevalence .................................................. Effectiveness of JD control program ...................................... Detemrining which management practices are effective in JD control programs ............................................................. Model fit analysis for univariable and multivariable logistic regression ..................................................................... Effect of dam’s JD test status on JD test status of offspring. . . . . . . . 3.6. Conclusion ....................................................................... CHAPTER 4 ECONOMIC EVALUATION OF JOHNE’S DISEASE CONTROL PROGRAMS IMPLEMENTED ON SIX MICHIGAN DAIRY FARMS .............................. 4.1. Abstract ........................................................................... 4.2. Introduction ...................................................................... 4.3. Materials and methods ......................................................... Study design .................................................................. Farms ............................................................................. Questionnaire used for economic data collection. Other data collected ......................................................... Data analysis .................................................................... A. JD prevalence ................................................... B. Cost of JD control program .................................... C. Losses due to JD ................................................ Decreased milk production ............................... Loss of future income due to premature culling ....... Benefits of JD control program ............................................... Sensitivity analysis ........................................................... 4.4. Results ............................................................................ Individual herd reports ...................................................... A. Herd l ............................................................ Farm background .......................................... JD risk assessment ......................................... JD control plan ............................................. Descriptive statistics ....................................... JD prevalence ............................................... vii 114 114 115 118 124 124 124 124 128 129 131 I35 I38 139 141 142 I43 I45 I45 I46 146 I47 147 I47 147 150 150 151 154 156 157 157 158 158 159 159 160 I61 Cost of the JD control program 2003-2007 ............ Economic losses due to JD 2003-2007 ....................... NPV calculation ............................................ Results of sensitivity analysis ............................ Producer perception of the JD control program. . . . . B. Herd 2 ............................................................ Farm background ............................................ JD risk assessment ......................................... JD control plan .............................................. Descriptive statistics ....................................... JD prevalence ............................................... Cost of the JD control program 2003-2007 ............ Economic losses due to JD 2003-2007 ....................... NPV calculation ............................................ Results of sensitivity analysis ............................ Producer perception of the JD control program. . . . . C. Herd 3 ............................................................ Farm background .......................................... JD risk assessment ......................................... JD control plan ............................................. Descriptive statistics ....................................... JD prevalence ............................................... Cost of the JD control program 2003-2007 ............ Economic losses due to JD 2003-2007 ....................... NPV calculation ............................................ Results of sensitivity analysis ............................ Producer perception of the JD control program ........ D. Herd 4 ............................................................ Farm background .......................................... JD risk assessment ......................................... JD control plan ............................................. Descriptive statistics JD prevalence ............................................... Cost of the JD control program 2003-2007 ............ Economic losses due to JD 2003-2007 ....................... NPV calculation ............................................ Results of sensitivity analysis ............................ Producer perception of the JD control program. . . . . E. Herd 5 ............................................................ Farm background .......................................... JD risk assessment ......................................... JD control plan ............................................... Descriptive statistics. . . . . . . JD prevalence ............................................... Cost of the JD control program 2004-2007 ............ Economic losses due to JD 2004-2007 ....................... viii I62 163 164 164 I66 166 166 168 168 169 I71 171 173 I74 174 I76 I76 I76 177 I78 178 180 180 182 183 I83 185 185 185 186 187 188 189 189 I91 I92 194 195 195 I95 I96 197 198 I99 199 200 NPV calculation ............................................ 201 Results of sensitivity analysis ............................ 202 Producer perception of the JD control program. . . . . 203 F. Herd 6 ............................................................ 203 Farm background .......................................... 203 JD risk assessment ......................................... 205 JD control plan ............................................. 206 Descriptive statistics. 207 JD prevalence ............................................... 207 Cost of the JD control program 2003-2007 ............ 208 Economic losses due to JD 2003-2007 ....................... 210 NPV calculation ............................................ 211 Results of sensitivity analysis ............................ 212 Producer perception of the JD control program. . . . . 213 Summary results ............................................................... 213 A. Farms and JD prevalence ...................................... 214 B. Cost ofthe JD control program 216 C. Economic losses due to JD .................................... 217 D. NPV results ...................................................... 219 E. Results of sensitivity analysis ................................. 219 F. Producer perceptions of JD control program ................ 222 4.5. Discussion ..................................................................... 224 4.6. Conclusion ...................................................................... 234 CHAPTER 5 OVERALL SUMMARY ..................................................................... 235 5.1. Introduction ...................................................................... 236 5.2. What is the extent of the MAP infectious burden in the environment of infected farms in relation to the JD burden in the herd, and does it change over time? .................................................................................. 237 5.3. Do farm management practices, designed to limit the transmission of MAP infection, actually decrease the JD burden in a herd over time? ....................................................................................................... 239 5.4. What specific management practices are the most effective in decreasing the JD burden in a herd? ...................................................... 241 5.5. Are management practices to control JD cost effective? ........................ 243 5.6. Conclusion ....................................................................... 245 APPENDICES ................................................................................. 246 Appendix A: Johne’s disease risk assessment .................................... 247 Appendix B: Economic questionnaire ............................................. 252 Appendix C: Farm input data for OptiCowTM Model ........................... 258 REFERENCES ................................................................................. 259 ix LIST OF TABLES TABLE 1.1: Factors associated with an increased risk of Johne’s disease in dairy herds .............................................................................. 38 TABLE 1.2: Factors associated with a decreased risk of J ohne’s disease in dairy herds .............................................................................. 40 TABLE 1.3: Qualitative summary on the impact of J ohne’s disease on dairy production and performance parameters. (References are sorted according to findings) ......................................................... 50 TABLE 1.4: Costs of commonly used Johne’s disease diagnostic test offered by the USDA certified J ohne’s testing laboratories in Michigan (prices current as of January 2008) ................................................... 58 TABLE 2.1: Distribution of Mycobacterium avium paratuberculosis (MAP) in the environment of seven Michigan dairy farms ............................... 72 TABLE 2.2: Percent of MAP culture positive cows and environmental samples by herd over time .................................................................. 76 TABLE 2.3: Univariable linear regression analysis using within herd Johne’s disease prevalence as outcome ................................................ 77 TABLE 2.4: Final multivariable linear regression model using within herd Johne’s disease prevalence as the outcome ................................. 78 TABLE 3.1: Herd size, breed and housing management of study herds ............... 104 TABLE 3.2: Kappa statistic for comparing agreement between Johne’s disease serum ELISA test and fecal culture .......................................... 107 TABLE 3.3: Cochran-Armitage test for “true” fecal culture prevalence — all cows. 108 TABLE 3.4: Cochran-Armitage test for “true” fecal culture prevalence — first lactation cows only ............................................................. 109 TABLE 3.5: Cochran-Armitage test for trend for ELISA prevalence — all cows. . 1 10 TABLE 3.6: Cochran-Amritage test for trend for ELISA prevalence — first lactation cows only ............................................................. l 11 x TABLE 3.7: Cochran-Amitage test for trend for prevalence based on Johne’s disease test status = positive — all cows ..................................... l 12 TABLE 3.8: Cochran—Armitage test for trend for prevalence based on J ohne’s disease test status = positive — first lactation cows only .................. 1 13 TABLE 3.9: Relative risk of exposure to Johne’s disease control program — Fecal culture as outcome ............................................................. 114 TABLE 3.10: Relative risk of exposure to Johne’s disease control program — ELISA status as outcome ................................................... 115 TABLE 3.11: Relative risk of exposure to Johne’s disease control program — JD test status = positive as outcome ........................................... 1 15 TABLE 3.12: Johne’s disease incidence rate over first three lactations for cows not exposed to the control program compared to cows exposed to the control program — using fecal culture as outcome .................. l 16 TABLE 3.13: Johne’s disease incidence rate over first three lactations for cows not exposed to the control program compared to cows exposed to the control program — using ELISA as outcome ......................... 117 TABLE 3:14: J ohne’s disease incidence rate over first three lactations for cows not exposed to the control program compared to cows exposed to the control program - using Johne’s disease test status = positive as outcome ........................................................................ 1 17 TABLE 3.15: Univariable logistic regression analysis of risk factors associated with cows testing positive for Johne’s disease ........................... 1 19 TABLE 3.16: Multivariable logistic analysis of risk factors associated with cows testing positive for Johne’s disease ....................................... 123 TABLE 3.17: Model fit analysis for univariable and multivariable logistic regression ..................................................................... 125 TABLE 4.1: Descriptive statistics for Herd 1 ............................................. 160 TABLE 4.2: Johne’s disease prevalence trends for 2003-2007 for Herd 1........... 162 TABLE 4.3: Cost of Johne’s disease control program 2003-2007 for Herd 1.. 162 TABLE 4.4: Economic losses due to Johne’s disease and assmned benefits of Johne’s disease control program for Herd 1 - 2003-2007.................. 163 xi TABLE 4.5: NPV of four scenarios for Johne’s disease (JD) control on Herd 1..... TABLE 4.6: Descriptive statistics for Herd 2 ............................................. TABLE 4.7: Johne’s disease prevalence trends 2003-2007 for Herd 2 ................ TABLE 4.8: Cost of Johne’s disease control program 2003-2007 for Herd 2. . . . . TABLE 4.9: Economic losses due to Johne’s disease and assumed benefits of Johne’s disease control program for Herd 2 - 2003-2007. . .. . TABLE 4.10: TABLE 4.1 I: TABLE 4.12: TABLE 4.13: TABLE 4.14: TABLE 4.15: TABLE 4.16: TABLE 4. I 7: TABLE 4.18: TABLE 4.19: TABLE 4.20: TABLE 4.21: TABLE 4.22: TABLE 4.23: TABLE 4.24: TABLE 4.25: NPV of four scenarios for Johne’s disease (JD) control on Herd 2... Descriptive statistics for Herd 3 ............................................ Johne’s disease prevalence trends 2003-2007 for Herd 3 ............... Cost of Johne’s disease control program 2003-2007 for Herd 3 ...... Economic losses due to Johne’s disease and assumed benefits of Johne’s disease control program for Herd 3 - 2003-2007. NPV of four scenarios for Johne’s disease (JD) control on Herd 3... Descriptive statistics for Herd 4 ............................................ Johne’s disease prevalence trends 2003-2007 for Herd 4 ............... Cost of Johne’s disease control program 2003-2007 for Herd 4 ...... Economic losses due to Johne’s disease and assumed benefits of Johne’s disease control program for Herd 4 - 2003-2007. . . . . . NPV of four scenarios for Johne’s disease (JD) control on Herd 4... Descriptive statistics for Herd 5 ............................................ Johne’s disease prevalence trends 2004-2007 for Herd 5 ............... Cost of Johne’s disease control program 2004-2007 for Herd 5 ...... Economic losses due to Johne’s disease and assumed benefits of Johne’s disease control program for Herd 5 - 2003-2007 ............... NPV of four scenarios for Johne’s disease (JD) control on Herd 5. .. xii I65 I70 171 I72 173 175 179 180 I81 182 184 188 189 190 I92 I93 I98 I99 200 201 202 TABLE 4.26: TABLE 4.27: TABLE 4.28: TABLE 4.29: TABLE 4.30: TABLE 4.31: TABLE 4.32: TABLE 4.33: TABLE 4.34: TABLE 4.35: TABLE 4.36: TABLE 4.37: Descriptive statistics for Herd 6 ............................................ Johne’s disease prevalence trends 2003-2007 for Herd 6 ............... Cost of Johne’s disease control program 2003-2007 for Herd 6 ...... Economic losses due to Johne’s disease and assumed benefits of Johne’s disease control program for Herd 6 - 2003-2007 ............... NPV of four scenarios for Johne’s disease (JD) control on Herd 6... Herd size, breed and housing management of study herds ............. Johne’s disease test prevalence (fecal culture and/or ELISA positive) for herds during observed study period 2003-2007 ......... Annual cost of Johne’s disease control programs implemented by study herds 2003-2007. All costs represent $/cow in herd ............ Annual losses due to J ohne’s disease for study herds 2003-2007. All values represent $/cow in herd ......................................... Net present value of Johne’s disease (JD) control program after 20 years for four different scenarios, excluding testing. All values are $/cow in herd .................................................................. Break-even point for cost of Johne’s disease control program for scenarios 1 and 2. All values are in $/cow/year ........................ Net present value of Johne’s disease (JD) control program after 20 years for four different scenarios, including testing. All values are $/cow in herd .................................................................. xiii 207 208 209 211 212 215 215 216 218 220 221 223 FIGURE 1.1: FIGURE 2.1: FIGURE 3.1: FIGURE 3.2: FIGURE 3.3: FIGURE 3.4: FIGURE 3.5: FIGURE 3.6: FIGURE 4.1: FIGURE 4.2: LIST OF FIGURES “Iceberg” effect: Relative proportion of cattle infected with Mycobacterium avium paratuberculosis distributed through the four stages of the disease .......................................................... Percentage Mycobacterium avium paratuberculosis (MAP) positive environmental samples by within herd Johne’s disease prevalence... Trend for “true” fecal culture within herd prevalence — all cows. . Trend for “true” fecal culture within herd prevalence — first lactation cows only ........................................................... Trend for ELISA prevalence — all cows .................................... Trend for ELISA prevalence — first lactation cows only ................ Trend for prevalence based on Johne’s disease test status = positive — all cows ..................................................................... Trend for prevalence based on J ohne’s disease test status = positive — first lactation cows only ................................................... Average costs of Johne’s disease control program 2003-2007 broken down by category. Values are shown in $/cow/year ........... Average losses due to Johne’s disease 2003-2007 broken down by category. Values are show in S/cow/year ................................. xiv 19 75 108 109 110 111 113 217 218 INTRODUCTION J ohne’s disease (JD) is ranked as one of the top three health issues affecting dairy cattle in the US (Wells, et a1, 1998), and is predicted to be the most economically important infectious disease in the dairy industry unless control practices are implemented (Collins, 2003). Johne’s disease is prevalent worldwide, and is becoming increasingly so in the US. Based on a national survey in 1996, it was estimated that JD cost the US dairy industry between $200 — 250 million annually (Ott, et a1., 1999). Since that time, the estimated reported prevalence of JD infected dairy herds in the US has tripled (USDA, 1997; USDA, 2008). It is likely the economic burden of ID has increased too. Aside from the substantial cost of JD to both the herd and economic health of dairy farms, there is also a potential human health risk. In recent years, there has been emerging evidence linking Mycobacterium avium paratuberculosis (MAP), the causative agent of JD, to Crohn’s disease in people (Feller, et a1. 2007). If this link is ever proven, or perceived, to be causal, the economic damage to the dairy and meat markets, both foreign and domestic, not to mention the loss of consumer confidence, could be unimaginable (Hansen and Rossiter, 1999). Consequently, understanding JD, along with its management and control, has become a priority for the US livestock industry (Linnabary, et a1., 2001). Johne’s disease is a chronic disease, characterized by an untreatable, slowly progressive, granulamatous enteritis (Sweeney, 1996). Cows generally become infected as young calves, but do not develop clinical signs of the disease until they become adults, two to five years later (Sweeney, 1996; Collins, 2003). Due to the long incubation period and subclinical stage of the disease, infected cattle are difficult to identify using currently available diagnostic tests, yet can become infectious at any time, potentially spreading the infection to susceptible herdmates (Whitlock 1992; Whitlock and Buergelt, 1996). As a result, testing and culling positive cattle, in and of itself, is relatively ineffective in controlling JD. Instead, control of JD must focus on implementing farm management practices that minimize the transmission of MAP to susceptible animals (Collins, 2003; McKenna, et al., 2006). Recommended farm practices to control JD are all based on what is currently known about MAP infection and its pathogenesis. Validation of these control practices in the field is limited, because of the time and expense it would require to collect the data (Groenendaal and Galligan, 2003). Instead, farm practices to control JD have been simulated, using existing theory for control, expert opinion, and the limited field data available (Collins and Morgan, 1992; Groenendaal, et al., 2002; Dorshorst, et al., 2006; Kudahl, etal., 2007). Field validation of these control practices is still warranted. Equally important as determining which farm practices are effective in limiting the transmission of JD, is estimating what it costs to implement those practices. Voluntary, widespread adoption of JD control programs will only occur if they are proven to be cost-effective. Numerous studies have attempted to quantify the costs of JD in terms of lost milk production and herd performance (Ott, et al., 1999), but studies on the cost to control the disease are lacking. Only when both sides of the equation are known (what the disease is costing the farm in lost production, and what it will cost to implement changes to control the disease) can producers make sound economic decisions regarding JD control. In summary, the control of JD is dependent on the implementation of farm practices that minimize the transmission of the disease and are cost-effective. Further farm-based studies are necessary to validate the effectiveness of practices recommended for JD control. Above all, there is a need to quantify the costs of implementing these control practices. This dissertation documents a five-year observational study of JD control practices implemented on seven commercial dairy herds in Michigan. The objective of this study was to answer the following four questions: 1) What is the extent of the MAP infectious burden in the environment of infected farms in relation to the JD burden in the herd, and does it change over time? 2) Do farm management practices designed to limit the transmission of MAP infection actually decrease the JD burden in a herd over time? 3) What specific management practices are the most effective in decreasing the JD burden in a herd? 4) Are management practices to control JD cost effective? CHAPTER 1 LITERATURE REVIEW 1.1. Introduction J ohne’s disease (JD) is a chronic enteric disease caused by Mycobacterium avium paratuberculosis (MAP). Although recognized primarily in cattle and other ruminant species, JD has been diagnosed in many other species of domestic and wild animals (Thoen, et al., 1975; Chiodini, and VanKruiningen, 1983; Williams, et al., 1983; Stehman, 1996; Buergelt and Ginn, 2000; Beard, et al., 2001; de Lisle, et al., 2002; Daniels, eta1., 2003a; Daniels, et al., 2003b; Corn, et al., 2005; Davidson, et al., 2004; Palmer, et al., 2005; Raizman, et al., 2005). It is prevalent worldwide, and has been predicted to be “the single most economically important single etiology infectious disease of dairy cattle” unless control measures are instituted (Collins, 2003). Johne’s disease is not a new disease. It was first diagnosed in a cow in Germany in 1895 by Drs. Johne and F rothingham (Olsen, et al., 2002). The first report of JD in the US occurred in Pennsylvania in 1908 (Pearson, 1908). Since that time it has spread across the country, and prevalence is increasing. In 1996, the National Animal Health Monitoring Service (NAHMS) estimated that 21.6% of dairy herds were infected with JD (USDA, 1997). In 2007, that estimate increased to 68% of dairy herds infected (USDA, 2008). Other regional estimates of the prevalence of MAP infected. dairy herds range from 50 - 96% (Collins, et al., 1994; Thorne and Hardin, 1997; Johnson-Ifearulundu and Kaneene, 1998; Hirst, et al., 2004; Keller, et al., 2004; Berghaus, et a1. 2006; Lombard, et al., 2006). In fact, JD is ranked as one of the top three health issues affecting dairy cattle (Wells, et al., 1998). The most recent national survey of beef herds estimated that 8% were infected with JD (Dargatz, et al., 2001), although other estimates vary from 4 — 76% depending on the method used to detect the disease, study design, and region of country (Thorne and Hardin, 1997; Hill, et al., 2003; Pence, et al., 2003; Keller, et al., 2004; Roussel, etal., 2005). Economic losses to the US cattle industry due to JD have been estimated to exceed $1.5 billion annually (Stabel, 1998; Harris and Barletta, 2001). Because of its substantial economic impact, and the potential public health issues should MAP be definitively linked to Crohn’s disease in people, understanding JD along with its management and control has become a priority for the US livestock industry (Linnabary, et al., 2001). The purpose of this paper is to review what is currently known about JD: the causative agent; pathogenesis as it relates to disease transmission; epidemiology in terms of environmean reservoirs, risk factors for transmission, and recommended control practices; and the economics of the disease and its control. Particular emphasis will be given to management, control, and economics of JD on dairy farms. 1.2. M ycobacterium avium paratuberculosis Mycobacterium avium paratuberculosis (MAP) is a gram-positive, acid-fast bacterium (Stabel, 1998; Harris and Barletta, 2001; Whittington and Sergeant, 2001; Olsen, etal., 2002; Tiwari, etal., 2006). It possesses a thick waxy cell wall made up of approximately 60% lipids (Rowe and Grant, 2006). This cell wall provides MAP with a survival advantage, both inside and outside of the host. It is sticky which leads to clumping of the bacteria (Klijn, 2001; Grant, 2003) and provides it with increased resistance to disinfectants (Whan, et al., 2001) and physical processes such as pasteurization (Grant, et a1. 1996; Grant, et al., 1998; Ellingson, et al., 2005). However, the cell wall also restricts the uptake of nutrients, making MAP the slowest growing of the cultivable mycobacteria (Rowe and Grant, 2006), with a generation time, under optimal growth conditions, exceeding 20 hours (Lambrecht, et al., 1988). MAP is differentiated from other members of the M. avium complex phenotypically by its dependency on mycobactin, an iron-binding siderophore necessary for growth; and genotypically by multiple c0pies of an insertion element, IS900 (Harris and Barletta, 2001; Olsen, et al., 2002; Tiwari, et al., 2006, Rowe and Grant, 2006). Similar to other pathogenic mycobacteria (M. tuberculosis, M. bovis, M. leprae), MAP is an obligate intracellular pathogen, surviving and multiplying within macrophages (Rastogi and David, 1988). Viable mycobacteria have been identified in three different sites inside the macrophage: phagosome, phagolysosome, and the cytoplasm (Rastogi and David , 1988), suggesting multiple mechanisms for evading the killing mechanisms of the macrophage. Most of these mechanisms appear to center around the bacterial cell wall (Ryter, et al, 1984; Rastogi and David, 1988; Frehel, et al., 1989). The lipid-rich cell wall of MAP is difficult to permeate (Rastogi and David, 1988, Chiodini 1996). Also common among intracellular mycobacteria, including MAP, is the ability to inhibit the fusion between phagosomes and lysosomes, thereby avoiding degradation by lysosomal enzymes (Ryter, et al., 1984; Frehal, et al., 1986; Frehal, etal., 1989). In other instances, mycobacteria develop a protective capsule composed of mycosides which allow the bacterium to survive in the hostile environment of the phagolysosome (Ryter, etal., 1984; Rastogi and David, 1988, Frehel 1989). While all mycobacteria, in general, have the ability to employ each or all the above evasion mechanisms, recent work suggests there are important strain differences, and some strains of MAP are more capable than others of surviving inside the macrophage; hence making them more pathogenic (Gollnick, et al., 2007). As an obligate intracellular pathogen of animals, MAP does not replicate outside the host in the environment (Whittington and Sergeant, 2001). It has been suggested that in addition to its cell wall, survivability of MAP is aided by recently identified gene sequences that, in the absence of adequate nutrients, allow it to enter a dormant, “viable- noncultivable state” and revert to a vegetative form when conditions again become favorable (Whittington, et al., 2004; Greig, 2005). Regardless, survival time of MAP in the environment, even with the ability to become dormant, is finite in the absence of an animal host. 1.3. Transmission of Johne’s disease In the domestic livestock industry, transmission of MAP from an infected to an uninfected herd ahnost always occurs through the purchase or introduction of infected animals (Sweeney, 1996; Step, et al., 2000; Whittington and Sergeant, 2001). The more important aspect of JD is how it is maintained, or spreads within a population or herd. The primary route of MAP infection is fecal-oral, with ingestion of the bacterium occurring via exposure to contaminated feedstuffs or environment (Sweeney, 1996; Step, et al., 2000; Whittington and Sergeant, 2001; Olsen, et al., 2002; Greig, 2005). Young animals are more susceptible to MAP infection than older animals (Hagan, 1938; Rankin, et al., 1961; Larsen, et al., 1975; Sweeney, 1996; Whittington and Sergeant, 2001). For example, in an observational study, 13 of 26 (57%) of calves born and raised on an infected farm either died of JD or had MAP lesions; while only one of six (17%) calves introduced into the herd when less than one year old, and none of six heifers introduced at greater than one year of age, had evidence of JD (Hagan, 1938). Also, when cattle were experimentally challenged with the same dose of MAP, the tissues of calves exposed at one month of age contained more MAP and pathologic lesions than did those of calves exposed at nine months of age or adult cows (Larsen, etal., 1975). However, that does not mean adult cattle are immune to MAP infection. Evidence suggests adult cows can become infected if repeatedly exposed to high doses of MAP (Rankin, 1962, Sweeney, 1996; Kovich, et al., 2006), but the extended incubation period characteristic of MAP makes it unlikely the disease will manifest itself during their productive lifetime (Sweeney, 1996; Whittington and Sergeant, 2001). The mechanism for increasing resistance to MAP infection with age is unknown. It has been hypothesized to be due do to the incomplete development of the immune system in young ruminants and/or easier access to the intestinal mucosa due to the “open gut” during the first 24 hours of life, which allows the absorption of macromolecules such as colostral immunoglobulins, and perhaps MAP (Sweeney, 1996; Olsen, et al., 2002). In addition, JD transmission is facilitated by the fact that shedding of MAP by infected cows is precipitated. by parturition (Harris and Barletta, 2001; Greig, 2005); thereby increasing the probability of exposing and infecting the next generation of herd replacements. While fecal-oral transmission is the most common, MAP infection can also be spread directly from dam to calf through in-utero infection or from contaminated colostrum and milk. Transplacental infection has been reported in multiple studies. The incidence of fetal infection occurring in cows in the clinical stages of JD ranges from 20- 40% (Pearson and McClelland, 1955; Lawrence, 1956; McQueen and Russel, 1979; Seitz, etal., 1989). In asymptomatic, MAP infected cows, transplacental infection occurred in only 8.6% of fetuses, and all occurred in cows classified as heavy fecal shedders (Sweeney, et al., 1992a). Collectively, in a recent meta-analysis, it was estimated that in-utero MAP infection occurred in 9% of fetuses from subclinically infected cows and in 39% of fetuses from clinically infected cows (Whittington and Windsor, 2007 ). Johne’s disease can also be transmitted fi'om dam to offspring through colostrum and milk. In one study, MAP was isolated from the colostrum of subclinically infected cows, as identified by fecal culture; 36% from cows classified as heavy shedders and 9% from light shedders. Isolation of MAP from colostrum was nearly three times of that in milk (Streeter, et al., 1995), and may again be due to the propensity for infected cows to shed MAP at parturition. MAP has been isolated in the milk of up to 35% of cows with clinical JD (Taylor, et al., 1981), 19% of asymptomatic heavy shedders, and 3% of asymptomatic light shedders (Sweeney, et al., 1992b). Thus, evidence suggests that as JD advances, the more disseminated the infection becomes, and the more likely it is for infected dams to pass MAP to their offspring, rather it be in-utero, directly in colostrum or milk, or from fecal contamination of the environment. Other potential routes of infection include: semen from infected bulls, embryo transfer, wildlife reservoirs, and fomites. MAP has been isolated from the semen and accessory sex organs of naturally infected bulls, providing the potential for infecting the uterine environment of cows (Ayele, et a1. 2004). However, following experimental inoculation into the uterus near the time of insemination, MAP was not cultured in the uterus or any extra-uterine organs beyond three or four weeks; leading to the conclusion 10 that MAP in the semen of bulls is more likely to be destroyed in the uterus rather than establishing a systemic infection in the cow (Merkal, et al., 1982). Regarding embryo transfer, in-utero infection of an embryo fiom an uninfected cow placed into a MAP infected recipient has been documented (Manning, et al., 2003). However the reverse, infection of the recipient after implanting an embryo from an MAP infected cow, has not been proven. MAP has been isolated from the uterine horns of naturally-infected donor cows as well as embryos collected from them (Kruip, et al., 2003; Bielanski, et a1. 2006). Yet, when embryos were washed according to the procedure established by the International Embryo Transfer Society and placed in uninfected recipients, none of the recipients, or the resulting calves, developed MAP infection over a period of five years (Bielanski, et al., 2006). It was concluded, therefore, that the risk of embryo transfer transmitting MAP infection from an infected donor to the recipient or the calf is very small (Sweeney, 1996, Kruip, et al., 2003; Bielanski, et al., 2006) Aside from domestic ruminants such as cattle, sheep and goats, MAP has been cultured from a variety of other domestic and wild animals including: swine (Thoen, et al., 1975), South American camelids (Belknap, etal., 1994; Stehman, 1996); multiple species of deer (Chiodini and VanKruiningen, 1983; Williams, etal., 1983; Davidson, et al., 2004; Raizrnan, et al., 2005), bighom sheep (Williams, et al., 1983), Rocky Mountain goats (Williams, et al., 1983), elk (Williams, et al., 1983), bison (Buergelt and Ginn, 2000), rabbits (Daniels, et al., 2003a; Raizman, et al., 2005); feral cats (Palmer, et a1, 2005) as well as a variety of wild birds (Beard, et al., 2001; Daniels, et al., 2003b; Corn, et a1, 2005) and non-ruminant wildlife, both predator and prey (Beard, et al., 2001; ll deLisle, et al., 2002; Daniels, et al., 2003b; Corn, et a., 2005). Many of these animals are dead end hosts; meaning that although they are infected, they do not excrete MAP at sufficient levels to be infectious, or their feces (such as that of foxes or stoats that eat infected rabbits) are repulsive causing avoidance by grazing ruminants (Grieg, 2005). It has been demonstrated that MAP can cross the wildlife — domestic species barrier (Williams, et al., 1983, Cetinkaya, et al., 1997; Daniels, et al., 2001; Judge, at al., 2005). However, the volume of MAP shed by infected wildlife is several times lower than that of infected sheep or cattle (Daniels, et al., 2003a; Corn, et al., 2005), and the pelleted nature of the feces of many of these species makes widespread dissemination of MAP into the environment unlikely (Sweeney, 1996). Moreover, the confined housing systems commonly used on many livestock, particularly dairy, operations limits the potential for domestic livestock to commingle with MAP infected wildlife or graze the same pasture. While the contamination of stored feedstuffs by infected wildlife on these operations remains a possibility (Beard, et al., 2001; Daniels, et al., 2003b; Palmer, et al., 2005), transmission of MAP by this route is negligible compared to the contamination of the environment by infected domestic ruminants (Corn, et al., 2005). A fomite is an object that serves to transfer infectious organisms from one individual to another. MAP is an organism that can readily adhere to objects such as boots, clothing, feeding equipment, vehicles, even other animals, and be transported to different areas within a herd or between herds (J ohnson-Ifearulundu and Kaneene, 1998; Grieg, 2005; McKenna, et al., 2006). Observations supporting this have been made during ongoing JD research at Michigan State University. MAP was isolated from the boots of four out of four different people, after walking through the holding pen on an 12 infected dairy farm (Grooms, unpublished 2008). Failure to wash and disinfect the boots before feeding calves could transport MAP to a population of highly susceptible animals. In another instance, MAP was isolated from skin swabs of the brisket, hock, and/or teats on 7 out of 10 cows housed in the close-up dry cow and maternity pens on an infected dairy farm. Concurrent fecal culture on all ten cows was negative, suggesting these cows were either not infected with MAP or were not actively shedding the bacterium at detectable levels at the time (Bolton, et al., unpublished 2006). Even though these periparturient cows may not have been infected, they potentially could infect their calves with MAP simply because they were carrying the bacterium on their bodies in areas commonly nuzzled by newborn calves. Transmission of MAP between cows by veterinary procedures such as rectal palpation has been suggested, but the ability of MAP to penetrate the rectal mucosa compared to the mucosa of the ileum remains unknown (Sweeney, 1996). While these and other breaks in biosecurity represent potential routes of MAP transmission, they are rarely implicated because they can almost always be traced back to an infected animal in the herd (Sweeney, 1996). 1.4. Pathogenesis of Johne’s disease Infection is defined as the invasion and colonization of pathogens in an organism. Disease is defined as the abnormal functioning of an organism Disease can result from infection. However, often disease is not directly due to the pathogen, but is rather the result of the body’s attempt to rid itself of the infecting pathogen. Such is the case with JD. The granulamatous lesions characteristic of JD are the result of the immune system’s battle to contain and rid itself of the infecting MAP (Chiodini, 1996; Coussens, 2004). 13 Unfortunately, as happens in many battles, that which is being protected also sustains damage, and has to deal with unintended consequences. While an in depth discussion of the molecular immunology of MAP is beyond the scope of this paper, a brief description of what occurs at the cellular level is provided, as it the basis for understanding the clinical manifestations of JD. Immunology of MAP infection MAP gains entry through the intestinal mucosa, primarily in the ileal region of the small intestine (Gilrnour, et al., 1965; Momotani, et al., 1988; Chiodini, 1996; Sweeney, et al., 2006b). There are three potential mechanisms by which MAP can penetrate the mucosal barrier: (1) paracellular route in which MAP passes between enterocytes despite tight junctions and the intact bacterium reaches the underlying lamina propria; (2) transcellular route in which the MAP is taken up by enterocytes by endocytosis and broken down with antigens being processed and presented on the basolateral cell surface in association with class II molecules that activate the intraepithelial, or lamina propria, lymphocytes; and (3) M cell route in which MAP is transported intact through the these specialized cells overlying the Peyer’s patches to be presented to the underlying immune cells. The route by which MAP is transported across the intestinal mucosa plays an important role in determining the type of immune response mounted by the host (Chiodini, 1996). Evidence suggests that the primary portal of entry for MAP is through the M cells overlying the Peyer’s patches in the ileum (Momotani, et al., 1988); and therefore, the following discussion will focus on the immune response resulting fi'om this route of infection. However, other routes such as the paracellular and transcellular routes, 14 cannot be excluded due to the occurrence of lesions in areas outside of the ileum, such as the colon, that are devoid of M cells (Chiodini, 1996). Unlike the surrounding enterocytes, M cells lack brush border microvilli, and they do not produce digestive enzymes or mucous; thus they provide an easily accessible surface for microorganisms such as MAP (Featherstone, 1997). MAP is transported across the M cell by transcytosis and deposited intact on the basolateral side of the cell. There it is phagocytized by resident macrophages or dendritic cells within the lamina propria (Chiodini, 1996; Stabel, 2000; Storset, 2003). At this point, MAP either, evades the macrophage’s killing mechanisms and multiplies, or it is processed and presented to T-lymphocytes, thus starting the immune process. The immune response to MAP is typical, and similar to the responses documented for other pathogenic mycobacteria (Coussens, 2004). It is paradoxical in nature; starting predominately as a cell-mediated response and transitioning in the latter stages of the infection to a humoral response. At the very end stages of the disease, immune anergy has been reported and there is no detectable immune response, either cell-mediated or humoral, allowing the infection to disseminate unchecked throughout the body (Chiodini, 1996; Stabel, 2000). The immune response to MAP is carefully choreographed by cross-talk between immune cells using a series of complex cytokine signals (Stabel, 2000; Coussens, 2004). Upon initial insult, MAP infected macrophages send a signal to the underlying Peyer’s patches, activating T-lymphocytes. These lymphocytes bind to the infected macrophages and either process the bacterial antigens for further immune processing (CD4+ cells), or kill the infected macrophages (CD8+ cells), releasing viable MAP into the surrounding 15 tissues where the immune process begins again (Stabel, 2000; Storset, 2003). At the same time these T-cells are releasing cytokine signals, the primary one being gamma interferon (IFNy) (Zubrick, et al., 1988; Waters, et al., 2003; Buza, etal., 2004; Khalifeh and Stabel 2004). [EN]! recruits blood monocytes to the infection site where they become activated macrophages to aid in controlling the spread of MAP (Zubrick and Czuprynski, 1987). IFNy also promotes CD4+ activity, and as the process repeats, more and more macrophages migrate to the infection site, causing the infected tissue to become inflamed. Gradually, this inflammation hinders function. At some point, perhaps due to tissue damage caused by the ongoing proliferative cellular immune response (Coussens, 2004), a signal is sent suppressing the release of IFNy. This slows the recruitment and influx of additional inflammatory cells into the infection site, and stimulates B-lymphocytes, which turn into plasma cells and produce antibodies (Stabel, 2000; Storset, 2003; Khalifeh and Stabel, 2004). Antibodies to MAP resulting from a natural infection do not protect the host from disease, and are ineffective in controlling the spread of the infection (Chiodini, 1996; Stabel, 2000, Coussens, 2004). This is because, by the time MAP antibodies are produced, the infection has become too well established, with the bacteria safely ensconced inside macrophages where they cannot be killed by the antibodies. In fact, the detection of MAP antibodies has been associated with the fecal shedding (Perez, et al., 1997; Storset, et al., 2001), and the onset of clinical disease. It is possible for there to be overlap in the cell-mediated and humoral immune responses (Chiodini, 1996). As the infection spreads, new foci of infection are formed within the intestinal wall. The earliest lesions may reach the humoral stage of the immune response, while the newer ones are 16 still in the cell-mediated stage (Storset, 2003). Over time, it is believed the constant exposure to MAP and its antigens overwhelms the immune system, resulting in complete anergy and the rapid dissemination of MAP throughout the body (Chiodini, 1996; Stabel, 2000). The immune response to MAP is often successful in controlling the infection. In endemically infected herds, it is likely most, if not all, animals would ingest or be otherwise exposed to MAP. Yet, usually only a small proportion of the herd is found infected (Chiodini, 1996), and only 10-15% of the infected animals develop clinical JD (Olsen, et al., 2002; Tiwari, et al., 2006). The ability of the immune system to completely eliminate the infection is not clear (Olsen, et al., 2002). In the case of M. tuberculosis, another intracellular mycobacteria that elicits an immune response similar to MAP, 95% of the exposed individuals are successful in eliminating the infection (Ellner, 1989). Evidence for the successful elimination of a MAP infection, is the observation that some animals identified as infected are later found to be MAP free (Chiodini, 1996). Others theorize that this phenomenon is the result of transient “pass through” of MAP in uninfected animals. In highly contaminated environments, animals ingest MAP; and it transits the gastrointestinal tract and exits in the feces where it can occasionally be detected by culture, but the animal itself does not become infected (Sweeney, et al., 1992c). In summary, the immune response to MAP and its outcome depends on many variables including: the number of exposures (Chiodini, 1996, Tiwari, 2006), the size of the infecting dose (Chiodini, 1996; Olsen, et al., 2002), the pathogenicity of the infecting strain (Miltner, et al., 2005; Gollnick, et al., 2007), the number of infectious foci l7 (Chiodini, 1996), age of the host at exposure (Sweeney, 1996; Olsen, et al., 2002;) the immune status and capability of the host (Chiodini, 1996; Whittington and Sergeant, 2001; Olsen, et al., 2002; Tiwari, et al., 2006); and the host’s genetic susceptibility (Koets, et al., 2000; Olsen, et al., 2002; Coussens, 2004). Clinical manifestations of Johne ’3 disease Infection with MAP has been divided into four stages depending on the severity of the clinical signs, the potential for shedding MAP into the environment, and the ease in diagnosing the disease (Whitlock, 1992). At any given time, the number of MAP infected animals in a population decreases in each subsequent stage of disease, resulting in the so—called “iceberg” effect of JD (Figure 1.1). In MAP infected herds, for every animal in the advanced clinical stage of JD, it is likely there are as many as 25 more animals infected; and only 15-20% of these infected animals will ever be detected, even. with the most sensitive testing techniques (Whitlock and Buergelt, 1996). A. Stage I: “silent” infection This is the earliest stage of the disease. It is called “silent” because there is no way to distinguish MAP infected animals in this stage from uninfected herdmates. They have no clinical signs of infection. There are no measurable subclinical effects in terms of retarded grth or weight gain; and there are no cost effective diagnostic tests to detect the infection (Whitlock and Buergelt, 1996; Tiwari, et al., 2006). The only way to detect animals in this stage of JD is through the demonstration of MAP in the tissues, either through culture or histologic examination of the affected intestine and/or associated 18 Figure 1.1: “Iceberg” effect: Relative proportion of cattle infected with szcobacterium avium paratuberculosis distributed through the four stages of the disease STAGE 2-5% IV Diarrhea 10% 111 Weight loss 35% II Occasionally detectable by testing, but no clinical signs I 50% Infected, but not detectable Total Number of Infected Cattle (Adapted from Whitlock, 1992) 19 lymph nodes. However, these animals may shed MAP into the environment intermittently, and at extremely low levels, below the detection threshold (Whitlock and Buergelt, 1996). This stage tends to contain the largest number (>50%) of MAP infected animals in a population, and lasts the longest, often months to years. The animals in this group ofien include calves, replacement heifers, as well as adult cows (Whitlock, 1992). B. Stage II: subclinical infection Stage H of JD consists of animals, generally adults, in the subclinical stage of infection. Only a small proportion (15-25%) of animals at this stage of the disease is detectable by currently available diagnostic tests. These animals do not have overt signs of JD, such as weight loss or diarrhea, but inflammation resulting fiom the cell-mediated immune response starts to affect intestinal tract function. Nutrient absorption is less than optimal, resulting in a lower nutritional plane, impairing performance and production. As a result, many of these animals are culled from the herd for reasons other than JD, never being identified as infected (Whitlock and Buergelt, 1996; Tiwari, et al., 2006). Frequently, these animals will be shedding MAP into the environment, potentially infecting other susceptible animals. Many MAP infected cows remain in the subclinical stage for years before progressing into the third, or clinical, stage of JD; and some may mount a successful immune response, such that they never progress to the clinical stage (Whitlock, 1992). It is generally believed the transition from a predominately cell- mediated immune response to a humoral response, with the production of antibodies against MAP, occurs at the end of stage II, and precedes the onset of clinical signs (Chiodini, 1996; Tiwari, et al., 2006). 20 C. Stage III : clinical infection The onset of clinical signs of JD follows an extended incubation period of 2-10 years (Whitlock and Buergelt, 1996; Collins, 2003). The first sign is generally weight loss in spite of a normal or sometimes increased appetite. This sign is often missed because the onset of clinical signs is often precipitated by parturition (Harris and Barletta, 2001; Whittington and Sergeant, 2001; Greig, 2005). Cows normally lose weight during early lactation, and that weight loss is unlikely to draw attention if a healthy appetite is maintained. The weight loss is a consequence of the progressive impairment of the functioning of the intestinal mucosa due to inflammation, hindering the absorption of nutrients. A firrther consequence soon follows in the form of diarrhea that is malabsorptive in nature. The diarrhea may be intermittent initially, with periods of normal manure consistency, but eventually becomes persistent. Thirst may be increased in these cows, but otherwise all other vital signs (appetite, temperature, heart and respiratory rate) remain normal (Whitlock and Buergelt, 1996). Only 10-15% of cows survive to the clinical stage (Olsen, et al., 2002; Tiwari, et al., 2006). Cows rarely remain in stage III longer than 3-4 months before progressing to stage IV, or, more likely, being culled (Tiwari, et al., 2006). Most cows in this stage of the disease will test positive on fecal culture and have detectable antibodies. On gross pathology, the small intestines of these cows will have the characteristic corrugated cardboard appearance, and the associated mesenteric lymph nodes will be enlarged (Whitlock and Buergelt, 1996; Olsen, etal., 2003). 21 D. Stage IV: advanced clinical infection Most cows are culled prior to reaching this stage. These cows are emaciated, weak, lethargic, and have the “pipe-stream” or “water-hose” diarrhea characteristic of JD. At this point, the damage to the intestinal tract has become so extensive that it has essentially ceased to function and absorb nutrients. This necessitates the utilization of body stores of fat and protein for survival; leading to cachexia, and the development of hypoproteinemia, resulting in submandibular edema, or bottle jaw. The condition of cows at this stage deteriorates rapidly, generally within a period of days; and often they cannot be salvaged and die as a result of dehydration and cachexia (Whitlock and Buergelt, 1996). For all intents and purposes, these cows starve to death Cows with advanced clinical JD may not test positive for antibodies to JD, as they may have reached immune anergy (Chiodini, 1996; Stabel, 2000). Without the immune system to hold the MAP in check, it rapidly disseminates throughout the entire body, and is readily detectable on culture and histopath (Whitlock and Buergelt, 1996). 1.5. Control and prevention of Johne’s disease Due to the insidious nature of MAP and its complex pathobiology, control and prevention of JD is extremely challenging (Sweeney, et al., 2006b). Johne’s disease control programs are multifaceted, and consist of any combination of the following: (on rare occasions) treatment, vaccination, diagnostic testing to identify MAP infected cattle, and (perhaps most importantly) the implementation of farm management practices aimed at preventing infection. Each will be discussed in turn, with the greatest emphasis placed on management including: the role of the environmental burden of MAP in sustaining JD 22 within a herd; risk factors associated with the spread of JD within a herd; and the proposed practices to prevent the transmission of MAP to susceptible cattle. Therapeutic treatrnent of MAP infections Treatment of MAP infection in production livestock is generally unrewarding and not practical due to the cost of the drugs, the hassle of continued daily administration, and protracted drug-residue withholding times for both milk and meat (St. Jean, 1996). Occasionally, there have been instances where treatment of MAP has been attempted using various antimicrobial agents that have demonstrated effectiveness in treating other mycobacteria] diseases (tuberculosis and leprosy) such as rifampin, clofazirnine, rifabutin isoniazid, pyrazinamide, and streptomycin, alone or in combination. However, in all cases, the drug protocol was unsuccessful in eliminating the infection. It only succeeded in temporarily alleviating clinical signs, and did not prevent shedding of MAP into the environment (St. Jean, 1996; Stabel, 1998; Belloli, et al. 2001). Treatment of MAP is, therefore, generally reserved for companion animals with strong sentimental value, or animals with high genetic value, in an effort to alleviate symptoms long enough to harvest embryos (St. Jean, 1996). More recently, monensin, a common feed additive in ruminant diets, has been associated with reducing the severity of lesions caused by MAP (Brumbaugh, et al., 2000), decreasing the odds of testing positive for MAP (Hendrick, et al., 2006a), and marginally reducing shedding of MAP from infected cows (Whitlock, et al., 2005; Hendrick, et al., 2006b). Monensin is an ion0phore antibiotic that modifies biological cell membrane permeability (Merck, 1991; Prescott, et al., 2000). It also alters the 23 proportion of volatile fatty acids produced in the rumen to favor proprionic acid production, which, in turn, improves feed efficiency and hence production (Merck, 1991 ). It is unknown if the mechanism for the positive effect of monensin on JD is due to increasing the permeability of the bacterial cell wall allowing for easier bacterial cell destruction (Whitlock, et al., 2005; Hendrick, et al., 2006b); or if improved feed efficiency puts the cow on a better nutritional plane, thereby allowing the maintenance of an active immune response for a longer period of time; or some combination of both. In reality, the suppressive effect of monensin on the progression of JD is a beneficial side- effect for most dairy producers. It has long been used as a coccidiostat in replacement heifers (Merck, 1991); and with its approval for use in lactating cows in November 2004 (FDA, 2004), it is now widely included in rations to enhance milk production For all practical purposes, JD disease is untreatable in production livestock (Wells and Wagner, 2000). If and when treatment is attempted, clinical improvement should not be confused with cure of the disease (Belloli, et al., 2001). Vaccination for Johne ’5 disease Vaccination for the control of JD is controversial (Collins, 1994; Stabel, 1998). Multiple experimental and field studies have demonstrated that vaccination reduces fecal shedding, the number of clinically affected animals, and the severity of pathologic lesions (Cramwell, 1993; Korrnendy, 1992; Juste, et al., 1994; Korrnendy, 1994; Wentink, et al, 1994; van Schaik, et al., 1996; Gwozdz, et al., 2000; Rast and Whittington, 2005; Reddacliff, et al., 2006). It does not, however, completely prevent infection or the spread of disease to susceptible animals (Konnendy, 1994; Wentink, et al., 1994; van Schaik, et 24 al., 1996; Reddacliff, etal., 2006). Use of vaccine is regulated in the US due to the potential for cross-reactivity leading to false positive tests for bovine tuberculosis (M. bovis) in vaccinated animals (Stabel, 1998; Harris and Barletta, 2001). Vaccinating for JD precludes the use of serological tests for diagnostic purposes, and granulomatous lesions can result at the injection site in cattle (Spangler, et al., 1991), as well as in people in the event of accidental self-inj ection (Patterson, et al., 1988). Vaccination may have a beneficial role in herds heavily infected with JD by alleviating symptoms of the disease and reducing economic losses, but it must always be used in conjunction with improved management practices to control further transmission of the infection (Harris and Barletta, 2001). Diagnostic testing to identify MAP infected animals Multiple diagnostic tests have been developed to diagnose JD. These tests fall into one of two categories; those that detect the actual bacterium, or those that detect the immune system’s response to it (Collins, 1996; Tiwari, et al., 2006). Given the pathobiology of a MAP infection, the efficacy of a test to correctly identify a MAP infected animal is dependent upon the stage of the disease process (Whittington and Sergeant, 2001). Almost without exception, the tests do very well confirming MAP infection in animals in the more advanced, or clinical, stages of the disease. They do not, however, do a particularly good job identifying animals in the early, or subclinical, stages of the disease (Collins, 1996; Step, et al., 2000; Olsen, et al., 2002; Dieguez, et al., 2008). The most commonly used diagnostic tests currently being used for JD will be briefly discussed and summarized. 25 A. Tests that detect MAP Bacterial culture Culturing MAP from infected tissues is the most definitive method for diagnosing JD (Collins, 1996; Stabel, 1998; Whittington and Sergeant, 2001). However, collecting tissue samples for an antemortem diagnosis can be problematic; and instead fecal culture is performed more commonly and is often used as the “gold standard” for confirming a diagnosis of JD (Collins, et al., 1991; Sockett, et al., 1992; Collins, et al., 1994; Sweeney, et a1, 1995; Whitlock, et al., 2000; Dargatz, etal., 2001; Stabel, et al., 2002; van Schaik, et al., 2003a; Collins, et a1. 2005; van Schaik, et al., 2005; Nielsen and Tofi, 2006; Tiwari, et al., 2006). Aside from necropsy and tissue biopsy, fecal culture is the most sensitive of the diagnostic tests currently available for JD (Whittington and Sergeant, 2001; Collins, et al., 2006). This is because shedding of MAP often occurs before the production of measurable antibodies (Whitlock and Buergelt; 1996; Whittington and Sergeant, 2001; Sweeney, et al., 2006a). The disadvantages of fecal culture arise from the slow-growing and fastidious nature of MAP. It takes 8-16 weeks to grow MAP in- vitro and requires special, mycobactin enriched media (Collins, 1996). Also, because contamination is often a problem when culturing feces, an aggressive decontamination procedure is necessary to prevent overgrowth of other fungal and bacterial microorganisms. This procedure is labor intensive and inadvertently decreases the number of viable bacteria in the sample, adding to the time it takes for detection (Stable, 1998; Readdacliff, et al., 2003). Additionally, bacteria grown on culture need to be verified as MAP by acid-fast staining procedures and/or polymerase chain reaction 26 (PCR), generally for the 18900 gene sequence (Collins, 1996; Tiwari, et al., 2006). Because of the time, special media, and experience required to culture MAP, fecal culture is relatively expensive as compared to other tests (Collins, 1996; Kalis, et al., 1999; Stabel, 1998; Tiwari, et al., 2006). The recent development of automated liquid culture systems has reduced the amount of time required to detect MAP to about half that required using Herrold’s egg yolk solid culture (from 16 weeks to 6-8 weeks), and improved sensitivity from 50% to 60-65% (Kim, et al., 2004; Motiwala, et al., 2005; Collins, et al., 2006; Rajeev, et al., 2006). However, it has not changed the decontamination procedure and requires additional specialized equipment; therefore, it has not reduced the cost ($15-23/sample; Michigan USDA certified Johne’s laboratories, 2008). Despite its better sensitivity, due to the cost, individual animal fecal culture is not recommended for routine screening of herds (Nielsen, et al., 2002a; Wells, et al., 2002b; Collins, et al., 2006). Instead, for herd screening purposes, culturing pooled fecal samples (mixing fecal samples from 5-10 cows together) or environmental samples from high-trafl'rc adult cows area, have proven to be valid and cost-effective methods to identify infected herds and get a rough estimate of within herd JD prevalence (Wells, et al., 2002a; van Schaik, et al., 2003b; Wells, et al., 2003; Kalis, et al., 2004; Raizrnan, et al., 2004; Tavompanich, et al., 2004; Berghaus, et al., .2006; Lombard, et al., 2006; van Schaik, et a1, 2007). In short, culturing for MAP remains a mainstay for diagnosing JD in infected animals and is used as an aid in herd control programs. 27 Genetic probe The evolution of PCR technology has made it possible to identify MAP antigens in samples using genetic probes. The genetic element most commonly used for the diagnosis of JD is a highly conserved insertion element of MAP, 18900; with multiple copies often present within each bacterium (Vary, et al., 1990; Collins, 1996; Harris and Barletta, 2001). While PCR is most frequently used to confirm the identification of MAP in cultured samples (Collins, 1996; Tiwari, etal., 2006), it can also be used on samples obtained directly from the animal (Stabel, 1998). Similar to culture, PCR has a specificity of >99%, but a much lower sensitivity (~30%) due to its inability to detect MAP antigens in animals shedding low numbers of bacteria (Stabel, 1998; Harris and Barletta, 2001; Collins, et al., 2006). In a study comparing direct PCR to fecal culture, PCR only identified 60% of cattle positive on fecal culture (Whipple, et al., 1992). Studies investigating different genetic probes to improve sensitivity while maintaining specificity are ongoing, but are not yet commercially available (Stabel, 1998). The advantage of PCR over culture is its speed; requiring only three days for test conrpletion. However, the required skills and equipment necessary to conduct the PCR test makes it as, or more, expensive as culture; hence, prohibiting it use for routine herd screening purposes (Collins, 1996). B. Tests that detect the immune response to MAP 28 Serological or Mbodv tests Three techniques have been developed to detect antibodies to MAP: complement fixation (CF), agar-gel immunodiffusion (AGID), and enzyme-linked immunosorbent assay (ELISA) (Collins, 1996). There are difierent types of commercially available ELISA tests, and all are superior in sensitivity to either the AGID or CF. The ELISA is, therefore, currently the most commonly used assay to detect MAP antibodies (Olsen et al., 2002; Tiwari, et al., 2006). The advantages of ELISA tests include: ease of sample collection (serum or milk), availability of results within days, and relatively low cost ($6/sample; Michigan USDA certified J ohne’s laboratories, 2008) (Collins, 1996, Tiwari, et al., 2006). The main disadvantage of the ELISA test is its overall lack of sensitivity (30%) (Collins, et al., 2006). The humoral immune response to MAP, with the production of antibodies, generally does not occur until well after infected animals start shedding MAP (Chiodini, 1996; Whitlock and Buergelt, 1996; Whittington and Sergeant, 2001); making the ELISA less effective in detecting subclinically infected animals than individual fecal culture (Dargatz, et al., 2001; Tiwari, et al., 2006; van Schaik, 2007). So typiCally one would expect that ELISA positive animals would be fecal culture positive, but that is not always the case. In one study 30 out of 33 cows (91%) positive on serum ELISA were negative on concurrent fecal culture (Pinedo, et al., 2008). In studies where cows with positive serum ELISA’s were followed up with fecal culture, 6% and 20% respectively, were fecal culture negative (Stabel, et al., 2002; Muskens, et al., 2003bb). These contradictory JD test results may be partially explained by the intermittent shedding that is not uncommon with MAP infections during the subclinical stages of the disease (Whitlock 29 and Buergelt, 1996). Another possibility is the potential for false positive ELISA test results. While the ELISA is generally assumed to have excellent specificity (Collins, et al., 2006), it is not perfect, and false positive tests do occur (Hendrick, et al., 2005b). There are documented cases with a disproportionate number of false positive ELISA teSts thought to be the result of exposure to other environmental mycobacteria (Grooms, et al., 2006; Roussel, et al., 2007). Also documented, is substantial variation in the level of antibodies upon serial testing, possibly due to stage of lactation and status of the immune system (Hirst, et al., 2002; Nielsen, et al., 2002a; Barrington, et al., 2003; van Schaik, et al., 2003a). Moreover, it has been reported there is very little to only moderate agreement between concurrent milk and sermn ELISA results (Hardin, et al., 1996; Hendrick, et al., 2005a); suggesting MAP antibody levels can vary between different tissues within the same cow on the same day. Because of all these things, it is advocated that ELISA test results be interpreted. quantitatively, rather than as simply positive or negative; taking into consideration the origin of the sample (milk vs. serum), the clinical presentation of the individual animal, and the JD history of the herd, (Adaska, et al., 2002; Collins, etal., 2005). The bottom line is the low cost and quick turn around time for results has made the ELISA test, the JD test of choice for many producers and veterinarians, despite its many drawbacks. As with culture and PCR, the accuracy of the ELISA tests improves as the disease progresses (Collins, 1996; Whitlock, et al., 2000; Stabel, et al., 2002; van Schaik, et al.,2003a). It is probably best used as a cost-effective method for screening purposes to identify infected herds, monitor disease burden over time, and aid in the identification and removal of the most infectious animals in a herd; although confirming 3O the diagnosis with follow-up fecal culture is recommended before making decisions regarding individual animals (Dargatz, et al., 2001; Wells, et al., 2002b; van Schaik, et al., 2003a; van Schaik, et al., 2007). Tests to detect the cell-meflted immune response Key to any disease control program is the accurate, early identification and removal of infected animals before they have a chance to transmit the disease to others. The consistent problem with the diagnostic tests for JD discussed so far is their inability to detect cows in the early and subclinical stages of the disease; they only detect animals after they have become infectious. The earliest stage of MAP infection is characterized by a cell-mediated immune response (Chiodini, 1996; Stabel, 2000; Storset, 2003; Coussens, 2004). It is believed this response occurs prior to bacterial shedding, and its waning contributes to shedding and the progression of the disease (Chiodini, 1996; Stabel, 2000). Being able to accurately identify the cell-mediated immune response to MAP would identify animals prior to them becoming infectious, and would go a long ways toward controlling JD (Stabel and Whitlock, 2001 ). Cell-mediated immune function can be assessed by the following two methods: antigen-specific delayed-type IV hypersensitivity reactions, and in-vitro T lymphocyte proliferation and cytokine stimulation assays (Stabel and Whitlock, 2001). The most commonly used test for the delayed-type IV hypersensitivity reaction is the skin test, where pathogen specific antigens are injected intradermally. If the animal is infected, swelling will occur at the injection site over a period of three days. Skin testing has been 31 the cornerstone for diagnosing tuberculosis in both people and cattle. Skin testing for JD has not been successful, most likely due to cross-reactivity with other ubiquitous mycobacteria in the environment (Collins, 1996; Olsen, et al., 2002). The primary cytokine responsible for modulating the cell-mediated immune response is IFNy (Stabel, 2000; Storset, 2003). IFNy assays have been successfully developed and used to diagnose and control bovine tuberculosis (Wood, et al, 1990). Likewise, IFNy assays have been developed for diagnosing MAP (Collins, 1996). Unfortunately, MAP shares many antigens with other mycobacteria commonly found in the environment resulting in cross-reactivity and unsatisfactory test sensitivity and specificity. Studies optimizing the antigen formulations used for the JD IFNy assay to improve test sensitivity and specificity are ongoing (Stabel and Whitlock, 2001; Jungersen, et al., 2002; Kalis, et al., 2003). Aside from diagnosing MAP infected cattle prior to the onset of shedding, an accurate lFNy assay for JD would be valuable for routine monitoring of young heifers as an aid in evaluating the effectiveness of control programs (Jungersen, et al., 2002). C. Summary of testing In general, diagnostic tests used for identifying MAP infected animals have excellent specificity, but only marginal sensitivity when used for screening populations (Collins, et al., 2006). Across the board, test sensitivity improves dramatically when used to confirm a diagnosis in an animal in the clinical stages of JD (Collins, 1996; Whittington and Sergeant, 2001; Tiwari, et al., 2006). The reason for the less than desirable sensitivity is due more to the pathobiology of MAP, than any innate fault of the 32 respective tests. MAP has an extremely long incubation period and subclinical stage during which infection cannot be detected The majority of infected cows in a herd are in the silent or subclinical stage of the disease, while only 10-15% of cows reach the clinical stage (Whitlock, 1992). In some ways JD is similar to cancer; the more advanced the disease, the easier it is to diagnose, but the worse the prognosis for the patient. It is not uncommon for the results of different JD tests, run concurrently, to disagree (Pinedo, et al., 2008). This is a function of both the tests and the pathobiology of the bacteria. Take, for example, fecal culture and the ELISA test. Fecal culture detects the actual bacteria, while the ELISA test detects antibodies, or the immune system’s response to the bacteria. The onset of MAP shedding does not necessarily coincide with the production of antibodies. Test agreement will only occur when these two events overlap. It is important to keep this in mind when choosing which tests to use, and interpreting the results (Rossiter and Burhans, 1996). Testing for JD is expensive and often represents the biggest cash cost of a control program (Rossiter and Burhans, 1996). Multiple testing strategies have been proposed for diagnosing JD, including the pooling of samples to reduce cost, or running different tests in parallel or sequence to improve overall sensitivity (Collins, 1996; Rossiter and Bruhans, 1996; Wells, et al., 2002a; Kalis, et al., 2004; Tavompanich, et al., 2004; Tavompanich, et al., 2008; van Schaik, 2007). Each strategy has its own merit, and there is no one best strategy to fit all. Choosing which test(s) and testing strategy to use needs to be made on a case-by-case basis (Rossiter and Burhans, 1996); taking into consideration: the purpose of testing (confirming a diagnosis vs. screening for control and management), costs, and the goals and capabilities of the operation (Collins, et al., 33 2006). Testing can play an important role in a JD control program, but only if the test results are utilized. If JD test status does not guide action to prevent the spread of the disease, testing as part of a control program is useless and a waste of money (Rossiter and Burhans, 1996). Implementation of farm management practices to control Johne ’s disease Regarding the control and prevention of JD, treatment using therapeutic agents is not practical or efficacious (St. Jean, 1996). Vaccination is controversial, only partially protective, and is generally considered a band-aid at best for JD control (Stabel, 1998; Collins, 1994). Diagnostic testing and culling of test positive animals facilitates the removal of the most infectious animals from the herd, and reduces disease burden (Holmes, et al., 2004; Jubb and Galvin, 2004); but is not very effective in eliminating the disease (Groenendaal, et al., 2002; Collins, 2003; Dorshorst, et al., 2006; McKenna, et al., 2006; Kudahl, et al., 2007). Instead, control of JD must focus on implementing farm management practices that minimize the transmission of MAP to susceptible animals (Thoen and Moore, 1989; Collins, 2003; Hoe and Ruegg, 2006; McKenna, et al., 2006). Before farm management changes for the control of JD can be recommended, a full understanding of the disease, its reservoirs, and the factors associated with increasing or decreasing the risk of infection is necessary. The pathogen and the disease have already been discussed Attention will now focus on the environmental reservoir and risk factors for JD specific to dairy herds, along with a brief discussion of recommended control practices. 34 A. Environmental reservoir offilAP on dairy herds It is generally accepted that the primary route of MAP infection is through the ingestion of bacteria from a contaminated environment (Sweeney, 1996; Step, et al., 2000; Whittington and Sergeant, 2001; Olsen, et al., 2002; Greig, 2005). Thus, the environment is a major reservoir for infection. Understanding how long MAP can survive, under what conditions, and the areas of the farm that are commonly contaminated is critical for developing strategies to minimize or eliminate exposure of susceptible animals to the bacteria. As previously discussed, MAP is an obligate intracellular pathogen and does not replicate outside the animal host (Whittington and Sergeant, 2001). The thick bacterial cell wall of MAP enables it to withstand exposure to environmental elements for extended periods of time. Substrate (feces, urine, water, milk), temperature, and pH are all factors that influence the length of time MAP will survive in the environment. (McKenna, et al., 2006). Documented survival times in farm environments include: river water — 163 days; pond water — 270 days; feces incorporated with black soil — 1 1 months; urine — 7 days; low ambient temperatures (<14 C) - >1 year (Chiodini, et al., 1984). While MAP has been cultured on pasture for more than a year following the removal of all livestock, the capacity for infectivity declines significantly after six months, provided MAP is not continuing to be excreted into the environment (Whittington, et al., 2003). While MAP may be hardier than many other pathogens, it is still susceptible to long-term desiccation, large fluctuations in temperature; repeated freeze-thaw cycles, exposure to sunlight, and soils with alkaline pH or low iron content (Richards and Thoen, 1977; Johnson-Ifearulundu and Kaneene, 1997; Johnson- 35 Ifearulundu and Kaneene, 1999; Whittington, et al., 2003; Ward, et al., 2004; Grewal, et al., 2006; McKenna, et al., 2006). On infected dairy farms, MAP has been isolated fiom many different areas including: return alleys from parlor; holding pens; high-traffic alleyways, sick cow pens; maternity pens; post-weaned calf pens; manure storage areas; and manure handling equipment (Raizrnan, et al., 2004; Berghaus, et al., 2006; Lombard, etal., 2006). The areas most commonly contaminated with MAP were manure storage areas, holding pens, and high-traffic cows areas where manure accumulated from adult cows on a daily basis (Raizman, et al., 2004; Lombard, et al., 2006). Also, there was a positive association between the distribution of MAP contamination in the environment and within herd JD prevalence (Raizrnan, et al., 2004; Fyock, et al., 2005; Berghaus, et al., 2006). To summarize, MAP is capable of surviving for extended periods of time in the environment of infected dairy herds, serving as a reservoir of infection for susceptible cattle. It is widely distributed in the environment of dairy farms. MAP is often found in areas where adult cows, the animals most likely to be shedding the bacteria, are housed. It is not uncommon for it to also be found in areas to which young calves, the animals most susceptible to infection, have access, such as the maternity and weaned heifer pens. Finally, the greater the environmental reservoir of MAP, the greater the infectious burden in the herd. B. Risk factors for Johne ’s disease on dairy herds The identification of factors or practices associated with increasing or decreasing the risk of MAP infection is vital information when assessing farm operations and 36 designing JD control programs. Once identified, they need to be carefully evaluated in an I attempt to explain the association; determine whether they are biologically plausible; and, most importantly from a disease control standpoint, decide if something can be manipulated to mitigate further transmission of the infection. Multiple studies have investigated the risk factors for JD. Factors and/or practices associated with an increasing risk for JD are listed in Table 1.1, and those associated with a decreasing risk of JD are summarized in Table 1.2. Increasing age or parity as a risk factor for MAP infection is consistent with the pathobiology of the bacterium and available diagnostic capabilities. The further the infection progresses, the more likely it is to be detected, and the older the animal. Several studies have associated an increased risk of JD with large herds. One explanation for this association is that the higher cattle density of larger herds contributes to a higher bacterial load in the environment, increasing the infection pressure of susceptible calves and promoting infection (Daniels, et al., 2002; Muskens, et al., 2003bb). Large herd size has also been associated with the purchase of cattle (USDA, 2005). The addition of purchased cattle is considered the primary method that JD is transmitted between herds (Sweeney, 1996). Thus, the association of herd size with JD may also be a reflection of the introduction of disease through purchased cattle, with the infection being subsequently sustained within the herd (Hirst, et al., 2004). Since the primary means of JD transmission between herds is through the purchase and addition of subclinically infected cattle, a possible way to mitigate this risk would be to screen purchased cows for JD prior to purchase. Due to the lack of diagnostic sensitivity for detecting subclinically infected cows, this has been proven to be 37 so. ..a a sewage Noam ...m 3 £255 voom ...m we .cumxutcoi ”Noom .._m 8 £250 ”32 ..E 8 593.530 Noam ...W a .__o=w2=< 588 ...MW e .5222 ”88 83.35. was s2 32 J“ 6 $8.530 coom in no .nomtox ”econ in no it: ”voom ...m B £85980 ”swoon ...m 3 .EU mooom .6:me 28 £03 Kam— .._m 8 6.3330 ”32 ...m we .536...on 602 .3 no .8380 38 ..a a $2320 ”.38 ..z a .2: ”88 ..a a £8.32 M.88 .._a a... $255 M8% new; 25 23% ”use ...A. a £530 coon .aoH can com—£2 68m in no anus—whom ”mmcom ..E no .5222 3:0: vocofimoc @0523 .9; 53880 EEoEEoU 323 8 Eamonxm 8.8m cook 8 $88 0:23, cog GEE 9.638: £553 9 oSmoaxm 5% e282 .22 $85 32:25 can 3&2 2:3 @8285 0N6 Ev: owes £390? @585:— 35.83. 3.3% ...—:2. aha—e E 33%.. 9223. he 5.: com—8.6.: :a :33 3.3983 333..— 34 «Eek. 38 moom r:W .0 .035. 5:: 833 9568“— Son £8. use com—2 Z 365w xoocm maze» :9: meow ..E «o .mEEmQ ”33 ..E 3 6.8330 8,339 :o 8:55 @2on .«o :ozmo=&< 88 doe 25 5222 M88 .3 3 Bags. 33 3.23 ”gem is .0 dome—£62“— mwai .280ch new aces—Emotiogng 2:58 .23me .32 oflaoxov maize: 38583: 3 $33.. GaoEaS—oo 3358850 .3285 boom dob Ea com—22 ”32 ..E 8 .oqcamBO co Sunni 958 :25 60¢ 38% 9 $28 .8 Eamoaxm 88. new“; 2:. 233 ”32 ..a a 55:50 :8 £832. 936 coon enema? 95 23>? 833 3:33-05 do wcszEéU boom :3 8 .Emmbfiox . .. . . .. . . . .. . 0:0 398 no mama 30:50 32.65 .38 .30 at: .38 .36 mace—m2). .32 .30 0.8330 . . . . 3:953— ..Boam v6.5.— .hzuc E 938:. 9252. he an: .5322: .3 .2: @3588.“ 332:— "325353 —._ ~35. 39 woa .ocoocux can sacs—anEéOmEE. 88 ..B a €3.32 88 ..3 a .8332 32 ..3 s 5.3330 moi 650:3. Ea suns—Emotéomésg ”32 ...m 3 6.3330 83me 9 2:: wfibaa< 0.. .«o 2%.; 3353 53» 958 we 35:8 25:3. Sum—mo. VEE mEcoom EB 38 H1630 wfigo—o Ens—gem mama $.58... mo wESwu—o E268“— 35.5—3— “Boar.— muuog his. E 83%.. £253. no 1%.. uomaouuuc a 5?» @8398»: €32:— "NA 935. 4O of minimal benefit (Carpenter, et al., 2004). Instead it is advocated, rather than testing individual cows that are to be purchased, performing some type of herd screening test on the herd of origin (environmental culturing; ELISA test on subset of adult cows), or buying cows only from herds with a known low prevalence (Carpenter, et al., 2004; Kovich, et al., 2006). The increased risk of MAP infection associated with the observation of clinical signs, or confirmed cases of JD (unless in purchased cows), is indicative of an established MAP infection within the herd. Cows with clinical JD are only the “tip of the iceberg”; they have likely been infectious for several months prior to the onset of clinical signs, potentially infecting several of their herdmates (Whitlock, 1992). Subsequently, routinely culling cows with clinical signs of JD has been associated with a decreased risk of JD (Muskens, et al., 2003b). The age of onset of clinical signs can be used as a predictor of the infection rate within a herd. Although relative, the observation of clinical signs of JD at an earlier age suggests a high infection rate in the herd where the cow was born; translating to high infection pressure for young calves (Collins, 2003). The main route of MAP transmission is fecal-oral (Sweeney, 1996). Therefore, any practice promoting the exposure of susceptible animals (young calves) to MAP contaminated feces (most likely from adult cows) will be associated with an increased risk of JD. This could be directly, particularly in group maternity pens (Cetinkaya, et al., 1997; Wells and Wagner, 2000), or indirectly through contaminated feed or equipment (Obasanjo, et al., 1997; Nielsen and Tofi, 2007). The application of stored manure to pasture has also been associated with an increased risk of JD, and again is consistent with fecal-oral transmission (Obasanjo, et al., 1997; Daniels, et al., 2002). Potential for 41 exposure to MAP contaminated feces also likely plays a role in the increased risk of JD with access to unrestricted group housing when compared to free-stall or tie-stall housing (Johnson-Ifearulundu and Kaneene, 1998; F redn'cksen, et al., 2004; Kobayashi, et a1, 2007; Nielsen and Tofi, 2007). In stall housing, the location of defecation is predetermined, there is less opportunity for fecal contamination of feed, and the stalls are cleaned more often than unrestricted exercise lots or manure packs (Obasanjo, et al., 1997; Johnson-Ifearulundu and Kaneene, 2000). To underscore the importance of fecal- oral transmission, frequent cleaning of pens is associated with a decreased risk of JD (Obasanjo, et al., 1997; Johnson-Ifearulundu and Kaneene, 1998). The commingling of preweaned calves as an increased risk for JD is not easy to explain. Again, it may be a consequence of some unknown characteristic or management practice on MAP infected farms (Collins, et al., 1994). More likely it is due to calf-to- calf shedding following initial transmission from an infected cow (Wells and Wagner, 2000). Recent studies have documented the shedding of MAP in young calves, even prior to weaning (Bolton, et al., 2005; vanRoermund, et al., 2005). These calves could easily spread the infection to other susceptible calves if housed together. Calves born to MAP seropositive dams were 6.6 times more likely to test seropositive in their lifetime (Aly and Thurmond, 2005). Another study determined that the darn’s JD test status was responsible for significant variation in the antibody level of their offspring (Nielsen, et al., 2002b). Whether this association is due to vertical or horizontal transmission is unknown, and academic in terms of control. MAP can be transmitted in-ulero (Whittington and Windsor, 2007), and is shed in the colostrum (Streeter, et al., 1995), milk (Taylor, et al., 1981; Sweeney, et al., 1992b) and feces 42 (Sweeney, 1996) of infected cows; all of which potentially exposes their offspring to infection. The protective effect of feeding milk replacer in decreasing JD is most probably due to the potential for MAP to be transmitted in milk (Taylor, et al., 1981; Sweeney, et al., 1992b). It has long been a common practice for farms that feed whole milk to calves, to feed waste milk. This is milk that cannot be sold for human consumption and would include transitional milk from fresh cows, mastitic milk, or milk from cows that were sick and contains antibiotic residues. Shedding of MAP into milk occurs more frequently just following parturition (Harris and Barletta, 2001) or when otherwise stressed (McKenna, et al., 2006). Hence, the practice of feeding whole milk to calves often results in feeding the most infectious milk to the animals most susceptible to becoming infected (Ridge, et al., 2005). Pastuerization of waste milk can be cost-effective (Godden, et al., 2005) and has been shown to eliminate (Stabel, et al., 2004) or at least significantly reduce (McDonald, et al., 2005) the infectious load; making it one option for JD control. The other, commonly used option to minimize the infectious dose of MAP in calves’ diets is to feed milk replacer. Certain breeds of cows, Jerseys and Guemseys in particular, have been associated with an increased risk for JD (Cetinkaya, et al, 1997). The reason for this is unknown, although several hypotheses have been proposed. One possible explanation is there is inherent variation in the susceptibility of cattle to MAP (Koets, et al., 2000), and these breeds are somehow genetically predisposed to infection. Another explanation suggests it has nothing to do with the genetic susceptibility of the breed, but rather with the management practices of herds with these breeds (McKenna, et al., 2006). Jersey and 43 Guernsey herds tend to be smaller and have a lower culling rate resulting in a higher average herd age, a factor also associated with JD (Cetinkaya, et al., 1997). Differences in herd management practices may also explain the increased risk of JD in commercial herds as compared to registered herds. Commercial herds tend to be larger than registered herds, and are more likely to purchase cattle on a routine basis (Obasanjo, et al., 1997); both risk factors for JD already discussed Other domestic ruminants, such as sheep and goats, can readily become infected with MAP, as can wildlife. Interspecies transmission has been documented (Williams, et al, 1983; Daniels, et al., 2001). So if infected with MAP, these animals could play the same role as subclinically infected cattle in JD transmission when cows are exposed to these animals. Exposure of cattle, or their feed stores, to wildlife to the extent of promoting interspecies JD transmission could also be an indirect reflection of other substandard herd management practices for minimizing the spread of MAP infection. Finally, multiple anecdotal reports have associated the application of lime to pastures and other cattle housing areas with decreasing the number of clinical cases of JD (Jansen, 1948; Kopecky, 1977; Richards, 1989). The mechanism for lime decreasing the incidence of JD is unknown, but is believed to be connected to an increase in the environmental pH (Johnson-Ifearulundu and Kaneene, 1998). Studies have shown that MAP survives better in acidic conditions (J ohnson-Ifearulundu and Kaneene, 1997; Ward and Perez, 2004). The theory is, as environmental pH increases, the bioavailabilty of iron is decreased. Iron is essential for MAP survival, so limiting iron exacerbates MAP destruction (Johnson-Ifearulundu and Kaneene, 1998). 44 C. Farm practices to control Johne ’s disease The list of farm management practices recommended for controlling JD is extensive (Rossiter and Burhans, 1996; Benedictus and Kalis, 2003), and it can be overwhelming to producers (Ridge, et al., 2005). In the end, the goal of each of those practices is the same; prevent, or minimize, the exposure of susceptible animals to MAP. Obviously, there are many ways to go about achieving that goal; and what works for one operation may not work for another. This means the management practices implemented as part of a JD control program need to be designed specifically for each operation, taking into consideration the JD burden in the herd, the risk for MAP transmission, the goals of the operation, and the resources available both in terms of money and manpower (McKenna, etal., 2006). The reason most JD control programs fail is because they were not designed to meet the unique needs and capabilities of the operation (Rossiter and Burhans, 1996; Collins, 2003). Designing a JD control program consists of three steps. The first step is to have an open and frank discussion with the producer, and determine what the operation’s goals are in regards to JD (Collins, 1994). Implementing a JD control program is long term commitment, and the producer must understand and be willing to make that commitment (Collins, 1994; Jubb and Galvin, 2004) The second step to is assess the risk of JD transmission on the operation (Collins, 1994). Over the years, different risk assessment tools have been developed for JD and assessed using a logical, scientific, systematic approach similar to that used in the successful beef and dairy milk and meat quality assurance programs (Pence, et al., 2004; Berghaus, et al., 2005; Raizrnan, et al., 2006). Recently, in the US, the National Johne’s Disease Working Group put together a 45 consensus risk assessment for JD, which has been approved by the USDA for use in the National Voluntary Johne’s Disease Control Program (Appendix A). The third, and final, step is to recommend farm management practices that are most likely to minimize the spread of MAP on that specific farm (Collins, 1994). Almost all the recommended farm practices for controlling JD are based on what is currently known about the pathogenesis of MAP, how it is transmitted, and factors associated with increasing or decreasing risk of infection. Validation of control practices in the field is limited, due to the chronic nature of the disease and the diagnostic difficulty in identifying infected animals (Ridge, et al., 2005). Instead, farm practices to control JD have been simulated, using existing theory for control, expert Opinion, and the limited field data available (Collins and Morgan, 1992; Groenendaal, et al., 2002; Dorshorst, et al., 2006; Kudahl, et al., 2007). The general consensus of all (expert research opinion, observational field studies, and simulated studies) is that improved calf hygiene is a critical component of any JD control program (Thoen and Moore, 1989; Collins and Morgan, 1992; Collins, 1994; Goodger, et al., 1996; Groenendaal, et al., 2002; Jubb and Galvin, 2004; Pence, et al., 2004; Ridge, et al., 2005; Dorshorst, et al., 2006; Kudahl, et al., 2007). Johne’s disease control starts with breaking the chain of infection (Kudahl, et al., 2007). Given that young cattle are more susceptible to infection with MAP than older cows, management should focus on roughly the first six months of life (Collins, 1994). This means eliminating, or minimizing contact of neonatal and young calves with colostrum, milk and/or feces from infected adult cows (Pence, et al., 2004; McKenna, et al., 2006). How this is done will likely vary by farm (Ridge, et al., 2005). Some of the more common and 46 easily implemented changes are: prompt removal of calf from darn after birth, cleaning maternity pen afier each use, housing calves in separate pens well away from contact with adult cattle, only feeding colostrum from JD test negative cows; and feeding calves milk replacer or pasteurized whole milk (Collins, 1994; McKenna, et al., 2006). Aside from calf management, other commonly recommended practices to control JD are to cull all cows with clinical signs of weight loss and, diarrhea and improve overall farm cleanliness; thereby removing the most infectious animals and reducing the environmental reservoir of MAP (Collins, 1994; Goodger, et al., 1996; McKenna, et al., 2006) In simulation models, improving calf hygiene was more cost-effective than testing (Groenendaal and Galligan, 2003; Dorshorst, et al., 2006); although improving calf hygiene and the use of a test-and-cull strategy provided the quickest means of control (Collins, 1992). Diagnostic testing for JD control purposes is not always necessary and ' should not be recommended in all herds (Dorshorst, et al., 2006). To support these simulations, a field evaluation of the Victorian (Australia) Johne’s Disease Test and Control Program (TCP), which consisted of testing and culling MAP positive animals along with improving calf management, found that within herd JD prevalence and the incidence of clinical cases did not decline significantly until the herds consisted mainly of cows born after the TCP was started (Jubb and Galvin, 2004). In short, farm management practices to control JD must focus on eliminating or minimizing exposure of calves to MAP (Collins, et al., 1994; Pence, et al., 2004; McKenna, et al., 2006). It is unlikely the success of a JD control program will depend on any single management change. Instead success will depend on a series of changes, each 47 with different degrees of importance, and that will be different from farm to farm (Ridge, et al., 2005). Thus, JD control programs must be designed specifically for each herd, only after understanding the goals and capabilities of the operation and an assessment of the areas at greatest risk for JD transmission on the farm is performed (Collins, 1994; Rossiter and Burhans, 1996; Ridge, et al., 2005). Finally, a JD control program is a long term commitment, and it may take years before the program has noticeable impact on within herd JD prevalence and/or incidence (Collins, 1994; Judd and Galvin, 2004). 1.6. Economics of Johne’s disease The economic costs of JD to producers can be divided into two broad categories: (1) the economic costs due to the disease as a result of impaired productivity and performance, and (2) the economic costs associated with diagnosing and controlling the disease. The decision to invest in a JD control program will often hinge on the magnitude of the difference between these two categories. One can look at the economic costs due to JD as an estimate of the potential benefits of controlling the disease (Groenendaal and Wolf, 2008 in press). In other words, if JD was eradicated, the producer could potentially realize an increase in revenue equal to the estimated losses caused by the disease. Ifthe cost associated with diagnosing and controlling JD is greater than the potential benefits of reducing or eradicating it, investing in a control program may not be a sound economic decision. In short, if controlling JD costs more than what the disease is costing the producer in terms of lost production and performance, it will be difficult to convince him to implement a control program. 48 There have been numerous studies attempting to define and quantify the economic costs due to lost production and performance as a result of JD (see Table 1.3). Few, have attempted to quantify costs of diagnosing and controlling the disease, and those that have are based on expert opinion and assumption; not on real farm data (Benedictus, et al., 1987). Production and performance losses due to Johne ’5 disease Few studies in the literature have attempted to quantify, in monetary terms, the production losses caused by JD because such economic indices are so unstable from year to year and across different regions (Hasonova and Pavlik, 2006). Instead, most have addressed production losses qualitatively, and then tried to estimate the magnitude of each qualitative loss. The reported impact of JD on common dairy production and performance indices varies greatly from study to study, and sometimes even within the same study (Spangler, et al., 1992; Hendrick, et al., 2005c). The most likely reason for the discrepancy of reported results is again, due to the chronic nature of JD and the associated difficulty of identifying MAP infected animals. The reported outcomes were dependent on: study design, the population being studied (cull cows vs. cows retained in the herd; subclinical cows vs. cows with clinical signs), prevalence of JD in the population, and the method used to identify infected cows; none of which were uniform or standardized Therefore, it would be a mistake to try to make direct comparisons between studies. Instead the literature has been evaluated qualitatively and the results summarized in Table 1.3. 49 mace 358m » 50.52— ouaouoE—c :2 3.2525— oouafihotofiguuaeea 395:5 3 9:23: totem 2a 82.2805 £335.23 8553:2— uea 5.335.:— mhae .8 038:. 9253. .3 82...: 2: :e FEES—a 333.35 x”.— 035—. 50 82 ...a a ..5 82 ..a .0 .=e§_§e.-§§o. 3.3 :6. .u .60 3a. :6. .o .66..=_an.:-..8=:2. $2 ...a a $20.85.. 86m :6. .o 66:60 omoom a... .0 £3.65: mom. ...m .o .6823 as. ....w a ..22.5. mg— .6855 6S. :owbsm boom :6. .o ......ENEM 88 r... .o 566223368662. £352 2:? .30 25E: 3:88... boom :6. .o .832 wnfl .6855 6.... .3925 so. ...... a .886 32 .._a a .9202 m8. ...... .... .332 582.23.. 88 .... a .2256 88 ...." .... .2353 32 ...a a 5.762 omoom r... .0 £2.65: who. £3.50 6.... .3925 82 ...a a 58:3 08. ...a a .2222 mg. ...a a .8622 8:32.68 ”6.8.338 3.85938 638...: ... .830: 638...._:.. 6.... 638,—... «6.85938 6890.... ... 8.5.5 :8an 8.5.5.56 62 639.63.... QOGGEHOV-oA—F—OmuOS—vchm . Amw=m6.._.. 8 966.83.. 68.5... 9... 82.9.0.3: 6.5.0.632. 855.83.. 6.... 5.8.6:...— f_a6 .... 3.68.6 92:6... .... .23.... a... ..e .3555... 253.130 262.5283 .3 «BE. 51 By far the most consistent finding in the literature was decreased milk production by MAP infected cows. In the four studies in which no production difference was noted between infected and uninfected cows (McNab, et al., 1991; Spangler, et al., 1992; Johnson, et al., 2001; Hendrick, et al., 2005c), all were based on comparing subclinically infected cows, as determined by serum ELISA, to test negative cows. In one study (Johnson, et al, 2001), the mean average parity of the study herds was <2, and it was hypothesized, that in “young” herds, subclinical MAP infection may have “little impact on milk production.” In two studies, when infection status was determined by fecal culture (Spangler, et al., 1992; Hendrick, et al., 2005c) or milk ELISA (Hendrick, et al., 2005c), milk production was significantly lower in test positive cows compared to test negative cows; yet, in the same studies, there was no production difference between infected and uninfected cows when infection status was determined by serum ELISA. This underscores the difference in sensitivity between different diagnostic tests, potentially leading to a lack of agreement when they are run concurrently. Only a few studies reported the magnitude of reduced milk production in MAP infected cows compared to uninfected cows. The milk production loss in cows with subclinical JD ranged from 2-6% (Nordlund, et al., 1996; Hendrick, et al., 2005c). For cows culled with clinical JD, the reported loss in milk production was 14% (Raizman, et al., 2007). One study compared milk production in cull cows for the current lactation and the two previous lactations respectively. Cows with subclinical JD produced 6% less milk in the lactation they were culled as compared to the next previous lactation, and 16% less milk than the second previous lactation. Cows with clinical JD produced 5% less milk in the lactation during which they were culled compared to the next previous 52 lactation, and 19.5% less than the second previous lactation (Benedictus, et al., 1987). In one study, the comparison was made based on fecal culture test status without reporting clinical status. Milk production loss in this study ranged from 6.6-14% (Wilson, et at., 1993). These results suggest that the drop in milk production gets worse as the severity of the disease progresses, and is consistent with an increasingly negative energy balance. Also consistent across studies, was a decrease in the productive lifetime of cows infected with JD (Buergelt and Duncan, 1978; Korrnendy, et al., 1989; Wilson, et al., 1993; Hendrick, et al., 2005c; Gonda, et al., 2007). This was generally reported as an increased risk of culling for infected cows. No attempt was made to quantify this loss in any of the studies, but was theorized to be significant due to suboptimal culling resulting in the lost future production of the cow culled and the cost of replacing her. Three studies reported a lower cull value for cows with JD. This was due to weight loss resulting in lower slaughter weight (Johnson-Ifearultmdu, et aL, 1999), or poorer body condition (Ott, et al., 1999). One study reported a 30% reduction in the cull value of cows with clinical JD (Benedictus, et al., 1987). In another, a 10% increase in serum ELISA JD test prevalence corresponded with a 33.4 kg (73.5 lb) decrease in the mean cull cow weight for the herd (Johnson-Ifearulundu, et al., 1999). Also uncontested across the studies evaluated, was the finding of an increased mortality rate in JD infected dairy herds. In one study the mortality rate was 3% higher in JD infected herd as compared to uninfected herds. In the 1996 US National Animal Health Monitoring Service (NAHMS) dairy study, the mortality rate in herds with a “low-clinical” rate of ID was 15% greater than in uninfected herds; while in herds with a 53 “hi gh-clinical” rate of JD, the mortality rate was 45% greater than in uninfected herds (Ott, et al., 1999). The impact of JD on the other parameters assessed in the studies was not so clear cut. Most controversial was the impact of MAP infection on udder health. Two of the three studies that reported an increase in the incidence of mastitis in infected cows were comparing fecal culture positive clinical cows to culture negative cows (Merkal, et al., 1975; Buergelt and Duncan, et al., 1978). This would suggest that the cows were in the more advanced stages of the disease. For the four studies in which MAP infection was found to have no significant effect on udder health or the incidence of mastitis, two of them were comparing subclinically infected cows to JD test negative cows (N ordlund, et al.., 1996; Hendrick, et al., 2005c); one did not report the clinical status of the infected group, but did report the body condition score of >85% of the infected cows was “normal” (Lombard, et al., 2005); while in the remaining study, the clinical status of the test positive cows was not reported (Gonda, et al., 2007). Thus, it would seem that the disease process in the infected cows in these studies was not as advanced. Perhaps the reason for the conflicting findings regarding udder health across these studies is due to the immune status of the infected cows. The more advanced the JD process, the more compromised the immune system becomes, making the infected cow more susceptible to other infections, such as mastitis. However, the reason for MAP infected cows having a reduced incidence of mastitis in one study remains unexplained (Wilson, et al., 1993); but the study consisted of only one herd, so it may be a phenomenon specific to that herd and should be interpreted accordingly. 54 The most controversial findings in the literature concerned the impact of MAP infection on reproductive performance. In many of the studies, reproductive performance was poorly defined as simply “infertility” (Merkal, et al., 1975; Buergelt and Duncan, 1978) or “poorer reproductive performance” (Raizman, et al., 2007) so it is difficult to contrast the studies and draw any conclusions. It seems reasonable as JD progresses, and the cow enters into an increasingly negative energy balance, that reproductive performance would be adversely affected, resulting in potential economic losses. To summarize, the economic impact of MAP infection on the production and performance of dairy cattle is due primarily to reduced milk production and cull value of infected cows, resulting in increased replacement costs (Ott, et al., 1999; Wells and Wagner, 2000). Other losses due to JD associated with concurrent disease and reproductive performance are possible, but the literature is less clear, and sometimes contradictory, regarding the direction of the infection’s impact Thus, these economic losses are probably minor compared to those resulting from reduced milk production and cull value. Finally, the magnitude of the economic costs caused by JD, are positively correlated with the stage of the disease; the more advanced the infection, the greater the economic costs (Ott, et al., 1999). Economic costs associated with production losses caused by Johne ’3 disease The estimated monetary costs of the production and performance losses caused by JD have been estimated at both the industry and herd levels. The estimates vary depending on study design and what was or was not included in the calculation. 55 Based on data collected in the 1996 NAHMS dairy study, it was estimated that JD costs the US dairy industry $200 — 250 million annually (Ott, et al., 1999). The estimated national prevalence of MAP infected dairy herds at the time was 21.6% (USDA, 1997). In the most recent NAHMS dairy study in 2007, the estimated prevalence of MAP infected dairy herds was 68.1% (USDA, 2008); a roughly three-fold increase. The associated estimated economic costs of JD have yet to be released, but it is probable that with more infected herds, the costs will have gone up as well. If a linear association is assumed between costs and prevalence, JD could currently be costing the US dairy industry $600 — 750 million per year. As with the assessment of JD control programs, simulated models have been developed to assess the economic costs associated with JD at the herd level. In a Canadian based model, the estimated cost of JD was (US equivalent) $33 per cow (Chi, et al., 2002a). A second simulation model estimated the cost of JD to an average rnidsize US dairy to be $30 per cow in inventory the first year following introduction of the disease into the herd, and increasing to $70 per cow per year by year 20 after infection in the absence of a control program (Groenendaal and Galligan, 2003). These data were in close agreement to those reported in field studies. Data from the 1996 NAHMS study was also used to calculate the cost of JD at the herd level (Ott, et al., 1999). Herds infected with JD lost $97 per cow in inventory per year as compared to uninfected herds. The economic costs increased to $245 per cow in inventory for JD infected herds with a high prevalence (>10% of cull cows having clinical signs). When aggregated across all cows in the US, the economic cost of JD was estimated at $22 — 27 per cow. The 56 economic losses due to JD reported in other studies ranged from $20 — 26 per cow after standardizing the milk price and cull value used in the calculation (Ott, et al., 1999). Economic costs of Johne’s disease diagnostic and control programs Diagnostic testing ofien represents the largest cash cost of a JD control program, and should only be done if the test results are going to be used to guide management decisions (Rossiter and Burhans, 1996). Testing is not a necessary component in all JD control programs. Although, it will aid in reducing the disease burden in the herd by identifying the most infectious cows so they can be culled (Collins, 1992; Rossiter and Burhans, et al., 1996). Cost is almost always the deciding factor on whether to undertake a JD testing strategy. For many herds, low costs tests are more useful than more sensitive, but more expensive tests (Dorshorst, et al., 2006). The current costs (2008) of the most commonly used J ohne’s diagnostic tests offered by the USDA certified Johne’s testing laboratories in Michigan are summarized in Table 1.4. Published, real farm data on the cost of implementing management practices to control JD, and their impact on the JD burden within dairy herds is lacking. The production losses caused by JD are substantial, and would seem to warrant disease control efforts. However, further research is needed on the costs of changing herd management to control JD, in terms of capital, supplies, and labor; before the cost effectiveness of control programs can be determined. 57 Table 1.4: Costs of commonly used Johne’s disease diagnostic tests offered by the USDA certified Johne’s testing laboratories in Michigan (prices current as of January 2008) ELISA Fecal Culture Lab PCR Serum Milk Standard Liquid Geagley Lab, Michigan Department of $6 NA $16 NA NA Agriculture Diagnostic Center for Population and Animal Health, $6 NA NA $23 NA Michigan State University . $30 — Antel Bro $6 $6 $35 NA $100,, NA: Not available * $30 for results in 2 weeks, $100 for results in 3 days 1.7. Zoonotic potential of MAP J ohne’s disease is classified as a reportable, but non-actionable, disease in many states in the US (Step, et al., 2000). It is classified by the Office of International des Epizooties (OIE) as a list B disease, meaning it has the potential for substantial socioeconomic or public health consequences (Wells, et al., 1998). Part of the reason these reporting classifications were made is the ongoing concern that MAP is a zoonotic pathogen. Other mycobacteria (M. tuberculosis, M. bovis, M. leprae, M. avium, M afi'icanum) are zoonotic (Hugh-J ones, et al.,l995), so it is not inconceivable that MAP is as well. More importantly, over the years, there has been a growing body of evidence 58 linking MAP to Crohn’s disease, a chronic granulomatous ileocolitis, in people (Chiodini and Rossiter, 1996). In a recent, comprehensive, meta-analysis, a specific and positive association was found between MAP and Crohn’s disease, but the causal role remained undetermined. MAP could be a causative agent of Crohn’s disease, a secondary pathogen exacerbating the disease, or an incidental colonist (Feller, et al., 2007). 1.8. Conclusion With most infectious diseases, eradication is the ultimate goal. However, many experts question if JD eradication is practical or possible (Collins, et al., 2006). Whether eradication is possible or not, the first step needs to be JD control. The increasing number of herds infected with JD suggests that the US cattle industry is a long ways from controlling JD. Thus, further research on how best to manage and control ID in a realistic and cost-effective manner is warranted. 59 CHAPTER 2 Pillars, R., Grooms, D.L., Kaneene, J .B., in press. Longitudinal study of the distribution of Mycobacterium avium paraluberculosis in the environment of dairy herds participating in the Michigan Johne’s Disease Control Demonstration Herd Project. Can Vet J. 60 CHAPTER 2 LON GIT UDINAL STUDY OF THE DISTRIBUTION OF M Y C OBA C TERI UM A VIUM PARA TUBERCULOSIS IN THE ENVIRONMENT OF DAIRY HERDS PARTICIPATING IN THE MICHIGAN JOHNE’S DISEASE CONTROL DEMONSTRATION HERD PROJECT 61 2.1 Abstract The objective of this study was to describe the distribution of Mycobacterium avium paratuberculosis (MAP) in the environment of infected dairy farms over time. J ohne’s Disease (JD) prevalence was monitored annually in seven Michigan dairy herds. Environmental samples were collected bi-annually and cultured for MAP. A total of 731 environmental samples were cultured, of which 81 (11%) were positive. The lactating cow floor and manure storage were the areas most commonly contaminated, representing 30% and 33% of positive samples respectively. When herd prevalence was >2%, MAP was cultured from the lactating cow floor and/or manure storage 75% of the time. When herd prevalence was 52%, MAP was never cultured from samples collected For every one unit increase in number of positive environmental samples, within herd JD prevalence increased 1.62%. Environmental contamination with MAP is consistent over time on infected dairy farms, and management practices to reduce environmental contamination are warranted. 2.2 Introduction Mycobacterium avium paratuberculosis (MAP), the causative agent of Johne’s disease (JD), is prevalent worldwide. The National Animal Health Monitoring and Surveillance (NAHMS) Dairy 2007 study, estimated 68.1% of the dairy herds in the US were infected with MAP (USDA, 2008). This is up from 21.6% reported in the NAHMS Dairy 1996 study (USDA, 1997). Based on data from the 1996 NAHMS study, annual economic losses for the dairy industry due to JD were estimated to range from $200-250 million (Ott, et al., 1999). With increasing prevalence, economic loses are likely to 62 increase. Cattle generally become infected with MAP as young calves, but do not exhibit signs of the disease until years later (Sweeney, 1996). Due to the chronic nature of the disease, and its long incubation period, testing and culling infected animals as a method for controlling JD has been relatively ineffective by itself (Collins, et al., 2003; Dorshorst, et al., 2006; Kudahl, et al., 2007). Instead, strategies for controlling JD have focused on minimizing the exposure of calves, the animals most susceptible to becoming infected, to MAP, thereby preventing new infections. While calves can become infected with MAP in utero (Seitz, et al., 1989; Sweeney, et al., 1992), or through ingestion of colostrum or milk from infected cows, this generally only occurs when the dam is in the latter stages of the disease (Sweeney, et al., 1992; Streeter, et al., 1995). It is believed most post-natal infections occur through the ingestion of the bacterium from a contaminated environment (Sweeney, 1996; Harris and Barletta, 2001). Thus, factors playing a role in transmission include the amount of MAP being shed into the environment, the location contaminated, andthe length of time the bacteria survives in that environment. As an obligate intracellular pathogen, MAP does not replicate outside the host (Harris and Barletta, 2001), but it can survive for months to over a year in the environment (Whittington, et al., 2004). Wildlife, birds (Beard, et al., 2001; Corn, etal., 2005; Raizman, et al., 2005), even invertebrates such as flies and worms (Fischer, et al., 2001; Pavlik, et al., 2002; Fischer, et al., 2005) commonly found around dairy farms can become infected with MAP, and occasionally shed the bacterium into the environment. While the amount of MAP shed by these nontraditional hosts is negligible compared to that shed by cattle (Tiwari, et al., 2006), it does represent a way by which the bacterium 63 can persist and multiply outside of the primary host. Recently, the discovery of “dormancy-related genes” in the MAP genome suggests that, in the absence of essential nutrients, MAP may enter a state of dormancy and then return to a viable, infectious state when conditions again become favorable (Whittington, et al., 2004). Under field conditions in Australia, using the sheep strain of MAP, the bacterium was cultured from pasture twelve months after removing livestock from the property (Whittington, et al., 2003). Studies have been conducted to determine the extent of MAP contamination on infected dairy farms (Raizman, et al., 2004; Berghaus, et al., 2006; Lombard, et al., 2006). The bacterium has been found in numerous locations on dairy farms including calving pens and post-weaned calf pens (Berghaus, et al., 2006), both of which are high- risk areas for transmitting the disease to the next generation of herd replacements. The areas most commonly culture positive for MAP are those where manure accumulates from adult cattle, the animals most likely to be shedding the bacterium. These include manure storage areas (lagoons, manure Spreaders) and high-traffic, common cow areas (feed alleys, holding pens, return alleys, etc.) (Raizman, et al., 2004; Berghaus, et al., 2006; Lombard, et al., 2006). As a result, targeted culturing of these areas can be used to identify MAP infected herds. In the most recent revision of the USDA’s Johne’s Program Standards (USDA, 2005), targeted environmental culturing was approved as an entry-level screening test for dairy herds desiring to participate in the Voluntary Johne’s Disease Control Program Evidence also suggests the number of positive environmental cultures, and the amount of MAP in those samples, is positively correlated with the 64 within herd prevalence (Raizman, et al., 2004; Fyock, et al., 2005; Berghaus, et al., 2006) To date, studies investigating MAP contamination on dairy farms have been cross-sectional in nature, with the environment being sampled at only one point in time (Raizman, et al., 2004; Fyock, et al., 2005; Berghaus, etal., 2006). The temporal relationship between MAP environmental contamination and within herd JD prevalence remains undefined. There is limited information on how MAP contamination in the environment changes as within herd JD prevalence changes. Therefore, the objective of this study was to characterize the distribution MAP in the environment of infected dairy farms, and describe if, or how, that distribution changes as within herd prevalence changes. The intention being to identify areas on infected farms that consistently culture positive for MAP. By understanding what areas on infected farms are consistently contaminated with MAP, even in the face of changing herd prevalence, more focused and economical herd screening programs can be developed. 2.3 Materials and Methods Farms This study was part of the larger Michigan Johne’s Disease Control Demonstration Project. A total of seven Michigan dairy herds participated in this study. Herds were selected based on the following criteria: 1) herds were known to be infected with ID upon enrollment; 2) the producer was willing to participate in a longitudinal study for at least 5 years; and 3) the herd was representative of a typical Michigan dairy 65 farm in terms of herd size and housing management. Upon enrollment, and annually thereafter, a JD risk assessment was performed for each herd. Based on the risk assessment and the individual herd’s goals and management capabilities, a JD control program was implemented on each herd and updated as necessary throughout the study. Study herd size ranged from 94-513 adult cows. Only one herd expanded significantly (231 to 445 cows) during the course of this study. Herd size for the other six herds remained fairly consistent throughout the study period. Housing management practices consisted of total confinement (4 herds), combination of confinement and grazing (2 herds), and one rotational grazing herd which was confined during the winter months. Confinement housing consisted of free stalls (6 farms) or a combination of tie stalls and free stalls (l farm). Determination of Herd Prevalence Fecal culture was performed on all adult cows in each herd annually. Prevalence was calculated as the number of cows with positive fecal culture results, divided by the total number of cows tested that year. Environmental Sampling Every six months enviromnental samples were collected from each farm At each visit, one sample was collected from the feeding area, primary water source and floor from each of the following areas: pre-weaned calf, weaned heifer, maternity, and lactating cow. A sample from the primary manure storage area (generally a lagoon or manure spreader) was also collected. Thus, a total of 13 environmental samples were 66 collected at each herd visit. In addition, samples of pasture, pasture water sources, deer feces, and recycled sand bedding were collected and cultured when appropriate. An attempt was made to get as representative a sample from each designated area as possible. For feed and flooring samples respectively, a clean, gloved hand was used to collect 10 random “gr ” samples from various locations in the designated area. The samples were mixed together thoroughly and placed in 720 ml sterile Whirl-Pak bags. A composite sample from all sources (buckets, water tanks, automatic waterers, ponds, etc.) providing drinking water to cattle in a given area was collected in a sterile, 1L bottle. The water sample was thoroughly agitated before filling a 120 ml plastic specimen cup and submitting for culture. For manure lagoons, samples were collected 15 cm below the surface from 4-6 different locations and pooled to fill a 120 ml specimen cup. For manure Spreaders, a 120 ml sample was collected from the heaters (box Spreaders) or dispensing area (liquid spreaders). For recycled sand bedding and pastures respectively, five random “grab” samples were collected from the surface and five underlying the surface at depths varying from 6-24 cm. All samples from each respective area were mixed together in a clean bucket and a pooled sample placed in a 720 ml sterile Whirl- Pak bag for culture submission. During each farm visit, the farmstead, particularly around feed storage areas, pastures, fields and any adjacent woods where deer sightings were reported, were walked and samples of deer feces collected when found. Environmental samples were collected from January 2003 through November 2006. 67 Bacterial Culture All fecal and environmental samples were submitted for MAP culture to the Diagnostic Center for Population and Animal Health, Michigan State University, East Lansing, MI. Prior to June 2004, all samples were cultured on standard solid culture using Herrold’s Egg Yolk (HEY) media. Thereafter, samples were cultured using the ESP® culture system 11 (ESP II, TREK Diagnostics Systems, Inc., Cleveland, OH). Processing and decontamination of samples prior to inoculation of culture media was the same throughout the study, and consisted of a modification of the Cornell method described previously (Stabel, 1997). Briefly, 2 grams of each sample was added to 35 ml sterile distilled water. The sample was vigorously shaken for 15 seconds, and then allowed to set at room temperature for 30 minutes. Five ml fiom the center of the supernatant was pipetted into a centrifuge tube containing 25 ml l/zXBI-II-HPC (half strength brain heart infusion broth with 0.9% l-Hexadecylpyridinium) and gently mixed. Tubes were incubated at 35-3 7° C overnight. Samples were then centrifirged at 3000 G for 20 minutes at >22° C. The supernatant was decanted. One ml of antibiotic mixture (50 pg amphotericin B, 100 pg vancomycin, and 100 pg naladixic acid in 1/2 XBHI) was added to the sample and vortexed to resuspend the pellet for final decontamination. Samples were incubated at 35-37° C overnight before inoculating onto culture media. Culture positive samples were confirmed as MAP using Kinyoun’s acid-fast stain and real-time PCR for the 18900 insertion sequence. Real-time PCR was performed after 42 days on all signal negative ESP 11 samples. Samples were only reported as negative if they were signal negative on ESP II and negative on PCR 68 Descriptive Data Analysis Culture results were recorded in, and descriptive statistics generated, using a commercial computer spreadsheet (Microsoft Office Excel”, Microsoft Corporation, Redmond, WA). Statistical Data Analysis The number of environmental samples collected at each collection date varied across and within herds depending on the housing management (pasture vs. confinement), season (pastures were not sampled during winter months when cows were confined and/or access was restricted due to snow cover), and availability (deer feces were not consistently found on all farms). The association between the within herd JD prevalence and the number of culture positive enviromnental samples over time was therefore restricted to only those samples that were consistently collected on all farms (feed, flooring, and water from the pre-weaned calf, weaned heifer, lactating cow, and maternity areas, and manure storage area). Environmental samples were collected every six months, while herd prevalence was only calculated once every 12 months. Thus, for every year, two samples were collected from each area on the farm. For ease of analysis, environmental culture results were aggregated by calendar year and animal location (pre- weaned calf, weaned heifer, lactating cow, maternity, and manure storage areas). Using within herd JD prevalence as the outcome of interest, its association with time (study year) and the number of positive environmental samples was assessed using linear regression, controlling for repeated measures within herds using generalized estimating equations (GEE) using an exchangeable correlation structure. The regression model was 69 built starting with univariable analysis for study year and the total number of positive environmental samples each year. To determine if within herd prevalence was associated with MAP contamination in specific areas on the farm, similar univariable linear regression models were assessed using the number of positive environmental samples in the pre-weaned calf, weaned heifer, lactating cow, maternity, and manure storage areas as the independent variables, respectively. Area-specific variables with a p-value of >015 on univariable analysis were then considered in a multivariable linear regression model using step-wise backward selection. The final multivariable model consisted of only those variables with a p-value of <0.05. Model fit for all respective regression models was assessed using an extension of cumulative residuals as discussed in Lin, et al. (2002). Briefly, the cumulative sums of the residuals for each independent variable in the respective regression models were plotted, along with the residuals of 10,000 simulated realizations from a zero-mean Gaussian distribution. The Kolmogorov-type supremum test was calculated along with its associated p-value. This process was repeated with alternative functional forms of the variable based on the initial pattern of the currrulative sums of residuals in an attempt to improve model fit when warranted. The greater the Kolmogorov-type test statisitic and its p-value, the better the model fits the data, and p-values <0.05 were considered indicative of poor, or insufficient, model fit. All statistical analysis was performed using commercially available software (Proc Genmod, SAS 9.1, SAS Institute, Inc., Cary, NC, USA). 70 2.4. Results Herd Prevalence Initial apparent JD prevalence based on whole herd fecal culture in the study herds ranged from 2-1 1%. Over the four-year course of this study, apparent JD prevalence within these herds ranged from 0-42%. In one herd, the prevalence increased dramatically, from 7% to 42% in the second year of the study then gradually declined to 12% by year four. This occurred despite the herd having been closed for over 20 years and herd size remaining constant. Apparent prevalence within the herd that purchased cattle to double herd size increased slightly (9-11%) over the study period. Johne’s disease prevalence in the other five herds tended to decrease or plateau between years three and four of this study. Environmental C ulturing A total of 731 enviromnental samples were collected with 81 (11%) culturing positive for MAP. Culture results by location are summarized in Table 2.1. Over the four-year course of the study, positive environmental samples were identified on six of the seven farms. The one farm with no positive environmental samples had extremely low fecal culture prevalence, ranging fi'om 0-2%. The areas most commonly contaminated were the lactating cow floor and the manure storage area, representing 30% and 33% of the positive samples respectively. One or both of these areas was positive on 75% of the environmental collection dates. 71 Table 2.1: Distribution of Mycobacterium avium paratuberculosis (MAP) in the environment of seven Michigan Dairy Farms No. No. Location Total Location Samples Positive % % CalfFeed 51 2 3.9 2.5 Calf Floor 57 4 7.0 4.9 Calf Water 49 0 0.0 0.0 Heifer Feed 50 0 0.0 0.0 Heifer Floor 5 3 3 5.7 3 .7 Heifer Water 52 0 0.0 0.0 Maternity Feed 52 0 0.0 0.0 Maternity Floor 56 8 14.3 10.0 Maternity Water 54 5 9.3 6.2 Lactating Cow Feed 52 2 3.8 2.5 Lactating Cow Floor 54 24 44.4 30.0 Lactating Cow Water 53 2 3.8 2.5 Lagoon/Manure Spreader 53 27 50.9 33.3 Recycled Sand 5 4 80.0 4.9 Other 40 0 0.0 0 Total 731 81 11.1 100.0 72 Both of these areas were positive in the six herds with positive environmental samples at least once, and often multiple times, on different sampling dates. Ten percent of the positive enviromnental samples came from the maternity floor and 6% from maternity water samples. The maternity area was positive for MAP at least once in four of the six herds. Fecal culture prevalence in those herds at the time the maternity area was positive ranged from 54-42%. The pre-weaned calf area was found contaminated in three of the six herds. Apparent prevalence of MAP shedding in those herds at the time ranged from 86-17%. On one of the farms, the calves were housed in a group pen across an alley from a contaminated maternity pen, with the potential for cross contamination. On the other two farms, the calves were housed in separate barns, well away from any possible contamination or run-off from adult cattle. Recycled sand bedding represented 5% of the positive environmental samples; however, these samples came from only one farm with fecal culture prevalence ranging from 12-42% at the time the samples were collected. The majority of environmental samples contaminated with MAP originated from flooring or manure storage (n=70) as compared to feed (n=4) or water (n=7). Two of the positive feed samples came from the calf area adjacent to a contaminated maternity pen on a farm when within herd JD prevalence was 14%. The other two positive feed samples came from fence-line feed alleys in free stall barns housing lactating cows. All of the MAP positive water samples originated from adult cow areas, with five occurring in the maternity area and two in the lactating cow area. 73 When compiled, the number of positive environmental samples decreased as herd prevalence decreased (Figure 2.1). Once herd prevalence fell to below 2%, MAP was never cultured in the environment of any area sampled. When herd prevalence was >2%, MAP was cultured from the lactating cow floor and/ or manure storage areas 75% of the time. All the positive samples in the 2-5% herd prevalence category originated from either the lactating cow floor or manure storage areas. When herd prevalence exceeded 5%, MAP began to be isolated from areas in addition to the lactating cow floor or manure storage areas, with the most common area being the maternity floor. Within individual herds, the trend for decreasing MAP environmental contamination (based on the percent of culture positive environmental samples) with decreasing within herd JD prevalence was not always as obvious (Table 2.2). Over the course of this study, the JD prevalence within each herd changed, and the herds moved up and down across the prevalence categories outlined in Figure 2.1. For example, the <2% category represents data from three different herds; the 2-5% category, five herds; the 6-15% category, six herds; and the >15% category, three herds. 74 Figure 2.1: Percentage Mycobacterium avium paratuberculosis (MAP) positive environmental samples by within herd J ohne’s Disease prevalence Positive MAP Samples (% of samples collected from herds In each category) 25 20 15 10 <2% 2-5% 6-15% >15% Apparent Within Herd Johne's Disease Prevalence 75 Table 2.2: Percent of MAP culture positive cows and environmental samples by herd over time 2003a 2004 2005 2006 Herd . Cows Environment Cows Environment Cows Envrronment Cows Environment PM (“/o) PM We) (”/o) W») ("M (“/o) 1 10.3 0 14a 24.18 20.3 18.4 4.4 6.5 2 10.2 6.7 4.1 3.6 2.9 7.4 1.9 11.5 3 8.6 14.3 5.4 17.9 10.6 21.4 11 11.5 a a 4 10.6 0 6.4 14.8 2 0 4 7.1 5 NT NT 5.3 19.2 5 7.7 6.3 19.2 6 7 6.7 42.1 17.2 16.9 28.6 12.1 23.8 7 1.8 0 0.6 0 0.6 0 2 0 a Indicates samples cultured on Herrold’s Egg Yolk (HEY) NT Herd. not tested Statistical data analysis The results of univariable linear regression models to assess the association between herd prevalence over time and the number of positive environmental samples overall and in each respective area are summarized in Table 2.3. The results of the final multivariable linear regression model assessing the association of within herd JD prevalence with MAP environmental contamination in specific areas of the farm are shown in Table 2.4. 76 00.0 09806 0.808 .Nd memo... mwd flood- :4.— .. 6038.6 880800.300 03.60.. .00802 00... 380.08 and $8.0 an. 2 K... $0 1 6038.6 880880.300 02.60.. .00832 00... 300 @830... _ ..o nmnmd 3N mwd- 8.. I m0.08.... 300800.680 02.60.. 008:2 00... 0.20.. 608.03 mm... mm 3.0 mm... 0. ._ 5.0 I 603806. 300800.300 03.60.. .00802 00.0 :00 608.03 cm... wooed 00.6 :0.— _o.m -0... I .0388 3808:0880 02.60.. 008.2 mm... SoodV Ned mm... no; :30. - 60388 380800.300 0360.. 008:2 MW... EN... .2. .... . - 8...- so. .86 02.3... 60. .085 .030. 8080...... 0.....->0.0w08_0v_ 03.3-.. :83 00:06:80 $3 0.0865 0602 0800...: 00 02.0.0.0...— 08086 90.50.. 6.0.. 85.3 9.6.. 06.62... 560050.. .00.:— 0_._.._..a>_.5 .m.~ 030,—. 77 00.0 b.8808 o... w .8... .... o. .. .... I 5.95. 0.055555 2.0.8. 5.052 00.0 8:0. 00:00... ...... 28... 6.. N... a. .0 u 8.95. 5525.35 3.52. 5.052 00.0 ..00 000003 xv... .ooodv no.0 mm.m N . .m -0... I 60.0806 880800.300 03.60.. .0..8..2 00.0.... .00.... .0301. 60. 8080...... 0gb->0.0w08.0v. 00.3-.. 9.8... 00:03.00“. $3 0.086%. 030...; 0800.00 0... 00 02.0.0.6... 0.00.... 0.08.3. 0.0.. 0......» 9.6.. .0008 0060050.. .00.... 030202.38 .08..— .vd 030,—. 78 The regression estimate for the association between within herd JD prevalence and study year was negative, suggesting that the prevalence in these herds declined over time, even though that decline was not statistically significant. Regardless, there was a significant association between decreasing JD herd prevalence and number of positive environmental samples. For every one unit increase in the number of annual positive environmental samples, the within herd JD prevalence increased 1.62%. (p= <0.0001). When contamination within specific areas of the farm were assessed, for every one unit increase in the number of positive environmental samples in the pre-weaned calf, weaned calf, and maternity areas, within herd JD prevalence increased by 5.12%, 6.19%, and 5.68% respectively. Environmental contamination in the lactating cow and manure storage areas were not statistically associated with increasing JD prevalence because these were the areas that were consistently contaminated on the farms, even when within herd prevalence was very low. 2.5. Discussion The strength of this study is its longitudinal nature, such that changes in the distribution of environmental MAP contamination could be monitored as within herd JD prevalence changed on infected dairy farms following the implementation of on-farm JD control programs. MAP was cultured consistently (75% of the time) over time in the manure storage area and/or the lactating cow floor when within herd culture prevalence was >2%; indicating a consistent reservoir of MAP contamination, even when relatively few cows in the herd are actively shedding. However, once the number of cows shedding the bacterium in the herd fell to <2%, MAP was not cultured from any location sampled. 79 Logically, the fewer cows shedding MAP, the less contamination there is in the environment; and the less likely it is for an environmental sample to be collected containing MAP at a level detectable by currently available culture methods. It is also possible manure management and sanitation practices implemented by the herds for JD control purposes resulted in a less manure accumulation, thereby decreasing the potential for MAP environmental contamination. The fact MAP was never cultured from environmental samples of one herd that consistently had low within herd JD prevalence (<2%) does not mean the environment on this herd was not contaminated with MAP. More likely the level of MAP contamination was minimal and below the detection threshold of the sampling protocol used in this study. Herd prevalence had to increase only slightly to 5% before MAP was cultured in areas in addition to the lactating cow floor and manure storage areas, with the most common area being the maternity floor. This is not surprising, as this is an area populated with adult cows. From a JD control standpoint, it is concerning because calves, the animals most susceptible to becoming infected with MAP, are being born in those areas. It emphasizes the importance of maternity pen management in any JD control program. A surprising finding in this study was the positive environmental samples in the pre-weaned calf area on three different farms. While it was possible to explain cross contamination from a contaminated maternity pen across an alley in the same barn on one farm; the other two farms with positive calf areas had separate calf barns, located well away from adult cattle. It is possible these areas became contaminated through farm personnel or feeding/cleaning equipment traveling between cow and calf barns. The 80 other possibility is some of the calves on these farms were shedding MAP. While it has been traditionally thought newly infected cattle do not start shedding MAP for several months, or until adulthood; recent reports suggest calves may indeed shed MAP, albeit transiently, and typically, at low levels (Bolton, et al., 2005; vanRoermund and deJong, 2005). Regardless, the finding of MAP in the pre-weaned calf area should be considered a risk for infection to the calves housed there, and appropriate precautions taken. Isolating MAP in four of five (80%) samples of recycled sand bedding (although originating from only one farm) was also an interesting finding, and raises the issue of where that bedding should be used. If the traditional JD paradigm that cattle become less susceptible to infection with age is accepted, using this sand to bed the adult herd likely represents minimal risk for spreading the infection. However, care should be taken to ensure it is not used in calf, young heifer, or maternity pens. Our findings were similar to those reported in previous studies (Raizman, et al., 2004; Berghaus, et al., 2006; Lombard, et al., 2006), in that the areas most commonly contaminated with MAP on infected dairy farms were those where there was the greatest concentration of manure (lactating cow floor and manure storage) from adult cows, the animals at greatest risk of shedding the bacterium. Also, as in previous studies (Raizman, et al., 2004; Fyock, et al., 2005; Berghaus, et al., 2006), there was an overall tendency for the amount of MAP in the environment to increase as within herd JD prevalence increased (Tables 2.3 and 2.4, Figure 2.1). The difference between this study and those referenced, was this study was longitudinal in nature while the others were cross- sectional. The significance being that these findings were consistent over time in the face of increasing and decreasing JD prevalence within the same herds. Thus, adding strength 81 to the importance of environmental contamination as a source of MAP transmission to susceptible cattle. At the individual herd level, there was not an obvious consistent downward trend in MAP environmental contamination as within herd JD prevalence decreased in all herds (Table 2.2). Factors to consider are: the potential for one or two “super-shedders” in the herd, the relatively long time MAP survives in the environment and the diligence each herd gave to sanitation. The “super-shedder” phenomenon in regards to MAP infection is a recently introduced theory in which one infected cow sheds billions of bacterium into the environment each day (Whitlock, et al., 2006). Thus, one or two “super-shedders” in a herd could disproportionately contaminate the environment, resulting in a high environmental load of MAP when compared to the absolute number of shedders in the herd. The prolonged survivability of MAP in the environment may result in a lag period following the removal of cows actively shedding MAP during which the environmental load of the bacteria in the herd’s environment is maintained. That environmental load does not decrease to undetectable levels until the MAP finally dies, or it is physically removed. As in the general population, some of the herds in this study did a better job at cleaning and manure removal than others. Subj ectively, the herds that did not follow the expected pattern of decreasing environmental MAP contamination in conjunction with decreasing within herd JD prevalence, were the ones that were less diligent in cleaning. Regardless, as long as MAP is detectable in the environment, susceptible cattle in the herd are at risk of becoming infected. In contrast, linear regression analysis did demonstrate a statistically significant association between within herd JD prevalence and the number of positive environmental 82 samples. As the number of contaminated environmental samples increased, so did within herd prevalence. Also, the culturing of MAP from areas other than the lactating cow and manure storage areas was likewise associated with increasing herd prevalence. The reason a similar association was not found in the lactating cow and mnure storage areas is likely due to the fact that these were the areas most consistently contaminated over time, even when prevalence was relatively low. Potential factors influencing the results of this study were sample collection and culture system used. An attempt was made to collect representative environmental samples from the same areas on each farm throughout the course of the study. All the study herds were infected with MAP. It is certainly possible any given area may have been contaminated with MAP, but the sample collected either did not contain the bacteria, or did not contain enough of it to be detectable by the culture methods used Therefore, due to sampling error, the MAP contamination on these dairy herds is likely to be more extensive than reported. During the course of this study, the lab switched from solid culture on HEY media to the ESP II liquid culture system. Subsequently, the number of positive environmental cultures increased, while JD prevalence on most of the study herds was on a downward trend. The most likely explanation for this is the ESP 11 culture system is more sensitive than HEY and able to detect MAP at lower levels (Rajeev, et al., 2006). Quantitative analysis of the level of MAP contamination in environmental samples over time, in conjunction with changing within herd JD prevalence would have strengthened this study. A high volume of contamination in the environment is associated with increased infection pressure, and subsequently higher within herd 83 prevalence (Collins, 2003). In a cross-sectional study, Raizman, et al. (2004), reported a positive correlation between the volume of MAP isolated from environmental samples and herd prevalence. Unfortunately, the laboratory reporting of quantitative culture results was inconsistent over the course of this study, precluding such analysis. The lack of quantitative analysis of MAP environmental contamination on these herds over time is a limitation of this study, and something that remains to be pursued in the future. In this study, when herd fecal culture prevalence was >2%, MAP was isolated from the lactating cow floor and/or manure storage area 75% of the time. If samples cultured using only the ESP 11 culture system are considered, the lactating cow floor and/or the manure storage area was positive 81% of the time. Thus, culturing these two areas was a sensitive method for determining the presence of, and, to a lesser degree, the extent of MAP in a dairy herd. This protocol could be adapted to monitor the progress of JD control programs at the individual herd or regional level. Periodically culturing these areas on an individual farm could provide some indication whether or not the herd’s JD control program is working. As within herd JD prevalence declines, eventually the lactating cow area and manure storage should consistently culture negative for MAP. At the state, regional, or national level, culturing the lactating cow and manure storage areas could provide an economical and efficient method for determining the number of dairy herds infected with JD, which in turn, could help direct the allocation of JD control resources. In conclusion, MAP was widely distributed in the environment of the Michigan dairy farms participating in this study. The lactating cow floor and manure storage area (areas with the greatest concentration of manure from the greatest number of adult cows), 84 were the locations on the farms that most commonly cultured positive for MAP. Periodic targeted sampling of these areas may provide an efficient and economic tool for monitoring the progress of JD control programs at the individual herd and regional levels. An increasing number of MAP positive environmental cultures was associated with increasing within herd JD prevalence. It was not uncommon for MAP to be cultured from the maternity and pre-weaned calf areas, areas where there is a high risk for transmitting JD to the next generation of herd replacements. This underscores the need to emphasize cleanliness in these areas when recommending JD control programs. Finally, as long as MAP is present and detectable in the herd environment, susceptible cattle are at risk of becoming infected with JD. Thus, when using testing as part of a JD control program, targeted testing of the environment may be as important as testing individual cattle. 85 CHAPTER 3 LON GITUDINAL STUDY TO EVALUATE THE EFFECTIVENESS OF MANAGEMENT PRACTICES IMPLEMENTED TO CONTROL JOHN E’S DISEASE ON INFECTED DAIRY FARMS IN MICHIGAN 86 3.1. Abstract A five—year longitudinal study was conducted on seven Michigan dairy herds infected with Johne’s disease (JD) to evaluate the effectiveness of management practices implemented to control the disease. The JD incidence and prevalence was monitored in each herd annually by serum ELISA and/or fecal culture of all adult cows. A JD control program was designed specifically for each herd based on the results of an initial risk assessment. The risk assessment was repeated annually and the control program updated as needed. Herd risk assessment scores were used as a measure of which control practices were implemented and to what extent. The risk assessment scores were extrapolated to each cow consistent with when she was present in each assessed area as determined by her birth date. To assess the overall effectiveness of the control programs in preventing new infections, the JD incidence rate by lactation was compared between cows born after implementation of the JD control program to cows born prior to the control program. Over the first three lactations, the incidence of JD was consistently lower in cows exposed to the JD control program as calves than in cows not exposed to the control program as calves; providing evidence that, overall, the control programs implemented on these herds were successful in preventing new infections. The effectiveness of specific management practices in preventing JD infection was evaluated using logistic regression; controlling for clustering of cows within herds with generalized estimating equations (GEE). Univariable and multivariable models were built. The final multivariable model consisted of the following two variables: exposure to adult cows other than dam at birth (OR = 1.09, 95% CI: 1.06 — 1.13), and feeding colostrum from one cow to multiple calves (OR = 1.10, 95% CI: 1.09 — 1.12). Thus, for every one point 87 increase in the risk assessment scores in each of these areas, the odds of a cow testing positive for JD was increased by 9% and 10% respectively. Conversely, lowering the risk assessment score will decrease the odds of a cow testing positive for JD. When designing JD control programs, implementing management practices that minimize the exposure of newborn calves to Mycobacterium avium paratuberculosis being shed by infected adult cows should take priority. 3.2. Introduction J ohne’s disease (JD) is a chronic disease of cattle and other ruminants caused by Mycobacterium avium paratuberculosis (MAP). It is prevalent worldwide and is becoming increasingly so in the US dairy industry. In the most recent National Animal Health Monitoring Survey (NAHMS) in 2007, it was estimated that 68. 1% of US dairy herds were infected with MAP (USDA, 2008). This is up fi'om 21.6% reported in the NAHMS Dairy 1996 survey (USDA, 1997). The economic costs are likewise substantial. Based on data from the 1996 NAHMS study, it was estimated that JD cost the US dairy industry, on average, $22-27 per cow, or $200-250 million annually (Ott, et a] 1999); due primarily to reduced milk production, premature culling and reduced cull value (Wells and Wagner, 2000). While cost estimates from the 2007 study have yet to be released, with an increase in the number of herds infected with MAP, it is likely the economic losses due to JD will also have increased. Due to its significant impact on herd productivity, along with the potential public health consequences should MAP be linked to Crohn’s disease in humans, understanding JD and how best to manage and control its 88 spread within and between herds is a priority for the US livestock industry (Linnabary, et al, 2001). Infection with MAP results in a slowly progressive granulamatous enteritis, causing the walls of the intestine to become thickened; which in turn, impairs the absorption of nutrients from the gastro-intestinal tract. Clinically, what is observed as the disease progresses is decreased production; weight loss despite a good appetite; intermittent, progressing to chronic diarrhea; and ultimately death if the cow is not culled prior (Whitlock and Buergelt, 1996). There is no approved or practical treatment for JD in production animals (St. Jean, 1996; Wells and Wagner, 2000). Cows generally become infected with MAP as young calves, through ingestion of bacteria from a contaminated environment or from contaminated colostrum or milk. The susceptibility of becoming infected seems to decrease as the animal ages. Typically it takes 2-5 years before infected cows develop clinical signs of the disease. However, as the disease progresses, the risk of shedding MAP, thereby becoming infectious (even in the absence of clinical signs), increases (Sweeney, 1996). Because of the prolonged pre-patent period, testing and culling of MAP infected animals is not very effective in eliminating the disease (Groenendaal, et al., 2002; Collins, 2003; Dorshorst, et al., 2006; McKenna, et al., 2006; Kudahl, et al., 2007). Instead, control of JD focuses on implementing management practices that minimize the transmission of MAP to susceptible animals (Thoen and Moore, 1989; Collins, 2003; Hoe and Ruegg, 2006; McKenna, et al., 2006). While irrrplementing management changes to minimize JD transmission may sound simple, its actual practice is more problematic. The list of recommended practices to prevent MAP infection is lengthy and complex (Rossiter and Burhans, 1996; 89 Benedictus and Kalis, 2003). When presented in generic form, JD control recommendations often fail because they are not designed to meet the unique needs and capabilities of each individual farm (Rossiter and Burhans, 1996; Collins, 2003). Moreover, they are often presented to producers in their entirety, and not prioritized in any way to allow for selective or progressive adoption (Ridge, et al 2005). This can overwhelm producers to the point they believe they cannot successfully implement all the recommendations to control JD, so they decide not to implement any. Part of the reason there is no ranking, or prioritizing, of management practices to control JD is because the current recommendations are all based on hypotheses of what is known about MAP infection and pathogenesis. Research to confirm recommended management practices actually work, or to identify which practices are more effective than others, has been limited This is due mainly to the prolonged course of the disease, and the difficulty in identifying infected cattle in the early stages of the disease, which makes such studies costly and time consuming (Groenendaal and Galligan 2003). For example, to evaluate the effectiveness of a management practice to prevent MAP infection in newborn calves, the study would have to last a minimum of 3-5 years to allow adequate time for the calf to mature; and, if infected, give the disease time to progress to the point it can be detected. Also necessary, would be extensive, and repeated, individual animal testing to determine if the JD incidence decreased following the implementation of such a control practice. Yet, being able to present producers with a concise and specific plan for JD control, that is not overwhelming in its breadth, is necessary for the widespread adoption of JD control programs on farms. 9O To date, management practices for the control of JD on dairy farms have been mainly evaluated using computer simulation models. Although the scenarios and assumptions varied across the studies, each concluded that improving and maintaining strict calf hygiene, thereby breaking the primary MAP infection route, was the most critical and cost effective component of a JD control program (Groenendaal, et al., 2002; Groenendaal and Galligan, 2003; Dorshorst, et al., 2006; Kudahl, et al., 2007). Simulation studies are commonly performed, particularly in the case of a chronic disease such as JD, because they are less expensive and require less time than field studies. The major disadvantage of a simulation study is that it can be difficult to validate. The input data on which the simulation is run is often based on field data supplemented with expert opinion. Therefore, a simulation study cannot be isolated from the field, and, in fact, can only truly be validated with observations made under real farm conditions with a field. study. The primary objective of this five-year longitudinal study was to evaluate the effectiveness of specific management practices implemented to control JD on seven MAP infected dairy herds in Michigan. Specifically, the study was designed to determine which JD control practices were most effective in reducing the incidence of JD. This information could then be used by veterinarians and producers in prioritizing or selecting necessary practices to implement when designing JD control programs for their own operations. Secondary objectives of the study included: comparing the level of agreement between the serum ELISA and fecal cultures (HEY and ESP II) used to monitor within 91 herd JD prevalence, and assessing the effect of the dam’s JD test status on her offspring’s JD test status. Critical to any disease control program is the early detection of infected animals so they can either be removed from the herd, or managed in a way to mitigate disease transmission to susceptible herdmates. Due to the prolonged incubation period of JD, identifying infected cattle in the early stages of disease is difficult with cru'rently available diagnostic tests (Collins, 1996; Stabel, et al., 2002; Motiwala, et al., 2005; Dieguez, et al., 2008). Culturing MAP from fecal or tissue samples remains the “gold standard” for definitively diagnosing JD (Collins, 1996; Stabel, et al., 2002). . However, ELISA tests to detect antibodies to MAP has become widely accepted for use in routine JD testing because it is inexpensive ($6/ sample for ELISA compared to $23/sample for culture, Diagnostic Center for Population and Animal Health, Michigan State University, 2008), and results are generally available within days as opposed to weeks for culture. Multiple studies have assessed the performance of fecal culture and the various commercially available serum ELISA tests for MAP. Findings vary depending on study population, diagnostic test used, and laboratory performing the test, but the general consensus of JD researchers and experts is as follows. Fecal culture detects infected cows earlier in the course of the disease (Sweeney, et al., 2006). Newer, broth-based culture systems are more sensitive and likely detect cows in earlier stages of disease than standard culture on Herrold’s egg yolk (HEY) solid media (Kim, et al., 2004; Motiwala, et al., 2005). The ELISA reliably identifies cattle that are shedding large numbers of MAP into the environment Whitlock, et al, 2000; Stabel, et al., 2002; Sweeney, et al., 2006). However, subclinically infected cows cannot be differentiated from clinically 92 infected cows based solely on the magnitude of the ELISA test result (Spangler, et al., 1992). There can be substantial variation between serial ELISA tests on the same cattle (Hirst, et al., 2002; Barrington, et al., 2003; van Schaik, et al., 2003; Sweeney, et al., 2006). The sensitivity of the ELISA test decreases with repeated herd testing as the number of heavy shedders in the more advanced stages of the disease are culled (Whitlock, et al., 2000; Stabel, et al., 2002; Sweeney, et al., 2006). Recently, only slight agreement (kappa = 0.14 i 0.07) was reported between the results of fecal culture on HEY and serum ELISA (Pinedo, et al., 2008). Given this background, it seemed prudent to compare the different JD diagnostic tests used in this study. While external, or environmental factors, are often the primary focus of JD control programs, vertical transmission should also be considered. Multiple studies have demonstrated that in-utero transmission of MAP does occur (Doyle, 1958; Lawrence, 1956; Seitz, et al., 1989; Sweeney, et al., 1992; Manning, et al., 2003). Also important in the transmission of JD is the fact that cows are more likely to shed MAP in their feces, colostrum, and milk around parturition (Harris and Barletta, 2001), the time when there is the closest, and to some extent, unavoidable, contact between calves (the animals most susceptible to infection), and infected cows. On one farm, cows with dams that were MAP ELISA positive were 6.6 times more likely to be ELISA positive themselves (Aly and Thurmond, 2005). Whether it is due to in-utero transmission, or occurs through direct contact between infected dams and their offspring, understanding the risk of calves with MAP infected dams becoming infected is important for deciding how to manage infected cows, as well as their calves, if they are to remain in the herd. 93 3.3. Materials and methods Herds Seven Michigan dairy herds were selectively enrolled in this study January 2003 through May 2004, and followed through December 2007. To qualify for this study, each herd had to be infected with JD, be representative of the MiChigan dairy industry in terms of herd size and housing management, and the herd owner willing participate in the study for at least five years and implement a JD control program. Within herd JD prevalence and incidence Data to calculate and monitor the within herd JD prevalence and lactational incidence rate was obtained by annual whole herd testing of all adult cows. Blood samples were collected, and serum analyzed for MAP antibodies using a commercially available ELISA test kit (Parachek®, Prionics AG, Schlieren-Zurich, Switzerland). Fecal samples were also collected and cultured for MAP. The laboratory switched MAP culture systems midway through the study. Through June 2004, cultures were performed on standard solid culture using Herrold’s Egg Yolk (HEY) media. After June 2004, all cultures were performed using an automated liquid culture system (ESP II, ESP® Culture System 11, TREK Diagnostic Systems, Cleveland, OH). Herd risk assessment A herd risk assessment (RA) was performed annually on each herd. The risk assessment tool used was that approved by the USDA for use in the National Voluntary 94 Johne’s Disease Control Program. It consists of a subjective score given to each of 32 “risk factors” divided into six different areas of the farm: maternity, pre-weaned calf, weaned calf, bred heifer, lactating cow, and additions/ replacements. For the purpose of the RA, it is accepted that some risk factors are more important in the transmission of JD than others. For example, exposure to manure as a newborn calf is considered to be more important than exposure to manure as an adult cow. To account for this, the maximum score possible for factors in the maternity area is higher than that in the lactating cow area. In other words, a score of four in the maternity area is not the same as a score of four in the lactating cow area. Individual cow data Individual cow data was collected for every cow tested on each annual whole herd test and included, when possible: identification, date of birth, lactation number, dam identification, JD test status, and. whether or not she was purchased or raised on the farm. Implementation and monitoring of JD Control Program The initial visit to each herd consisted of whole herd testing (as described above) to determine baseline prevalence. Also during this visit, a JD risk assessment was performed; and a JD control program developed based on the risk assessment, farm goals, and capabilities. This RA was repeated annually, and the control program updated as needed Variation of scores within and between herds was minimized by having the same person (DLG) perform the RA throughout the study. 95 Although the RA is subjective, it was the best tool we had available to monitor what JD control practices each herd put in place, and the extent to which they were following those practices over time. While the RA scores were assigned at the herd level, the scores were extrapolated to the individual cow level. Each cow was assigned the RA score for each area consistent with when she would have been present in that area based on her birthdate. The initial RA was used as baseline. For analysis purposes, the JD control program was considered to have been implemented on the date of the initial herd visit. Cows were classified as having been “exposed” to the control program if they were born after the initial visit, with the exception of purchased cattle. All other cattle were classified as “unexposed.” The vast majority of purchased cattle were introduced into these herds as springing heifers or adults, and information on the conditions under which they were raised was unavailable. As a result they were not assigned any RA scores for those areas and were classified as being “unexposed” to the JD control program They were dropped from all analyses, except for that comparing JD in purchased cows vs. cows raised on the farm. Statistical analysis A. Comparing JD serum ELISA to fecal culture Agreement between the serum ELISA test and fecal culture was compared using the Kappa statistic. Fecal culture has commonly been used as the “gold standard” for diagnosing JD (Milner, et al., 1990; Collins, et al., 1991; Cox, et al., 1991; Socket, et al., 1992; Collins, et al., 1993; Sweeney, et al., 1995; Dargatz, et al., 2001). The Kappa was 96 calculated comparing the ELISA to both HEY and ESP 11 culture systems as well as culture status regardless of system Kappas were also calculated for each herd and across all herds. B. Calculating JD prevalence. The annual JD prevalence within each herd was calculated several ways. ELISA prevalence was calculated as the number of cows testing “positive” (defined as an adjusted OD reading 2 0.1 per manufacturer’s recommendation) that year divided by the total number of cows tested. The apparent fecal culture prevalence was calculated as the number of cows culturing positive divided by the total number of cows tested. Because the lab switched culture systems midway through the study, and the ESP II system is reported (and appears) to be more sensitive than the HEY system, an adjustment for these differing and imperfect test sensitivities and specificities was made by calculating the “true” fecal culture prevalence using the following equation (Smith, 1995, pp.82): Apparent prevalence + Sp - 100% Se + Sp — 100% TRUE PREVALENCE = The Sensitivity (Se) and Specificity (Sp) used for each culture system is as follows: Culture System Se Sp Reference HEY 50% 99% Sockett, et al., 1992 ESP II 65% 99% Kim, et al., 2004 97 Prevalence was also calculated based on JD test status. Cows testing positive, (either ELISA positive and/ or PC positive) were divided by the total tested This was performed because the agreement between the two tests ranged from poor to moderate depending on herd and culture system. Because JD is a chronic disease with a long and varied incubation period averaging 2-5 years, the prevalence at the herd level may not be an accurate reflection of the effectiveness of the JD control program. The first lactation cows may provide the best indicators of whether or not the control program is working. Thus the ELISA, fecal culture (F C), and JD test prevalence for first lactation cows was calculated by taking the number of first lactation cows testing positive by each respective test and dividing by the total number of first lactation cows tested that year. The Cochran-Armitage test for trend was calculated for each binary outcome (ELISA, F C, and JD test status) across the years of the study at both the herd level and for first lactation cows only. A test statistic greater than zero suggested prevalence increased over the study period, while a test statistic less than zero suggested decreasing prevalence over time. A p-value of <0.05 was considered statistically significant. C. E flectiveness of JD control program Before specific management changes for preventing JD could be evaluated, it first had to be determined if the JD control programs implemented on these farms were effective in reducing the JD burden in the respective herds. The effectiveness of the overall JD control program for each herd was evaluated in two ways: by calculating the relative risk (RR) of cows “exposed” to the control program to those “unexposed,” and 98 by calculating the JD incidence rate by lactation for cows “exposed to the control program as compared to those “unexposed.” Due to the longitudinal nature of the study, many cows were tested for JD multiple times. Johne’s disease is a chronic, slowly progressive disease; and while the diagnostic tests used in this study may not have the best sensitivity (SO-60%) when applied to random cows within the general population, they do have excellent specificity (>99%) (Collins, et al., 2006). Thus, it was assumed that once a cow tested positive (regardless of test), she was infected with MAP, and remained so for life. As a result, only a single observation to define JD status was kept for each cow. The observation retained was either the first time the cow tested positive for JD; or, for cows always testing negative, the last test performed. To compare the prevalence of MAP infected cows born prior to the JD control program being implemented to that of cows born and raised with the control program in place, a 2 x 2 table was constructed and the RR and Pearson chi-square calculated for each herd. Another way to consider the effectiveness of the JD control program is to compare the age at which cows “unexposed” to the control program test positive for JD to that of cows “exposed” to the control program To do this, the incidence of JD for each group (unexposed and exposed) by lactation was calculated. For lactation 1, the incidence was calculated as the proportion of cows testing positive in their first lactation divided by the total number of first lactation cows tested. The incidence for subsequent lactations was calculated as the number of cows in that lactation testing positive (provided they had not tested positive in a previous lactation) divided by the total number 99 of cows in that lactation at risk of testing positive for the first time. The fisher’s exact test was used to test for statistically significant differences in the incidence between cows exposed and unexposed to the JD control program. For both the RR and incidence rate calculations, a p-value of <0.05 was considered statistically significant. D. Determining which management practices are effective in the JD control program Individual cow data, which included the assigned risk scores, was used to determine which management practices significantly affected JD status. The risk of acquiring JD decreases with age, becoming minimal by adulthood. Testing to detect JD did not begin until adulthood. Thus, for practical purposes, the risk of acquiring JD remained unchanged for each cow throughout the observation period. A single observation to define JD status was retained for each cow as described previously. Due to the nature of the study, correlation resulting from clustering of cows within herds was controlled using generalized estimating equations (GEE) with an exchangeable correlation structure. The outcome of interest was JD test status (positive/negative). As JD test status was binary in nature, logistic regression was used to model the probability that the outcome was positive. The distribution of risk scores for each factor evaluated on the RA was analyzed. Factors with a distribution range of 5 3 points were dropped from fiirther analysis. This was deemed appropriate because the scores were subjective in nature, making it difficult to argue that a discrepancy of one point in either direction is biologically significant. Furthermore, there must be sufficient variation in the data to justify incorporating a 100 variable into a statistical model. To maintain the weighting inherent in the RA for risk factors during multivariable analysis, each risk score was multiplied by the maximum possible score allowed in the RA. For example, the maximum possible score for manure build up in the calving area is ten (Appendix A). So a cow born into a herd when the manure build-up in the maternity pen was rated for would be assigned a risk score of 40 (4 x 10). Accordingly, the maximum possible score for manure contamination in the lactating cow area is four. So an adult cow present in a herd when the manure contamination in the lactating cow area was four, would be assigned a risk score of 16 (4 x 4). This demonstrates that a score of four in the maternity area is not the same as a score of four in the lactating cow area. Formal interaction terms were not considered in the analysis. However, upon considering the list of risk factors assessed, it seemed likely some factors were linked. For instance, manure build up in the maternity pen was likely associated with manure- soiled udders and legs. In those instances the scores were summed together to form another “risk factor.” In the maternity area, the factors fell into three categories, exposure to manure from adult cattle, direct exposure to adult cattle (other than dam), and time spent with dam. In the calf area, the combined categories consisted of colostrum management (potential for colostrum from one cow being fed to multiple calves), feeding of unpasteurized pooled milk, and adult manure contamination of feed or water supplies. Combined terms in the other areas of the farm (weaned calf, bred heifer, and lactating cow) were not necessary because multiple factors in these areas were dropped from the analysis due to lack of variation across scores. To avoid collinearity problems during multivariable modeling, the combined term was used in place of the component risk 10] factors only when each respective factor was found to be statistically significant on univariable analysis. Potential confounding factors included lactation number and age. The reason for considering both age and lactation number, which could be used as a proxy for age, will be explained in detail later. Other variables evaluated included culture system (HEY vs. ESP 11), exposure to JD control program (yes/no), and source of cows (raised vs. purchased). Univariable and multivariable models were built using JD test status as the outcome. Variables with a p-value 5 0.10 on univariable analysis were considered in the multivariable model. Stepwise backward elimination was used to build the final model. The final multivariable model included only those variables with a p-value of 5 0.05. Model fit was assessed in two ways: a modification of the Hosmer-Lemeshow goodness of fit analysis as described in Horton, et a1 (1999), and an extension of model checking using cumulative residuals for marginal regression models as discussed in Lin, et al. (2002). Briefly, for the Hosmer-Lemeshow analysis, the data set was divided into groups of approximately equal size based on ordinal ordering of the expected probabilities of observations testing positive for JD. Dummy variables were assigned to each observation, defining into which group it belonged. The regression equation was rerun including the dummy variables. The null hypothesis for the model being, if the model fits the data well, the regression coefficients for all dummy variables will equal zero (Wald 12 p-value >005). For the residual analysis, the cumulative sums of the residuals for the each respective covariate in the marginal regression model were plotted along with the residuals of 10,000 simulated realizations from a zero-mean Gaussian 102 distribution. The Kolmogorov-type supremum test was calculated along with its associated p-value. This process was repeated with alternative functional forms of the covariate based on the initial pattern of the cumulative sums of residuals in an attempt to improve model fit when warranted. The greater the Kolmogorov-type test statisitic and its p-value, the better the model fits the data, and p-values <0.05 were considered indicative of poor, or insufficient, model fit Both the Hosmer-Lemeshow analysis and the cumulative residuals were assessed for the final multivariable regession model as well as for the respective univariable regression models for each covariate included in the final multivariable regression model. E. Effect of dam ’s JD test status on oflspring ’s JD status Another potential risk factor for a calf becoming infected with JD is its dam’s JD status. The dam’s identity for each cow tested was recorded when available. Over the five-year course of study, dam test information was available on a little over one third of the cows. It also meant there was no dam JD test information on 2/3 of the cows. Thus controlling for dam’s JD status in the above described regression analysis would have resulted in only a small proportion of the data being used. Therefore, the association between dam’s JD status and that of her daughter was analyzed separately using a 2 x 2 table and calculating the RR and Pearson Chi-square. All cows (dams and daughters) were classified as positive if they were tested at least once during the course of the study, and had a positive ELISA &/or positive F C. 3.4. Results 103 Descriptive data analysis A. Herds In six of the seven herds, both serum ELISA and fecal culture was performed annually. In the remaining herd only serum ELISA testing was performed. Six of the seven herds consisted of Holsteins, or predominately Holstein cows and one herd consisted of Jersey cows. Type of management varied from total confinement (N=3), combination of confinement and grazing (N=3), and rotational grazing with winter confinement (N=l). Two herds (herds 3 and 5) were actively expanding during the study period and were routinely purchasing cattle. Another herd (herd 1) was expanding internally, although occasionally cattle were purchased to improve genetics. The herd size of the other four herds was relatively consistent throughout the study, although three of the four had purchased cows within the five years prior to the start of the study. One herd (herd 6) had been closed for almost 30 years prior to the start of the study. These descriptive statistics are summarized in Table 3.1. Table 3.1: Herd size, breed and housing management of study herds Herd Ave. herd Herd size Breed Housing size Start End 1 1 91 l 70 2 1 5 Holstein Confinement/ grazing 2 125 103 137 Holstein Total confinement 3 378 2 1 8 458 Holstein Confinement/ grazing 4 73 75 68 Jersey Rotational grazing (organic) 5 53 l 484 641 Holstein Total confinement 6 155 145 167 Holstein Total confinement 7 184 209 168 Holstein Confinement/ grazing 104 B. Cows Over the course of this study, a total of 8,660 observations were made on 4,123 cows. There were 6,447 observations with concurrent ELISA and fecal culture (F C) results of which 227 (4%) were positive on both tests, 493 (8%) ELISA positive, and 586 (9%) FC positive. There were a total of 6,530 observations with PC results, of which 586 (9%) were positive; and 8,578 total observations with ELISA results, of which 493 (6%) were positive. A total of 787 (9%) observations originated from purchased cows, of which 178 (23%) had concurrent ELISA and FC results while 785 had ELISA results and 180 PC results. Forty-eight (6%) of those observations were positive on ELISA and 46 (26%) had positive FC. Of the 7,873 (91%) observations originating from cows born and raised on their respective farms, 445 (6%) were positive on ELISA and 540 (7%) FC positive. Eighty-two percent (7,137) ofthe observations fi'om raised cattle came from cows that were born prior to the implementation of the JD control program and 18% (1,523) from cows born and raised with the control program in place. Of the observations made from cattle born prior to the JD control program 421 out of 6,272 observations were ELISA positive and 472 out of 5,073 observations were FC positive. Of the observations from cattle born and raised after the JD control program was implemented, 24 out of 1,520 observations were ELISA positive and 68 out of 1,277 observations were FC positive. Of the 4,123 cows tested during the course of this study, 416 were purchased and 3,707 were born and raised on their respective farms. Fecal cultures were performed on a total of 2,999 cows of which 460 (16%) were FC positive at least once. Seventy-seven (17%) cows were F C positive two or more times. Serum ELISA tests were performed on 105 a total of 4,086 cows of which 359 (9%) were positive at least once. Sixty-eight (19%) cows were ELISA positive multiple times. Of the 4,123 cows tested 679 (16%) were ELISA and/or FC positive. Of the 416 purchased cattle, ELISA tests were performed on 414, with 40 (10%) having at least one positive ELISA result. Fecal cultures were performed on 96 of the purchased cows with 36 (3 8%) being positive. Overall 63 (15%) of the purchased cows were ELISA and/or FC positive. Seventy percent (2,610) of the 3,707 cows raised on the farm were born prior to the implementation of the JD control program, while the remaining 30% (1,097) were born and raised with the control program in place. Three hundred fifty-four out of 2,032 ( 17%) cows were FC positive and raised prior to the JD control program, while 68 cows out of 871 (8%) were FC positive and raised with the control program in place. Two hundred ninety-five cows out of 2,578 (11%) were ELISA positive and born prior to the implementation of the JD control program; while 24 out of 1,094 (2%) cows were ELISA positive and born with the control program in place. Overall, 535 out of 2,610 (20%) cows born prior to the JD control program were ELISA and/or FC positive, while 81 out of 1,097 (7%) cows born after the implementation of the control program were ELISA and/or FC positive. Statistical data analysis 106 A. Comparing serum ELISA test with fecal culture. The kappa statistics comparing the level of agreement beyond chance between the JD serum ELISA test and F C for each herd and across all herds are summarized in Table 3.2. Note kappa is only calculated for six herds because one herd (herd 5) was monitored by annual whole herd serum ELISA only. Table 3.2: Kappa statistic for comparing agreement between Johne’s disease serum ELISA test and fecal culture. Herd Kappa for HEY Kappa for ESP 11 Overall Kappa 1 0.1 1 0.46 0.33 2 0.43 0.28 0.37 3 0.58 0.38 0.43 4 0.58 0.24 0.44 5 N/A N/A N/A 6 0.16 0.46 0.41 7 0.06 0.15 0.07 All herds 0.33 0.39 0.37 Kappa level of agreement: SlighFO-Oz, Fair=0.2-0.4, Moderate=0.4-0.6 (Smith, 1995, pp. 149) N/A: Annual herd testing in herd 5 consisted of serum ELISA only B. Within herd JD prevalence over study period The within herd prevalence using fecal culture, serum ELISA, and JD test status as the outcomes of interest respectively are shown in the Figures 3.1-3.6 along with the accompanying Cochran-Amitage test for trend (Tables 3.3-3.8). 107 Figure 3.1: Trend for “true” fecal culture within herd prevalence - all cows \1 O PREVALENCE (%) 8 8 8 8 -o—Herd1 A\ +Herd2 ,' t - a -Herd3 f h -e -Herd4 , “ -0 -Herd6 IT \ —)It—Herd7 : ‘A‘ I /\\~ _ . 4 Table 3.3: Cochran-Armitage test for trend for “true” fecal culture prevalence — all cows Herd Statistic p-value 1 -0.5422 0.2912 2 -3.4548 0.0003 3 4.4486 <0.0001 4 -0.7 848 0.2163 5 N/A N/A 6 -3.3 703 0.0004 7 1.5504 0.0605 N/A Not applicable Whole herd fecal culture not performed on Herd 5 108 Figure 3.2: Trend for “true” fecal culture within herd prevalence - first lactation cowsonly 45 -o—Herd1 40 +Herd2 35 1‘ 'l -Herd3 A / ‘. A -o-Herd4 E530 4. \ -O-Herd6 g ’ , \ +Herd7 z 25 [l \ a . \ < 20 . ‘ E... I ~ \ :h a" K / Li ' -\- . ‘7‘ 1o ' ‘. x - - _ 5 V ' o ' ...-v ’ “k ‘ ‘ ‘4 ' z; I \ ‘9 I s 0 1 2 3 4 5 6 STUDYYEAR Table 3.4: Cochran-Armitage test for trend for “true” fecal culture prevalence - first lactation cows only Herd Statistic p-value 1 0.4751 0.3174 2 -3.6514 0.0001 3 1.9171 0.0276 4 0.0252 0.4899 5 N/A N/A 6 -2.2172 0.0133 7 2.3979 0.0082 N/A Not applicable Whole herd fecal culture not performed on Herd 5 109 Figure 3.3: Trend for ELISA prevalence - all cows 20 18 A; -o—Hem|1 , ‘ +Herd2 16 i, ‘ 'h'HerdS l \ d'HH’d4 §14 = ‘ -E1-Herd5 :12 J “ -e-Herd6 g \ +Herd7 $10 E 8 a. 6— 4 2 O Table 3.5: Cochran-Armitage test for trend for ELISA prevalence - all cows Herd Statistic p-value 1 0.2012 0.4203 2 -0.4743 0.3 176 3 -1.6445 0.0500 4 -2.3214 0.0101 5 -1.0306 0.1514 6 -3.8153 <0.0001 7 -4.7067 <0.0001 110 Figure 3.4: Trend for ELISA prevalence — first lactation cows only PREVALENCE (%) owe-moo 18 16 14 12 10 +Herd1 +Herd2 It 'Herd3 -o-Herd4 -B-Herd5 _, -.'Hetd6 Table 3.6: Cochran-Armitage test for trend for ELISA prevalence — first lactation cows only Herd Statistic p-value 1 -0.3967 0.3458 2 -2.3082 0.0105 3 -l .3991 0.0809 4 1.0933 0.1371 5 -1.3068 0.0956 6 -3.2688 0.0005 7 -3.9409 <0.0001 lll Figure 3.5: Trend for prevalence based on J ohne’s disease test status = positive all cows 000) DUI PREVALENCE (%) N or 20 1 5 1 0 5 0 1 2 3 4 5 6 STUDY YEAR Table 3.7 : Cochran-Armitage test for trend for prevalence based on Johne’s disease test status = positive — all cows Herd Statistic p-value 1 -1.0365 0.1500 2 ~2.5123 0.0060 3 2.9725 0.0015 4 -1.2732 0.1015 5 -1.0306 0.1514 6 -4.8501 <0.0001 7 -2.9047 0.0018 112 Figure 3.6: Trend for prevalence base on Johne’s disease test status = positive first lactation cows only 3O \ -o—Herd1 .' ‘ +Herd2 f -t-Herd3 25 f \ —o-Hend4 i— A ’ A -E-Herd5 £20 ' “ -.'Hetd6 __ ... I /{ \ 3...”... o I E 15 ' ‘ A \ , ‘ AZ \ \ w k- ---. Z a 10 'Y ’I %~ - i n. A , 5 kl’, ik/ 0 Table 3.8: Cochran-Armitage test for trend for prevalence based on Johne’s disease test status = positive - first lactation cows only Herd Statistic p—value 1 0.5612 0.2873 2 ~3.6809 0.0001 3 1.3797 0.0838 4 -0. 1961 0.4223 5 -1.3068 0.0956 6 -3.8195 <0.0001 7 -2.2360 0.0127 113 C. E ffectiveness of JD control program Relative risk of cows exposed to JD control program testing positive compared to cows not exposed to JD control program The relative risk for the following three outcomes: fecal culture, ELISA, and JD test status (fecal culture and/or ELISA positive) were calculated for each herd, and across all herds, along with the 95% confidence limits, and corresponding p—value for the Pearson chi-square and are summarized in Tables 3.9-3.11. All herds had a R < 1, and. all but one (Herd 7 - fecal culture as outcome, p= 0.16) was statistically significant (p < 0.05). Table 3.9: Relative risk of exposure to Johne’s disease control program—Fecal culture as outcome 0 2 RR P323: 1 0.4144 0.2420 0.7098 0.0005 2 0.0637 0.0089 0.4580 0.0001 3 0.7274 0.5307 0.9970 0.0440 4 0.2047 0.0486 0.8625 0.0137 5 Whole herd fecal culture not performed 6 0.0881 0.0284 0.2731 <0.0001 7 0.5279 0.2106 1.3233 0.1610 All herds 0.7505 0.6720 0.8382 <0.0001 114 Table 3.10: Relative risk of exposure to Johne’s disease control program—ELISA status as outcome 2 o Herd RR iiégonfidence 1232:: Piaf/$23361 1 0.1816 0.0666 0.4948 <0.0001 2 0.3227 0.1165 0.8938 0.0187 3 0.2905 0.1403 0.6014 0.0003 4 0.2900 0.0667 1.2621 0.0741 5 0. 1485 0.0468 0.4705 0.0001 6 0.1561 0.0495 0.4923 0.0001 7 0 ELISA positive cows exposed to control program All herds 0.1936 0.1285 0.2917 <0.0001 Table 3.11: Relative risk of exposure to Johne’s disease control program—JD test status = positive as outcome RR $53.12 1 0.3591 0.2196 0.5872 <0.0001 2 0.2161 0.0890 0.5245 <0.0001 3 0.6355 0.4685 0.8619 0.0026 4 0.2373 0.0734 0.7671 0.0067 5 0.1430 0.0452 0.4526 <0.0001 6 0.0889 0.0336 0.2356 <0.0001 7 0.2761 0.1135 0.6719 0.0016 All herds 0.3602 0.2883 0.4501 <0.0001 Incidence of JD Tables 3.12-3.14 summarize the JD incidence rate by lactation over the first three lactations using FC, ELISA, and JD test status as outcomes respectively. The incidence rate was not calculated beyond lactation 3, because at the end of the study, there were no 115 cows “exposed” to the JD control program exceeding their third lactation. Regarding calculations for F C incidence, all samples from cows in the “exposed” group were cultured using the ESP II system. In the “not exposed” group, lactation l, 52% (n=42) of positive samples were cultured using HEY and 48% (n=39) by the ESP II system; lactation 2, 11% (n=15) of positive samples were cultured using HEY and 89% (n=104) cultured using the ESP II system; and in lactation 3, all positive samples were cultured using the ESP II system. Table 3.12: Johne’s disease incidence rate over first three lactations for cows not exposed to the control program compared to cows exposed to the control program — using fecal culture as outcome Lactation Cows I 'd 1 In 'd 2 In . d 3 nor ence c1 ence c1 ence (%) N (%) N 0%) N N01 ExpOsed ‘0 JD 5.5 1485 9.8 1084 6.0 726 control program Exposed to JD control 4.9 860 12.5 345 0 42 program Fisher 5 exact test 057 0.16 0.16 p-value 116 Table 3.13: Johne’s disease incidence rate over first three lactations for cows not exposed to the control program compared to cows exposed to the control program - using E_L1_S_A as outcome Lactation Cows 1d 1 1nd 2 1nd 3 nor ence c1 ence c1 ence 1%) N 1%) N 00 N N01 EXP°Sed 1° JD 3.7 1822 3.9 1378 9.0 863 control program Exposed to JD control 1.3 1038 1.9 366 0 42 program Frsher s Exact Test 0.0008 0.08 0.04 p-value Table 3.14: Johne’s disease incidence rate over first three lactations for cows not exposed to the control program compared to cows exposed to the control program - using Johne’s Q’sease test status = positive as outcome Lactation Cows In 'd 1 In . d 2 In 'd 3 c1 ence c1 ence c1 ence (%) N 1%) N 0%) N NO‘ Exposed ‘0 JD 6.6 1822 10.6 1412 8.7 832 control program Exposed to JD control 4.7 1041 77 351 0 42 program F rsher s Exact Test 005 0.11 MM p-value 117 D. Determining which management practices were most effective in JD control program The results of the univariable regression analyses to determine the effect of various risk factors on JD status are provided in Table 3.15. The risk factors analyzed were those assessed in the RA. Due to lack of variation in scores (<3 points) throughout the study period the following factors were dropped from the analysis: All adult cow risk factors Bred heifer risk factors 0 Direct cow contact or pen contamination with cows’ manure 0 Possible manure contamination of feed: refused cow ration, stored feed, equipment, cows, traffrc splatter, people or runoff 0 Share pasture with cows 0 Manure spread on forage grazed/harvested same season Post-weaned heifer risk factors 0 Share pasture with cows 0 Manure spread on forage grazed/harvested same season Pre-weaned calf risk factors 0 Possible manure contamination of feed or water by cows, traffic splatter, equipment or people 0 Direct cow contact or potential manure contamination of pen: by cows, traffic splatter, equipment or people Sources of additions & replacements 118 2.8... 3.... no. 3.. .39 00.0 3.50.02 .25... v n... 3.... m . . .. Am... 0.....500. E0.0 5.3 32.0 0E... So... .m. m... m . .. 6.3 0:388. 038 :33 9 0.30096 08o... cm. X... . ... Am 0% m 0:388. 0.5.5.: 8 00:02.01”. 32.... mm. 0.... .m. Est 005: 00200 .w .oood v .v. v. .. um. .03 .‘NA: .. ..... .=.E omv. Ewe 5.3 >80 00300 0E... .5 Sec... v .m. cw... 2.... 000.0 300 .050 =. EB 00200 .3 Eco... v 3... .... mm. 03.0.0.6: 00.30 0.5.8.). .m 50:. 8.. 40.0 8.. 88032833 02.. 0882.. ... wwmod m... .... o...— A0>033 ......... 003:. 38 0.0.0 8.. .000: 00.0 005. ..m 28.0 a... 02. a _ ._ .63 are... . a... .583. 80805.8 a. 0.0: a. 2.3 2:52 .0 .25... m... N... R. 90.. 00.0320 00:0.0 ........... =2. 0.w=.m. 00: .0830 03.....2 .. 0.5.00..— 0.0.~. 00.3. 5.50.03. .0an .039. 020.»-.. MO 030.53 .0 $3 000003 0.053.. 8.. 03:02. @300. 0300 ....B 33.00000 0.580.. 0.0.. .8 0.0305. 5.00030. 0..0.wo. 030.3323. um . .m 030... 119 wowm... 0... .0... 3.. .08. 00.0 .0...00 000003.000. ammo... om. m0... m0... . 0.0000 .0 0000:. 0000.00 0E0... .0300 an. .0 0...» 00.000 00.03 .0000: .0 00:00.00 00 08.00.80.000 .0. .0..00.00 .m 0000.0 . 0.. 00... mm. . . ....000. .0 0.080 0000.00 0.0.0.. .0300 00080.08 000.. 00.0.0 02.0. 300 000000. “000.. ..0 02.00.80.000 0.0008 03.0000 .N 2.00... mm. P... .0. 0.0008 0.300 0..? 02.00.80.000 000 .0 .0088 300 .0000 .. 0.0.00...— ..0.¢ .020: 000003-000.— .ooo.o v 0 ... N. .. v. .. 3.. 003800. 00200 .0.0>00 0. w0.0w B8 000 80.. 82.00.00 00mm... mm. mm... 8.. . . 0.0000 .0 050.. 0.000.: .00>.0.. .0 u...8 .0 8.0.00.8 ..0 02.00.80.000 0.0008 03.0000 .0 Sec... V cm. 0.... 0 ... .58 00.000 000.000.0000: 00 ... .m .80... v 0m. 0. .. mm. 00>.00 .0.0>00 0. >60 3:03.00. 80.0 83.00.00 00 0 .0 Eco... v mm. 00.. mm. 88.00.00 00.000 000 .. 0.0.00.0 0.0.40— .0...0= 00000.50..— .0002 .038. 00.0.1. 00 030.05 .0 $3 00000.0 0.0002. .00 0>...000 w0..00. 0300 0..? 00.0.8000 0.0.00. 0.0.. ..0 0.00.000 00.00080. 0..0.m0. 03000303 A00:0..08. m . .m 030... 120 3.86 8a 35 a; 8; 3352 u 85cm 585 o; 33 as $8 = mm. H 52% 23.5 586v $6 9 _ .o and SS semen .228 Q 2 883m 83o o2 Ed 8.. E: &< 386 2: 3o 8.. 83 ESE 8:923 Z Ecuaaegv 33:8...— awwmd .2 93 2: .92 8; 5E 35 . . . . 0302. 8 toad 532% oEmb £38 3 mono o 2. N V0 0 No _ .5 :23 $0.32? “.533 3.2.5“: .5 com—Q95 mo cosmEEpficoQ how 33:305— .— ecsé in. 3%: can 5qu 53o:— o=_a>.a mo 032:; 6 $3 038% $250.. .8 02:8; mama“: 958 £3, 33888 808$ xmt no mam—£8 :BmmbB ocflwo. ointhD “€35.53 m_.m 03$. 121 Because all the adult cow risk factors were dropped due to lack of score variation, the risk profile for cows being tested did not change throughout the observation period, and only one observation was retained for each cow tested. Several cows tested positive on PC but were negative on the concurrent ELISA, and vice versa. Also, some cows would test positive on one or both tests one year, and then test negative on the same test(s) in subsequent years. Once a cow tested positive (regardless of test), she was assumed to be infected with MAP and remained so for life. Univariable regression analysis results of all confounders and non-risk factor variables are also shown in Table 3.15. Potential confounders included in the analysis were lactation number and age. Due to the chronic nature of JD, infected cows are more likely to test positive as they age and the disease progresses. Originally, the plan was to use lactation number as a proxy for age because we did not have accurate birthdates on all cows. However, upon closer examination of the data, it was discovered that a proportion of the cows had the same lactation number for two or more test dates. This suggested these cows had prolonged lactations for some reason, and meant their lactation number would not accurately reflect their age. As a result, it was decided to evaluate age as well, dropping cows with unknown birthdates (N=233) from the analysis. Univariable analysis of lactation number and age revealed both to be not statistically significant (p=0.97 and p=0.63 respectively). Modeling risk factors with and without the cows with missing birthdates resulted in slightly different regression coefficients, but both univariable and multivariable model interpretations were the same. The multivariable regression results for herd management risk factors shown in Table 3.16 are for the full dataset including 3,707 cows. 122 Table 3.16: Multivariable logistic analysis of risk factors associated with cows testing positive for Johne’s disease _ 95% CI Variable OR p-value Lower Upper Exposure to adult cows other than dam at birth Feeding colostrum from one 1.09 1.06 1.13 < 0.0001 . 1.10 1.09 1.12 <0.0001 cow to multiple calves Other variables evaluated were culture system (HEY vs. ESP II) used, exposure to the JD control program (yes/no), and source of cows (raised vs. purchased). Using ESP II as the referent, analysis of culture system resulted in an OR = 0.49 and a statistically significant p-value (0.0001) on univariable analysis. Likewise, using exposure to the JD control program as the referent on univariable analysis, resulted in an OR = 0.29 and p- value of <0.0001. Regarding the source of the cows, purchased cattle were generally bought as springing heifers or adult cows, and the JD status of the herd of origin was unknown, so risk scores for the areas evaluated could not be assigned and they were dropped from the above described analysis. A separate, but similar regression model to that of the risk factor analysis using GEE, was used to determine the effect of source of cattle (raised versus purchased, N = 4123 cows) on ID test status in the absence of all management risk factors. Purchased cows were 1.6 times (95% CI: 0.95 — 2.60; p = 0.08) more likely to test positive. 123 E. Regression model fit analysis Summary statistics for the model fit analysis are presented in Table 3.17. The p- values for all tests performed were >005, suggesting adequate fit of the respective regression models to the data. F. Effect of dam ’s JD test status on offspring ’5 JD status Dam JD test information was available for 1,486 cows. Cows born to JD test positive cows (ELISA and/or FC positive) were 1.4 times (95% CI: 1.09 - 1.86, p=0.01) more likely to test positive for JD (ELISA and/or FC positive) themselves than cows born to test negative dams. 3.5. Discussion Comparing the serum ELISA test with fecal culture The Kappa statistic was used to compare the amount of agreement beyond chance between the Johne’s serum ELISA test and FC. The Kappa does not specify which test is correct, only the level of agreement between the two tests. Because two different culture systems were used during the course of the study, the Kappa was calculated for each culture system separately as well as for all tests combined. While both the JD serum ELISA and PC are used to identify cows with JD, each test detects something different. The JD serum ELISA detects antibodies, or the immune system’s response to MAP, and possibly other closely related organisms. Fecal culture detects live MAP, the actual causative agent. As culture is the “gold standar ”, it is 124 03; 823 23:2: 3 38 mmdémd m w _ .o E .o de AX: 25 Set «SSS—co wficuon— A5 wwwAac m -8 o :2. 5% a 5% 55 35o €6.36 0 Ed mmd -34 $58 :33 8 8:8me $9 Ewan: Aowcfiv Gwen: Acme—Ev 0297a 0..—aim 3:80 8:80 2:91. 3.5% 023d 330% 838053 0&0 E=ansm 62 62 NR “:5; Mo Uwa NR V-flg Mo UNmm I>OHOWOE~OV~ abl>OH°woa—OV— .082 352 292:3 E52 055332.22 .322 03232:: ozflazzzz oiatmzca mix—SS mien—22‘ BosmoqucofimoI 22.28.. .«o 255 3:22:30 2233ch ouflwo— 053.3322: can 032.“:ng c8 mix—£8 E T602 ”:6 033. 125 assumed that PC is a more reliable test for definitively identifying MAP infected cows (Collins, 1996; Stabel, et al., 2002). However, due to the length of time it takes (6-8 weeks) to complete MAP culture and the higher cost, the use of the serum ELISA test can often be justified. Both the JD ELISA and FC suffer from a lack of sensitivity (ELISA, 30 i 5%; FC, 60 i 5%), but are generally considered to have specificities above 99% (Collins, et al., 2006). But, because each test detects something different (antibodies to MAP versus the actual MAP), it is likely each test will detect a different subpopulation of the diseased population. The amount of overlap in these two populations will vary depending on several factors including: route of infection, size of infecting dose, stage of infection, strain of MAP, and exposure to other Mycobacteria. Obviously variation in the amount of overlap in the populations identified by each test will impact the Kappa. The Kappa for comparing the serum ELISA test to FC using solid culture on HEY media ranged from 0.06 (slight agreement) to 0.58 (moderate agreement) across the individual herds. The Kappa for comparing the serum ELISA test to FC using’the ESP II liquid culture system ranged from 0.15 (slight agreement) to 0.46 (moderate agreement). When all observations with concurrent ELISA and FC results were combined, the overall Kappa ranged fi'om 0.07 (slight agreement) to 0.44 (moderate agreement) across individual herds, with the overall Kappa for all herds combined being 0.37 (fair agreement). There was no consistency in the direction of change in the value of Kappa between the two culture methods (Table 3.2), suggesting that something other than culture method was affecting Kappa. In all instances the overall Kappa fell between the Kappa values for HEY and ESP II. 126 Upon closer examination (Table 3.2, Figures 3.1 & 3.3), there appeared to be a trend between the Kappa for HEY and that for ESP 1] based on prevalence. If both ELISA and F C prevalence increased, the Kappa increased (herds 1 and 6). Conversely, if both ELISA and F C prevalence decreased, then Kappa decreased (herds 2 and 4). Also, if F C prevalence increased while ELISA prevalence decreased, the Kappa increased (herd 7, all herds). A possible explanation for this is, assuming that FC is a more reliable indicator of the true disease state, an increasing FC prevalence indicates a higher proportion of JD infected cows in the test population. This should theoretically increase the number of MAP infected cows available to be detected by ELISA. If there is any overlap at all in the populations being detected by the two tests, increasing the number of infected animals should increase the number of positive cows identified by each test, which in turn increases the probability that the tests will agree. Thus, when the true prevalence in a p0pulation increases, then the kappa between two diagnostic tests should increase if they are related at all and vice versa. In one herd (herd 3), the FC prevalence decreased at the time of the culture switch while ELISA prevalence increased, resulting in a decrease in the Kappa. Following the above conjecture, a decrease in prevalence would be associated with a decrease in Kappa However it does not explain the concurrent increase in ELISA prevalence. This leaves one to wonder if there was something particular about this herd, or the strain of MAP infecting the herd, resulting in a higher ELISA prevalence (perhaps proportionately more false positives, i.e. decreased specificity). This herd doubled in size over the course of the study, and did so through the purchase of a large number of cattle with little regard to JD status of individual cows or the herds of origin prior to purchase. 127 Within herd JD prevalence Within herd JD prevalence was calculated using three different outcomes: fecal culture prevalence, ELISA prevalence, and JD test status (ELISA &/or FC positive) prevalence. Prevalence was calculated at the herd level as well as for first lactation cows only. The reason for looking at prevalence in first lactation cows was due to the nature of the disease and the control programs implemented. Most of the management practices recommended to control JD started at birth. Moreover, it takes 2-5 years for JD to manifest itself. Thus, monitoring JD prevalence in first lactation animals may provide the earliest indication that the control program is working. The prevalence trends for each outcome are shown in Figures 3.1-3.6. The Cochran-Amritage test for trend (Tables 3.3-3.8) was conducted to determine if there was a significant change in prevalence over time and the direction of that change. A significant change was defined as a p-value < 0.05. Positive test statistics indicate an increasing trend in prevalence, while negative test statistics indicate a decreasing trend in prevalence. The direction of the trend and the level of statistical significance varied between and within herds depending on the outcome. The Cochran-Armitage test was calculated based on apparent prevalence. The change to the more sensitive ESP 11 culture system (Kim, et al., 2004) midway through the study may partially explain the unexpected increases in JD prevalence as well as some of the insignificant changes in prevalence trends. Being more sensitive, the ESP 11 culture likely identified JD infected cows earlier. Therefore, any prevalence dependent on F C culture (F C and JD test status) calculated based on results using the ESP 1] system were likely inflated as compared to those obtained with the HEY culture system. The ESP II system was put into use in 128 between the second and third annual herd tests. This corresponded to when cows born and raised with the JD control program in place began to be tested, and when prevalence was expected to decline due to improved management practices. However, regardless of the outcome, a large proportion of the herds had negative Cochran-Armitage test statistics. Thus, overall, there appears to be a general trend for decreasing JD prevalence despite a lack of statistical significance, suggesting that the JD control programs put in place on these herds are working. Effectiveness of JD control program Before specific management practices could be evaluated, it first had to be determined if the prevalence and/or incidence had changed in response to implementing the JD control program by each herd. If there was no change in the JD prevalence or incidence, the analysis could go no further. The Cochran-Armitage tests for trend calculated for JD prevalence across each respective herd as well as for prevalence in first lactation cows only provided preliminary support that the JD control programs implemented were working, but more definitive evidence was desired. One of the easiest ways to evaluate the impact of the respective JD control programs was to calculate the R of JD prevalence in cows exposed to the control program to that in cows not exposed to the program (Tables 3.9-3.11). For all herds, the RR comparing the JD prevalence of exposed cows to that of unexposed cows was < l and statistically significant (p <0.05), or approaching statistical significance. A R < 1 suggests that the JD control programs implemented on these farms did indeed have a protective effect. 129 Care must be taken when interpreting these RR’s, as they may overestimate the true effect. There is only 4-5 years worth of data on each of these herds. Cows “exposed” to the JD control program were only 2-4 years of age at the end of the study. Potentially, a proportion of these cows were indeed infected, but the disease had not progressed to the point where it could be detected by the tests being used. Meaning, the numerator could, in reality, be higher. Meanwhile, the denominator of the RR would not change much, as the youngest “unexposed” cows were at least four years of age at the end of the study, and it is unlikely a large number of those cows would test positive had the study been continued. However, as the RR’s were all rather small, (ranging from 0.09 — 0.64) it seems improbable that the “protective effect” of the JD control program would have been reversed had these cattle been followed fiirther. The diagnosis of JD at an early age is indicative of high infection rate and high infection pressure on young cattle in the herd of origin (Collins, 2003). It follows then, that if control practices are successfirl in preventing infection, the infection pressure on young cattle will decrease; which will, in turn, decrease the overall new infection, or incidence rate, and increase the average age at which infected cattle are diagnosed. When the ELISA or JD test status was the outcome, there was a lower incidence of JD in the cows exposed to the control program across all lactations analyzed (tables 3.13-3.14). The fact that this did not hold true when F C was the outcome of interest may be real or may be due to the change in culture systems used. Because it is more sensitive, the ESP II system may have been detecting infected cows earlier than HEY. As all the “exposed” cows were cultured using the ESP II system, it is possible a proportion of those cows culturing positive would have cultured negative with HEY, and not detected until some 130 later test date. This would explain the increase in the incidence of FC positive cows in lactation l in the exposed group as compared to the unexposed group, when it was expected that the opposite would occur if the JD control programs were effective. Furthermore, as JD test status consisted of both PC and ELISA test results, the magnitude of the decrease in the incidence rate between cows exposed and not exposed to the JD control program, may have been obscured by the culture system used to classify the JD test status of the cows. However, as already noted, when JD test status was the outcome of interest, the incidence rate was consistently lower in cows exposed to the control program, and that decrease in incidence was statistically significant in all but the second lactation, and was approaching statistical significance (p = 0.1) in lactation 2. In conclusion, after evaluating the RR and incidence rates of JD by lactation for cows exposed to the control program compared to cows not exposed to the control program, it was determined that the JD control programs implemented were effective in reducing the JD burden in these herds. Determining which management practices are effective in JD control programs Given the evidence that the JD control programs implemented on the study herds were successful in reducing the prevalence and incidence of JD in these herds, the next step was to determine which management practices (as measured by RA scores of risk factors) were most effective. Due to the way the outcome was modeled (probability that JD test status was positive), it was expected that all the OR would be >1. After all, the scoring system used was based on biologically proven risks of JD transmission—the higher the scores, the greater the risk of cows becoming infected with JD. The p—values 131 could then be used to sort out the importance of the respective risk factors to JD control. Also, the direction of the OR is more important than its actual magnitude. This is because the risk scores were subjective and specific to the herds in this study. It would be inadvisable to extrapolate the magnitude of the OR to herds outside of this study. However, the trends established in this study are valid and should extend to the general dairy herd population. Potential confounders, age and lactation number, were both statistically insignificant on univariable analysis and were not included in the multivariable analysis. Because of the characteristic slow progression of JD, it was expected age, or lactation number as a proxy for age, would be an important risk factor for testing positive for JD. This may be an artifact of the herds in this study. The herds with the highest JD prevalence in this study tended to be “younger” on average than herds with lower prevalence, particularly at the beginning of the study. The question then becomes whether this is a function of the JD process over time. Perhaps high levels of MAP contamination on these herds resulted in the cows being exposed repeatedly to high infectious doses, which accelerated the disease process allowing it to be detected at an earlier age and is consistent with previous observations (Collins, 2003). In spite of being statistically significant on univariable analysis, neither exposure to the JD control program or culture system was included in the multivariable analysis due to collinearity issues with the risk factor analysis. As expected, cows exposed to the JD control program had lower risk scores than cows that were not exposed to the program Also, the ESP II system was introduced two years into the study. This concurred with cows exposed to the JD control program entering the test population, 132 while all the cows tested with the HEY system were not exposed to the JD control program and had higher risk scores. Being exposed to the JD control program did decrease the probability of a cow testing positive for JD (Table 3.15). This was expected as it was the intent of the study to implement a control program that would decrease the incidence of JD on these farms. It also provides further evidence that the control programs are working. The magnitude of the estimate (OR=0.29) may be overestimated here. Cows exposed to the JD control program in this analysis were still relatively young, 2-4 years old at the conclusion of this study, and a proportion of them may have been infected but tested negative. In this analysis, cows cultured with the ESP II system were less likely to test positive for JD (OR=0.49) than those cultured with HEY. On the surface this would seem to contradict that the ESP H system is more sensitive than HEY and capable of detecting cows earlier in the course of the disease (Kim, et al., 2004). However, both culture systems were not being run concurrently during the study. It is also important to remember the ESP H system was not put into use until halfway through the study. This means the herds had at least one, and sometimes two years of testing using the HEY system By the time the ESP II system was put into service, most of the JD infected cows in the more advanced stages of the disease had been culled, and cows exposed to the JD control program were entering the herd. Thus, the overall prevalence in the population of cows cultured using ESP II was lower than that in the population of cows cultured using HEY. As expected, on univariable analysis, the majority of the risk factors did have OR’s >1, and many were statistically significant. However, on multivariable analysis, 133 only two factors remained in the model: exposure to adult cows other than dam at birth and feeding colostrum (pooled or not) from one cow to multiple calves (Table 3.16). Both seem biologically plausible and are similar to findings in previous studies (Thoen and Moore, 1989; Obasanjo, et al., 1997; Johnson-Ifearulundu and Kaneene, 1998; Wells and Wagner, 2000; Muskens, et al., 2003; Ridge, et al., 2005; Nielsen and Toft, 2007). Adult cows are the animals most likely to be shedding significant amounts of bacteria, so the more cows a calf on an infected farm has contact with, the greater the probability that one of those cows is infected and shedding. Likewise, if a cow is infected with JD and shedding MAP into her colostrum, feeding that colosn'um to multiple calves increases the likelihood of infecting all the calves. It was interesting that manure build up in the maternity pen and the cleanliness of the dam fell out of the multivariable model. Manure build up was borderline significant on univariable analysis (p=0.08), but manure soiled legs and udders was highly significant (p<0.0001). It was assumed that manure build up would result in more manure soiled legs and udders, and so the two were combined in the multivariable analysis, but fell out in the second round It was expected that a calf being born into a pile of manure to a dam coated in manure would have a high probability of testing positive for JD as an adult. The fact it fell out of the model in this study should not condone the neglect of maternity pen and cow cleanliness. It is more likely the result that, in general, the cows and maternity pens on the farms in this study were fairly clean. The fact that all factors relating to areas other than the maternity or pre-weaned calf areas either were not significant on univariable analysis or fell out of the final multivariable model, underscores the importance of disease transmission at or in the weeks 134 immediately following birth. It is also consistent with decreasing susceptibility to becoming infected with MAP with age (Larsen, et al., 1975). Unexpectedly, the results of the univariable risk factor analysis (Table 3.15) revealed one variable with an OR < 1 (contamination of post-weaned heifer water with manure from adult cows), however it was not statistically significant. Upon further investigation, the risk scores for this particular variable were higher in herds with the lowest JD prevalence than in herds with the highest prevalence in this study. As a result, there were proportionately more test positive cows with low scores in these areas than test positive cows with high scores. Thus, while statistically (irrespective of level of significance) this factor may appear “protective,” it was simply a function of the herds in this study, and is not biologically plausible given current knowledge of JD. Model fit analysis for univariable and multivariable logistic regression Regarding analysis of regression model fit, there were no statistically significant p-values on either the Hosmer-Lemeshow analysis or the cumulative sums of residuals analysis (Table 3.17) that would support a conclusion that the respective 1mivariable and multivariable regression models do not fit the data. In order to perform the Hosmer- Lemeshow analysis the data set needed to be divided into groups of approximately the same size in an ordinal manner. For this analysis, the data set was sorted by the expected probability that the observation would test positive for JD in the respective regression model. The way the risk scores were assigned to each observation in this study resulted in cohorts of cows having the same risk score profile, and thus the same expected probability of testing positive for JD. Unfortunately, the number of observations in 135 adjacent cohorts varied greatly across the data set. The grouping for the Hosmer- Lemeshow analysis was dictated by the size of these cohorts. The data set was divided into as many groups as possible while trying to retain roughly the same number of observations in each group. As demonstrated in Table 3.17, the data for variable A (exposure to adult cows other than dam at birth) was more equally distributed than that for variable B (feeding colostrum fiom one cow to multiple calves), which allowed the data set to be divided into more groups when performing the Hosmer-Lemeshow analysis of the univariable model for variable A. Across the respective models there was no data to suggest the models did not fit the data, but the Hosmer-Lemeshow test has limited power to detect departures from the assumed model. Therefore, non-significant p-values may not mean too much, and a cumulative residual analysis was performed as a more sensitive method for assessing model fit. Given the objective of this study, the regression analysis was set up as a marginal model rather than a subject specific model. Thus, traditional methods of residual analysis and assessing model fit do not apply, and the cumulative sums of residuals analysis has been proposed as a more appropriate method for assessing regression models with aggregated residuals resulting from clustered data (Lin, et al., 2002). From a purely statistical standpoint, the univariable model for variable A appears to fit the data the best (cumulative residual p-value =0.55), as compared to the models including variable B. The univariable model for variable B (p=0. 14) does not appear to fit the data particularly well, and contributes to the marginal fit of the multivariable model. Squaring variable B appeared to improve model fit statistically, but it still did not fit the data as well as the model with variable A alone. Reviewing the raw data, it was observed that the 136 distribution of variable B was highly skewed to the left. Furthermore, only 39% of the observations with the highest risk scores for variable B originated from the herd with the highest within herd JD prevalence. The other 61% of the observations with high variable B scores originated from the three herds with extremely low JD within herd prevalence. Meaning, while statistically there was some evidence to support that high scores for variable B were associated with cows testing positive for JD (as evidenced in the final multivariable regression model); there was also a substantial number of cows with high scores for variable B that tested negative JD. Conversely, in considering the raw data for variable A, there was an obvious trend for herds with the highest within herd JD prevalence to have the highest scores for variable A, leading to a higher proportion of JD test positive cows having high variable A scores than cows testing negative for JD. This is specific for this particular data set, and care should be taken in its interpretation. It does help explain why the models including variable B do not fit the data as well as the univariable model for variable A. From a practical standpoint, it simply suggests that when JD prevalence is low, or absent, a farm can get away with risky practices for JD transmission, such as feeding colostrum from one cow to multiple calves, at least for a time, because there is a lower probability the colostrum came from an infected cow and is contaminated with MAP on these farms. There are other issues affecting the results of this analysis. First, the majority (70%) of the cows analyzed in this study were born and raised prior to the implementation of JD control programs on these herds. Second, in order to evaluate something statistically you need variability. We could not evaluate all the risk factors assessed on the herd RA because there was little or no variation in the scores for the 137 herds across the years. This lack of variation, along with the rarity of positive test results, limited the power of the study. The lack of power in this study was evidenced by relatively wide confidence intervals. A formal power analysis was not performed for two reasons. The first being it would be a post-hoe analysis, which is generally frowned upon in most statistical and epidemiological circles. Second, calculating power for longitudinal studies of this design (with cows clustered within herds of unequal size) is still being debated with no apparent consensus. Effect of dam ’s JD test status on JD test status of offspring Cattle become infected with MAP through the ingestion of the bacteria from a contaminated environment, colostrum or milk. Calves can also become infected in-utero, but this occurs in small proportion of animals and generally only when the dam is in the more advanced stages of the disease (Sweeney, 1996). The R of JD test positive cows having a JD test positive darn was 1.4 and statistically significant (p=0.01). This would suggest that cows with JD positive dams are 40% more likely to test positive themselves. Again, care must be taken in interpreting the magnitude of this effect. Dam JD test information was available for only about one third of all cows tested. It is possible that some cows were born to a JD infected dam, but the dam was culled before she tested positive. Also possible, cows from infected dams were not followed long enough in this study for them to become test positive themselves. In both instances, the calculated RR would be underestimated. Another possibility is that this association, or some part of it, may be due to factors that are confounded with or interacting with dam test status. There was insufficient data to include dam status in the 138 regression analysis. The positive association between the JD test status of the dam and the JD test status of her offspring could be due to direct contact with the dam, through in- utero transmission of MAP, or ingestion of MAP from contaminated colostrmn or environment. It could also be reasoned that, because the dam was infected, the environment was so contaminated, or the farm management such, that MAP infection of the calf was likely regardless of dam’s test status. 3.6. Conclusion In summary, while the incidence of JD did decrease following the implementation of JD control programs on these farms, it is difficult to draw any conclusions regarding the ranking of specific risk factors on the transmission of JD. What is apparent is that risk factors associated with the maternity pen and pre-weaned calf areas are critical areas to focus control efforts. This is supported by the fact that the variables that were statistically significant in the multivariable model fell within those areas. Also, the scores for the risk factors in these areas had the greatest range of distribution. The JD risk assessment scores across the herds for the weaned heifer, bred heifer, and cow areas did not vary much, which precluded many of them from this analysis. This suggests that these herds, at least in terms of risk of JD transmission, were managed similarly in these areas. Yet the JD prevalence in these herds varied greatly. The number of purchased cows was small in all but two herds. In fact, the herd with the highest JD prevalence in this study had been completely closed for over 30 years. Thus, the difference in the JD prevalence in these herds must be due to different management practices, and the maternity and pre-weaned calf areas were the areas on these farms where management 139 varied the most. Finally, focusing JD control from birth to weaning is logical if one accepts that the susceptibility of calves becoming infected with MAP decreases as they mature. 140 CHAPTER 4 ECONOMIC EVALUATION OF JOHNE’S DISEASE CONTROL PROGRAMS IMPLEMENTED ON SIX MICHIGAN DAIRY FARMS 141 4.1. Abstract J ohne’s disease (JD) is an untreatable, chronic infectious disease that is becoming increasingly prevalent in dairy herds throughout the US and the world; resulting in substantial economic losses. However, information on the costs of controlling the disease is limited, yet necessary, if producers are to make sound decisions regarding JD management. The purpose of this paper is to describe a method for evaluating the cost- effectiveness of implementing management changes to control Johne’s disease on infected dairy farms. A five-year longitudinal study of six dairy herds infected with JD was performed. Each herd implemented a JD control program upon study enrollment. Prevalence of JD within each herd was monitored with annual testing of all adult cows using fecal culture and/or serum ELISA Individual cow production and culling information was collected to estimate the annual economic losses caused by JD. A questionnaire to collect economic data was developed and administered to each herd annually to estimate costs directly attributable to the JD control program Based on the costs of the control program, and using the losses to estimate the potential benefits, the net present value (NPV) of the control program was calculated for each herd during the study and projected into the future for a total of 20 years. The NPV was calculated for four different scenarios: (1) assuming a linear decline in losses following year five and disease eradication by year 20 with the control program; (2) assuming losses and JD prevalence remain constant at year five levels while continuing the control program; (3) assuming linear increase in losses at rate equal to that in scenario 1 with no control program; and (4) assuming losses remain constant at the same level as the beginning of the study with no control plan implemented. The NPV varied greatly across the herds. 142 For scenario 1, only three herds had a positive NPV; and only two herds had a positive NPV under scenario 2. In the absence of a control program, the NPV’s were always negative. When calculated across all cows in the herd, the costs of the JD control prograrrrs implemented on these herds averaged $30/cow/year with a median of $24/cow/year. The annual losses due to JD averaged $79/cow/year with a median of $66/cow/year. Investing in a JD control program can be cost effective, and doing something to control JD was always a better economical decision than doing nothing. 4.2. Introduction J ohne’s Disease (JD), an infectious disease of cattle and other ruminants caused by the bacterium Mycobacterium avium paratuberculosis (MAP), is becoming increasingly prevalent, especially in US dairy herds. In 1996, the National Animal Health Monitoring Service (NAHMS) Dairy study estimated the prevalence of dairy herds infected with JD in the US to be 21 .6% (USDA, 1997). In the 2007 NAHMS Dairy study, that estimate increased to 68.1% (USDA, 2008). Other estimates range fi'om 21- 93%, depending on region and testing method used to classify infected herds (Collins, et al., 1994; Obasanjo, et al., 1997; Thorne and Hardin, 1997; Johnson-Ifearulundu and Kaneene 1998; Johnson-Ifearulundu, et al., 1999; Adaska and Anderson, 2003; Hirst, et al., 2004). Several estimates of the economic impact of JD have been made. These estimates vary depending on study design, JD prevalence, and herd performance. Based on the NAHMS 1996 study, JD cost the US dairy industry an estimated US$ 200-250 million annually (Ott, et al., 1999), due primarily to reduced production and cull value of infected cows and increased replacement costs (Wells and Wagner, 2000). When spread across 143 all cows in the herd, estimates of losses due to JD range fiom USS 20-33/cow (Meyer and Hall, 1994; Ott, et al., 1999; Chi, et al., 2002). In a simulated study, the mean loss due to JD on a typical midsize US dairy herd started at $35/cow in the first year following JD introduction, and increased to > $72/cow by year 20 in the absence of a control program (Groenendaal and Galligan, 2003). As JD prevalence increases so too do the economic losses incurred by the disease. While quantifying the losses caused by JD is inrportant, it represents just one of the necessary components needed for making on-farm decisions regarding JD management and control. Also needed are estimates of future losses that can be prevented by implementing control practices, and how much those control practices will cost. From a producer perspective, it is difficult to justify investing in a disease control program if the cost of controlling the disease is greater than the costs being incurred it. Only when all three components (estimated losses with and without control and costs of the control program) are known, can return on investment be estimated, allowing producers to make informed economic decisions. Multiple computer simulation studies have attempted to estimate the benefits and costs of various JD control practices (Groenendaal, et al., 2002; Dorshorst, et al., 2006; Kudahl, et al., 2007). While simulation studies are commonly used and have the advantage of costing less and requiring less time as compared to field studies, they can be difficult to validate. The input data required for these simulations is often based on field data supplemented by expert opinion. Thus, a simulation study cannot be isolated from the field; and, in fact, can only truly be validated with observations made under real farm conditions once a field study is performed. 144 The chronic and insidious nature of JD is the most likely reason studies attempting to quantify the costs and benefits of JD control have been limited to simulations to date. A field study would have to be of sufficient length that changes in herd performance resulting fi'om a JD control program are observed. As most management practices to control JD focus on preventing young calves from becoming infected with MAP (Collins, 1994; Rossiter and Burhans, 1996), and infection generally does not become detectable until infected calves become adults (Sweeney, 1996), a field study would have to last a minimum of three years to see results fi'om the control program The study reported here, to our knowledge, is the first longitudinal observational field study investigating the costs of implementing management practices to control JD on infected dairy farms. The objective of this study was to estimate the net present value (N PV) of the JD control programs implemented on each of six Michigan dairy farms to aid producers in making more informed decisions regarding JD control and management. This chapter consists of a series of case reports. The economic analysis performed is described for each herd separately, and the results summarized across all herds. 4.3. Materials and Methods Study design This was a five—year longitudinal observational study of six Michigan dairy herds infected with MAP. 145 Farms This study was part of the larger Michigan Johne’s Disease Control Demonstration Project. A total of six Michigan dairy herds participated in this study. (Herd 7 discussed in previous chapters was not included in this part of the project. It was a university-owned research herd; and, as such, the economic decisions made on this farm were not necessarily consistent with those of other commercial dairies.) The herds were chosen based on the producer’s willingness to participate in a longitudinal study for at least five years, and because they were representative of typical Michigan dairy farms in terms of herd size and housing management. Types of housing management targeted included total confinement, combination of confinement and grazing, and rotational grazing. All herds were infected with JD and initiated a control program upon enrollment into the project. The JD control programs implemented were not standardized or static. Instead, the programs were designed specifically for each herd based on the operation’s goals and capabilities, and modified as necessary. The within herd prevalence of JD on these farms was determined annually by fecal culture and/or serum ELISA of all adult cows, and was used for monitoring the effectiveness of the respective control programs. Herds were enrolled in the study beginning in January 2003 with the last herd enrolled in May 2004. Questionnaire used for economic data collection A questionnaire was developed to collect data regarding costs directly attributable to the JD control program (Appendix A). This questionnaire was broken down into four sections specific to each farm’s management: supplies, management, labor, and capital 146 investments. The questionnaire was administered to the herd owner/manager beginning the first year following implementation of the JD control program and annually thereafter through calendar year 2007. Additionally, questions regarding the manager’s perception of, and satisfaction with, the effectiveness of the JD control program were asked the last time the questionnaire was administered in 2007. Other data collected Individual cow production and cull information was collected when available. Information was obtained from the Dairy Health Improvement Association (DI-11A) for four herds. One herd had computerized daily milk weights, and the remaining herd had hand-written records only with no individual cow production data. Data analysis A. JD prevalence Within herd JD prevalence was calculated for each herd based on annual whole herd fecal culture and/or serum ELISA. Johne’s disease prevalence was calculated as the number of cattle testing positive (regardless of test) divided by the total number tested. B. Cost of JD control program The cost of the control program was calculated based on information obtained from the questionnaire. 147 The supplies category consisted of things such as milk replacer, additional ear tags or other identification for JD test positive animals, change in volume of bedding, sanitation supplies, etc., that were a direct result of implementing the JD control program For herds that switched fiom whole milk to milk replacer, an adjustment was made for the sale of milk that would have previously been fed to the calves. It was assumed that one 50 pound (23 kg) bag of milk replacer was needed to wean one calf, and was equivalent to 400 pounds (181 L) of whole milk (Groenendaal and Galligan, 1999). It was further assumed that all milk previously fed to calves was marketable. The number of calves weaned per year was calculated assuming non-seasonal calving, a 14-month calving interval with 50% heifer calves born. All farms sold bull calves within 1-2 days of birth, and thus deacon calves were not included in the calculation. All herds belonged to the same milk marketing cooperative, although one farm became certified organic during the course of the study. The adjustment for additional milk sold was valued as the average base farm price paid by the cooperative for each respective year. For the herd that became certified organic, the cooperative base farm price was used for the adjustment up until certification, with the average yearly price received by the farm used thereafter. The intent of the management category was to account for the time (hours) herd managers spent coordinating the JD control program. This would include things such as testing, decision-making, and employee education. For herds that had a full-time herd manager (N=2) earning an annual salary, the approximate hourly wage was calculated as the total value of compensation (salary + benefits) divided by the average number of hours worked. In the other four herds where the owner was the herd manager, the value 148 of management time was based on the owner’s estimated value ($/hr) of their time given their education and experience (what they could be making doing something else), or what they estimated they would have to pay to hire a manager. Management costs were then calculated as the hours committed to the JD control program multiplied by the estimated hourly value of management. The labor category included the additional hours employees spent implementing the JD control program such as improved cow and calf hygiene, increased time spent caring for calves, etc. All farms had hired part-time labor paid hourly wages. Labor costs were calculated as additional hours spent performing tasks required by the JD control program multiplied by the hourly wage paid Capital investnrents included things such as a pasteurizer, calf hutches, fencing, skid-steer, and/or loader buckets purchased as a direct result of the JD control program. Capital investments were converted to annuities based on the useful life of the purchase (as determined by the producer), and those values used annually for the evaluation period. In other words, once the useful life of the investment expired, it would be replaced as necessary. The annual capital cost was calculated as the purchase price multiplied by the annuity factor. The annuity factor was calculated as i/ [1-(l+i)'“] (Olson, 2004, pp. 407); where i is the interest rate, assumed to be 7%, and n is the useful life of the investment. Also included in this category was any interest paid on capital purchases that were financed. The yearly costs by category were summed together to calculate total costs. The annual total costs were then divided by the number of adult cows tested in the herd to 149 estimate the cost/cow. The annual costs/cow were averaged over the observed study period to estimate the overall cost/cow/year. C. Losses due to JD The economic losses for the following three aspects were calculated for JD infected cows identified by annual testing: decreased milk production, loss of future income due to premature or suboptimal culling, and reduced cull value. Decreased milk production The mature equivalent 305 day (ME305) milk and butterfat production were obtained for all cows when available. These were used to calculate the 3.5% fact corrected milk (FCM) for each cow using the following equation: 3.5% FCM = (0.4323 x pounds of milk) + (16.216 x pounds of fat) (Hutjens, 2005). The average annual 3.5% FCM production was calculated for each of the following groups of cows annually in each respective herd, controlling for lactation number and days in milk (Proc GLM, SAS 9.1, SAS Institute, Inc., Cary, NC): cows culled due to clinical JD, cows testing positive for JD and culled for reasons other than JD, cows testing negative for JD and culled, cows testing positive for JD and remaining in the milking herd, and cows remaining in the milking herd testing negative. For all cows testing positive for JD, milk production lost due to the disease was estimated by subtracting each individual cow’s production from the average milk production of test negative cows remaining in the herd, and then summing the differences over each study year. For test positive cows culled when they 150 were too fresh for a valid ME305 to be calculated, the average for her respective group was used as a proxy. For the herd that did not have individual cow production data, the annual pounds of milk and butter fat sold was obtained, the 3.5% FCM calculated and divided by the average number of cows in the herd that year to estimate the yearly rolling herd average. Lost milk production for cows testing positive for JD was imputed based on milk production losses averaged across the other five herds in the study. Cows testing positive for JD but not exhibiting clinical signs were assumed to produce 12% less milk than test negative cows, while cows with clinical JD were assumed to produce 23% less milk than their test negative herdmates. The annual pounds of milk lost due to JD was calculated as the number of cows with subclinical and clinical JD each year multiplied by the product of the rolling herd average times 12% and 23% respectively, then summed together. The total economic losses resulting fi'om reduced milk production due to JD was calculated by multiplying the estimated number of pounds lost by the average price/cwt received by the producer for each respective year. The cost of milk production lost per cow in the herd was calculated by dividing the total economic value of milk lost each year by the number of adult cows tested in the herd that year. Loss olfilture income due to premature culling When cows are culled due to JD the producer often sustains losses from two aspects. First her net income stream is lost, provided she has not reached optimal culling age. Second, JD infected cows generally weigh less, especially cows exhibiting clinical signs, which lowers their cull value. Future productivity will be addressed first. 151 A recently developed computer spreadsheet model (OptiCowTM, Model v1.4, Center of Animal Health and Productivity, University of Pennsylvania, Kennett Square, PA) allows the retention pay-off (RPO) value of individual dairy cows to be calculated. The RPO-value of a cow is defined as the total additional expected profit if the cow is kept until her optimal age as compared to her immediate replacement. It is an economic index that can be used to rank cows by their future profitability; the higher the RPO, the more valuable the cow. A negative RPO means replacement is the preferred action (Groenendaal and Galligan, 1999). For herds with individual cow production data (N=5), the RPO values were calculated for cows culled due to clinical JD and/or test positive for JD throughout the study, and summed together to obtain the total RPO-value lost due to JD per year. The loss of future productivity per cow for the herd was estimated by dividing the total RPO-value of all test positive cows culled by the number of adult cows in the herd for each respective year. The loss of future productivity for the herd without individual cow data was not calculated. (The farm input data for the OptiCowTM Model are summarized in Appendix C). While the RPO-value estimates future production potential, it does not include slaughter value. Due to the pathogenesis of JD, infected cows lose weight and may exhibit diarrhea, depending on the stage of infection. Thus, they tend to weigh less than uninfected cows when culled, resulting in lower slaughter value. The economic loss due to reduced cull income was estimated. Previous studies report losses in slaughter value ranging from 10-37.5% (Benedictus, etal., 1987; Ott, et al., 1999), with higher losses occurring with the more advanced stages of the disease. 152 Cull cow data was available for all herds, although in some cases it was incomplete. When possible the reduction in cull cow income due to JD was calculated using the following guidelines. Test negative Holstein cull cows were assumed to weigh an average of 1400 pounds (636 kg) and Jerseys 800 pounds (364 kg). Cows that were culled due to clinical JD were assumed to weigh 30% less (420 pounds Holstein, 240 pounds Jersey) than test negative cows during the first year of the study, and 15% less (210 pounds Holstein, 120 pounds Jersey) thereafter. The loss in body weight due to JD was changed because, after enrolling in the JD demonstration project and the start of annual whole herd testing, the working definition of “culled due to clinical JD” changed. Producers were more cognizant of the disease and quicker to cull a cow as soon as she started to lose weight or developed diarrhea, especially if she happened to test positive for JD. Cows that tested positive, but were culled due so some reason other than JD, were assumed to weigh 10% less (140 pounds Holstein, 80 pounds Jersey) than test negative cows throughout the study. In some instances, records were complete enough to identify cows that were sold for slaughter and those that died. Obviously if a cow died, the producer did not realize any cull income. The loss in cull value due to JD was then calculated as the total weight lost due to cows culled with JD (either clinical or test positive) multiplied by the respective slaughter value for each year. In the event a test positive cow died, the loss was calculated as the entire value of the cow, equaling the average weight minus 10%, multiplied by the respective slaughter value for that year. The total losses for each year were then divided by the number of adult cows in the herd that year to estimate the lost cull value per cow in the herd. 153 Benefits of JD control program The economic benefits, or the reduction of losses, due to the JD control program were calculated as the difference between the annual economic losses due to JD fore each year following the implementation of the control program and a baseline measure of the losses caused by the disease prior to the control program The baseline was estimated as the average of the losses due to JD over the first two complete years of the study. It was believed that the completeness of the data for the first year of the study depended on when, in the course of that year, a herd was enrolled in the program (herds enrolled in the spring were more likely to have more complete data than herds enrolled in the fall). Furthermore, as the majority of the management changes implemented to control JD were intended to prevent new infections in young calves, no benefits resulting from the control program were expected until the third year of the study at the earliest. Averaging losses over the first full two years would, therefore, provide a more accurate baseline measure of losses caused by JD in the absence of a control program. Thereafter, the annual benefits of the JD control program were estimated as the losses due to JD at baseline minus the losses due to JD in the subsequent study years. Both costs and benefits were used to calculate the net present value (NPV) of each farm’s JD control program over the course of the study. Simply put, the NPV is the value of the expected future returns of an investment minus the value of expected future costs, discounted to current dollars. It is commonly used in economics as a method for appraising long-term projects. All cash flows used to calculate the NPV were discounted back to the first year of the study to account for the time-value of money and the risk of the investment. Additionally, the NPV was projected over a total of 20 years assuming 154 four different scenarios: (1) the economic losses beyond the observed study period follow a linear decrease with eradication of JD from the herd 20 years after the start of the control program; (2) the economic losses stay constant at a rate equal to that of the last observed year of the study while continuing to invest in the control program; (3) the economic losses increase from baseline equal to the rate of decrease in scenario 1, in the absence of a JD control program; and (4) the economic losses remain constant at the baseline level in the absence of a control program. The reason scenarios 3 and 4 were calculated was to demonstrate potential economic losses should the farm elect not to implement a JD control strategy. The NPV was calculated as follows: n C! Cf NPV = + 121 (1+ rY r(1 + r)"+1 Where: t = the index of time (year), n = the total number of periods (years) during which cash flows were estimated, r = the discount rate, C, = the net cash flow for period t, and C f: the constant net cash flow expected in years beyond n. To estimate the NPV beyond 2007 (t = 4), or the observed study period, some assumptions were made. For scenarios 1 and 2, the projected ongoing costs of investing in the JD control program were assumed to equal the average annual cost of the control program during the observed study period. In scenario 1, it was assumed the losses due 155 to JD would follow a linear decline until disease eradication in year 20, when the losses would equal zero. The observed loss in 2007 (t = 4) was divided by 16, to estimate the necessary annual decrease in JD losses resulting in disease eradication in t = 20. For years t = 5 to 20, this result was subtracted fi'om the previous year’s loss to calculate the loss for each respective year t. In scenario 2, the observed loss in 2007 (t = 4) was assumed to remain constant for years t = 5 to 20 respectively. For scenario 3, an amount equal to the annual decrease in JD losses for scenario 1 was added to the baseline beginning in year t = l, and increased by the same increment in all subsequent years until t = 20. In scenario 4, the baseline loss was held constant throughout the 20-year projection. The opportunity cost of capital is represented by the discount rate, r. The opportunity cost for capital used in agriculture is generally lower than that in other economic sectors. The discount rate was assumed to equal 8%, but was varied later in, a sensitivity analysis. The constant net cash flow for years beyond n, Cf was assumed to be equal to C 1 when t = n. Dividing Cf by r resulted in a terminal value, or perpetuity, calculation to reflect future benefits from investing in the JD control program. As with the net cash flows for each period, t, the terminal value, was discounted back to the start of the program. Sensitivity analysis A sensitivity analysis was performed on the NPV calculations to determine which inputs had the greatest influence on the final calculation. The discount rate, r, was varied from 5% to 10%. All other input factors were varied by i 10% including: overall cost of 156 JD control program, individual components of the JD control program (supplies, management, labor, and capital investments), milk price, cull price, and RPO-value. The break-even cost for the JD control program was calculated for the two scenarios (1 and 2) that included investing in a control program. Testing for JD was also included in the sensitivity analysis. Testing was provided free of charge to the herds during the observed study period and involved annual whole herd fecal culture and/or serum ELISA. It is unlikely the producers would have invested in such an intensive testing program had they incurred the testing costs, yet they used the test results to make management and culling decisions. Moreover, it is probable the producers will continue to do some JD testing after the study, so it was important to estimate the effects testing would have on the estimated NPV. Testing costs used in the sensitivity analysis were set equal to the laboratory costs being charged during the study period: fecal culture, $23/sample; and serum ELISA, $6/sample. The sensitivity analysis included the scenario for what actually happened, the costs of both tests run in parallel; as well as for serum ELISA testing only. 4.4. Results Individual herd reports 157 A. Herd 1 Farm background This farm has been owned and operated by the same family for over 100 years. At the beginning of the program, the herd was milking approximately 140 Holstein cows (167 total adult cows) with a rolling herd average (RHA) of 31,516 pormds (3.5% FCM). Bred heifers and heifer calves were occasionally purchased, mainly in an attempt to improve herd genetics, not because they were needed to maintain herd size. Considerable thought went into the purchase of these animals and they were purchased directly fiom herds which were at low risk for disease (including JD). Johne’s disease was first diagnosed in the herd in the mid-1980’s. In 2003, 5.5% of the cows culled were due to clinical signs of JD. The youngest animal to develop clinical JD was an 18 month-old home-raised heifer in January of 2002. She was the wake up call the farm needed to realize they needed to take steps to control JD. Aside from JD, this herd had very few other health problems. The producer reported that the annual incidence of all periparturient diseases combined, was less than 5%. There were occasional summer flare-ups of environmental mastitis, and the bulk tank SCC averaged 300,000. The long term goals for this farm at the beginning of the study were: 1. Expand (internally) to milk 200 cows 2. 30,000 lb RHA 3. Market dairy replacements 158 JD risk assessment Prior to the implementation of a JD control program, the area at greatest risk for disease transmission on this farm was the calving area The farm had individual maternity pens, but they were not always cleaned between each calving. No consideration was given to a cow’s JD status when placed in the pens. About 5% of calves were born in free stalls. Most of the calves (80%) were removed from the dams within two hours of birth. However, pooled colostrum was used for feeding. The other issue with the calving area was its proximity to weaned calves. It was an old basement barn. Weaned calves were kept in group pens just across a six foot alley fiom the maternity pens. Occasionally manure slurry ran from the maternity pens, across the alley and into the calf pens. Other risks for JD transmission prior to the control program included feeding left- over feed from the lactating herd to replacement heifers; and, in the summer, bred heifers had fence-line contact and shared a waterer with dry cows on pasture. JD control plan Upon enrolling in the Michigan Johne’s Disease Control Demonstration Project, annual testing of all adult cows for JD with sermn ELISA and fecal culture began All cows testing positive on either test were visually identified with a distinct ear tag and flagged in the computer for management. Cows positive on fecal culture and exhibiting clinical signs were culled immediately. Fecal culture positive cows not showing clinical signs were not bred back and were culled when they either began to exhibit clinical signs, or their milk production fell to below a break-even point defined by the farm. 159 A matenrity pen was designated for calving all JD test positive cows. A greater effort was placed on cleaning all maternity pens after each calving. Pooled colostrum was no longer used. Colostrum from JD test positive cows was not fed to heifer calves. Only colostrum from JD test negative cows was frozen to be used as needed Weaned calves were still housed across the alley from the maternity pens; but with more fiequent cleaning, the amount of manure contanrination from the maternity pens to the calf pens was reduced. A super hutch and fencing was purchased in the summer of 2005 to house calves that were just weaned, and relieve some of the crowding in the pens across from the maternity area. Feeding waste feed to replacement heifers was discontinued Although not done specifically for the JD control program, an existing barn was renovated for dry cows so they no longer had contact with bred heifers. Descriptive statistics Descriptive statistics for herd 1 are summarized in Table 4.1. Table 4.1: Descriptive Statistics for Herd 1 Year Herd size (“1331; Cull Rate cured dueto Mortality (adult cows) FCM) (%) Clmlcal JD (%) Rate (%) 2003 170 31,516 25.3 4.7 4.1 2004 176 29,666 38.1 11.9 5.1 2005 190 33,090 26.8 15.7 2.1 2006 204 33,744 27.5 12.5 2.5 2007 215 32,522 27.9 5.0 1.9 160 Herd size increased steadily over the study period in accordance with the farm’s stated goals. The overall cull rate, with the exception of 2004, was fairly constant over the five-year study period, and was lower than the average state cull rate of 37.7% (Hadley, et al., 2006). This is likely due to the herd trying to expand internally to bring existing facilities to full capacity of 200 cows milking. As the herd reaches this goal, it is likely the overall cull rate will increase. Mortality rate decreased over the course of the study. The number of cows culled for clinical JD increased initially and then decreased back to approximately the same level as when the study began. However, this may have been a larger decrease than suggested numerically. Following study enrollment, the producer was much more cognizant of cows with clinical signs. The farm also had access to the annual JD test results, which figured heavily into the culling decision process. Cows were culled as soon as they started exhibiting signs of weight loss and/or diarrhea. Cows testing positive for JD had one strike against them. As soon as they started exhibiting clinical signs, or developed another problem, they were culled. JD prevglence The within herd JD prevalence trend is outlined in Table 4.2. J ohne’s disease prevalence increased steadily over the first three years of the study, before declining dramatically in 2006. This pattern fit with what the producer reported seeing clinically. The number of cows exhibiting weight loss and diarrhea increased in 2004 and continued through the spring of 2006. Beginning in 2006, the producer began noticing a decline in clinical cases. The increase in prevalence in 2007 was unexpected and remained unexplained at the conclusion of this study. 161 Table 4.2: Johne’s disease prevalence trends 2003-2007 for Herd 1 Year Apparent JD prevalence (ELISA &/or PC positive) 2003 12.0% 2004 24.3% 2005 22.0% 2006 9.8% 2007 15.0% Cost of the JD control prograflg003-2007 The costs of the JD control program observed over the five years of this study are summarized in Table 4.3. As part of the Michigan Johne’s Disease Control Demonstration Project, the herd did not have to pay for any JD testing beyond the cost of labor to collect samples on the day of the annual test. It is unlikely the herd would have done the extensive testing that was performed if it had to pay for the testing, yet management decisions were made based on those test results, so for the sake of completeness, testing costs are included in the last column of the table. Table 4.3: Cost of Johne’s disease control program 2003-2007 for Herd 1 ($/cow) Year No. Supplies Management Labor Capital Total Total Cows Investments (+ testrng) 2003 170 $9.72 $1.17 $5.35 $0.00 $16.24 $45.24 2004 176 $9.27 $1.16 $5.85 $0.00 $16.29 $45.29 2005 190 $2.47 $0.47 $7.16 $0.77 $10.88 $39.88 2006 204 $2.21 $0.46 $6.67 $0.72 $10.05 $39.05 2007 215 $2.19 $0.45 $5.95 $0.68 $9.27 $38.27 Ave. 191 $5.17 $0.74 $6.20 $0.43 $12.54 $41.54 162 The supplies category included the purchase of pink ear tags to identify JD test positive cows, and colostrum replacer over the first two years of the program until the herd had enough colostrum from test negative cows banked to meet its needs. Also included was a charge for increased bedding used as a result of more frequent cleaning of the maternity and calf areas. Management costs were mame due to record management to keep track of test positive cattle. Labor costs were largely due to the increased time spent cleaning the maternity barn; although time spent assisting in the collecting of samples for JD testing was also included in this category. The capital purchase made by the farm was a super hutch and fencing in 2005 to keep just-weaned calves out of the maternity barn. Economic losses due to JD 42003-400 7 The annual estimated economic losses due to JD, along with the calculated, or assumed, benefits of the control program for Herd 1 are summarized in Table 4.4. Table 4.4: Economic losses due to Johne’s disease and assumed benefits of Johne’s disease control program for Herd 1 — 2003-2007 ($/cow) No. Milk Assumed Year RPO Cull Value Total Cows Value Benefits 2003 170 $87.03 $0.31 $1 1.61 $98.95 N/A 2004 176 $114.93 $0.00 $23.61 $138.53 -$19.79 2005 190 $45.84 $21.80 $12.33 $79.97 $38.77 2006 204 $49.41 $22.55 $11.60 $83.57 $35.17 2007 215 $50.09 $4.11 $6.02 $60.23 $58.51 Ave. 191 $69.46 $9.75 $13.04 $92.25 N/A N/A Not applicable 163 NP V cglcugtion The NPV was calculated for four different scenarios: (1) assuming a linear decline in losses following year five and disease eradication by year 20 with the control program; (2) assuming losses and JD prevalence remain constant at year five levels while continuing the control program; (3) assuming linear increase in losses at rate equal to that in scenario 1 with no control program; and (4) assuming losses remain constant at same level as the beginning of the study with no control plan implemented. The results are shown in Table 4.5. Due to the relatively small about of money invested in the JD control program as compared to the estimated potential benefits, the NPV for scenarios 1 and 2 became positive by the third year of the study. The NPV’s for scenarios 3 and 4 remain negative over the entire ZO-year projected study period because, in the absence of a control program, there was nothing to offset the losses caused by JD. Results of sensitivity analysis The break-even cost for the JD control program is an estimate of the amount of money a farm can invest in the control program and still “break-even”, or have a NPV equal to zero. The break-even cost calculated for scenario 1 was $75.64/cow/year, and $49.74/cow/year, for scenario 2. The NPV was most sensitive to the discount rate (r) used. After the discount rate, milk price, followed by cost of the JD control program had the greatest effect on the NPV. When the individual categories within the JD control program were evaluated, the results were ranked depending on which category represented the highest proportion of the 164 Table 4.5: NPV of four scenarios for Johne’s disease (JD) control on Herd 1 Year of Scenario“ Program 1 2 3 4 5 $80.49 $77.93 -$516.88 -$474. 10 10 $244.76 $202.85 -$9 1 9.80 -$796.76 20 $778.53 $459.88 -$1 ,944.65 -$1,460.67 Payback Year 3 3 N/A N/A N/A Not applicable Payback Year: Year of control program when NPV became positive * Scenario 1: Assuming linear decline in losses caused by JD after year 4 due to declining prevalence and eventual eradication after implementation of JD control program Scenario 2: Assuming losses caused by JD remain constant after year 4 while still investing in JD control program Scenario 3: Assuming losses caused by JD increase at a rate equal to the decline in Scenario 1 in the absence of JD control program Scenario 4: Assuming losses caused by JD remain constant at baseline level in absence of JD control program control program For this herd, changing the input costs of labor had the greatest impact on the NPV, followed by supplies, management, and capital investments respectfully. Including testing in the calculations increased the input cost of the JD control program, and hence decreases the NPV. For scenario 1, the NPV decreased from $779 to $705 when ELISA testing only was included, and to $434 when both ELISA and fecal culture was included. For scenario 2, the NPV decreased from $460 to $386 and $115 respectfully, for including ELISA testing only or ELISA and fecal culture together. 165 Producer perception of the JD control program Overall, the producer was very pleased with the results of the JD control program, and planned to continue investing in it after the end of the study. Subjectively, herd health improved over the course of the study, which in turn led to improved production and increased revenues. An additional value resulting from the JD control program on this farm was seen in the marketing of herd replacements. Other producers were willing to pay more for cattle raised on this farm because of the JD control program the herd had implemented and the amount of diagnostic testing that was performed. B. Herd 2 Farm background This is a second generation dairy farm that entered the Michigan Johne’s Disease Control Demonstration Project as a partnership in December 2002. At that time, the herd was milking around 85 Holstein cows (105 total adult cows) with a RHA of 26,569 pounds (3.5% FCM). Purchased cattle were last added to the herd in 2000 when 10 yearling heifers were bought from a single source with a low risk of JD (vouched for by the herd’s veterinarian). Prior to that, in 1999, six bred heifers were purchased fi'om a single source with unknown JD status. Johne’s disease was first diagnosed in the herd in 1992, within five years of purchasing six yearling heifers. The youngest clinical case of JD occurred in an 18 month-old bred heifer raised on the farm in 2002. In 2002, approximately 7% (N=7) of cows were culled due to clinical signs of JD. Aside from JD, 166 the herd historically has had problems getting cows bred back, but few overt health problems. The bulk tank SCC averages 210,000. The stated long term goals for the farm at the beginning of the study were (in order): 1. Maintain quality personal time for owners and employees 2. 60% of the adult herd pregnant at all times 3. 30,000 lb. RHA 4. Maintain or expand herd size In 2004, the partnership dissolved, and one partner was forced to buy out the other. This, along with the low milk prices at the time, put the farm in a tenuous financial position. As part of the buyout, all cows testing positive for JD were culled, regardless of production or reproductive status. This temporarily decreased the size of the milking herd, and cows that would have been culled previous to the partnership break-up were retained in order to meet cash flow needs. Labor was also an issue. Retaining quality, dependable employees was a challenge; and the farm had a high employee turnover rate, including the herd manager. As a result of these distractions, improving herd health and facilities were not top priorities, as it was a struggle to maintain the status quo. Beginning in 2007, things began to stabilize. Time and improved milk prices put the farm on better financial footing, the labor issues seemed to have been resolved, and focus was again being placed on the cows. 167 JD risk assessment Prior to the implementation of a JD control program, the areas at greatest risk for JD transmission were the calving and pre-weaned calf areas; although there were also some risks in the weaned and bred heifer areas. The farm had one maternity pen bedded with straw. It was cleaned infrequently (about once every 10 calvings), although fresh bedding was occasionally added between calvings. No attempt was made to segregate JD suspect cows and JD test negative cows in the maternity pen or the adjacent close-up dry cow area. Calves were generally removed from the cow within two hours of birth, unless they were born during the night, then they might remain with the cow for 6-8 hours. If multiple cows were calving, colostrum was pooled and fed to the calves. All calves were fed pooled, unpasteurized, whole milk. Weaned calves were fed hay in an alley adjacent to the lactating cow area, where feed could be contaminated by manure from adults. Bred heifers and dry cows were housed together in the same pen. Breeding age heifers were housed in a pen adjacent to the bred heifers/dry cows and shared the same water source. Breeding age heifers were also fed leftover feed from the lactating cows when available. JD Control Plan Upon enrolling in the project, annual testing of all adult cows for JD with serum ELISA and fecal culture began. All cows testing positive on either test were not bred back, and were culled when they developed clinical signs, or milk production decreased below some break-even point determined by the farm. Test positive cows were visually identified with a notch in their ear tag and calves born to test positive cows identified 168 with a blue ear tag with a “J”. An effort was made to clean the maternity pen more often, although not always after each calving. The maternity area was finally remodeled in October of 2007, with multiple calving pens, which theoretically will be cleaned after each calving. Colostrum from JD test positive cows was no longer used, and colostrum was no longer pooled. Extra colostrum from individual JD test negative cows was fiozen for use as needed. In the absence of colostrum, a colostrum supplement (Colostrixm) was used All heifer calves were fed milk replacer. Bull calves being raised as steers continued to receive pooled waste milk as available. Bottles used for feeding milk were sanitized after each feeding. An off-the-floor hay feeder was constructed in the weaned calf area. Feeding of leftover feed from the cows to breeding age heifers was discontinued. Descriptive statistics Descriptive statistics for herd 2 are summarized in Table 4.6. Herd size increased over the study period, and by 2007, the farm was milking approximately 120 cows, which was the desired herd size, with existing facilities at capacity. The overall cull rate fluctuated over the study period. It started off at 36%, which was close to the average state cull rate of 37.7% (Hadley, et al 2006). It increased dramatically in 2004 as a result of the dissolution of the partnership. This was followed by a substantial decrease in cull rate in 2005, likely due to the herd trying to recover from the buyout, and increase herd size to increase cash flow. The increase in the number of cows culled due to clinical JD in 2004 may be due, in part, to misclassification. Recall that all JD test positive cows were culled as part of the partnership settlement. So some of the cows in 2004 may 169 Table 4.6: Descriptive Statistics for Herd 2 Year (Hdelrldsize) (1:3; cog/1:86 fill-1810111663) $213: a tcows 00 c unca 00 oo FCM) 2003 103 26,569 35.9 6.8 2.9 2004 121 25,081 47.9 10.7 2.5 2005 134 23,854 21.6 0.7 3.0 2006 132 22,022 27.3 5.3 3.0 2007 137 25,231 25.5 1.4 6.6 have been culled due to “”JD but were not actually exhibiting clinical signs. Following the “JD cleansing” of the herd in 2004, it was not surprising the number of cows culled due to clinical JD dropped to less than 1% in 2005. The JD test positive cows, the cows in the most advanced stages of the disease, and therefore most likely to develop clinical signs, had been culled the previous year. In 2005, the herd was relatively young (as compared to previous years), with a small proportion of test positive cows, so there were fewer cows culled due to clinical signs. The number of cows culled due to clinical JD crept up again in 2006. This is likely a return to what it would have been in the absence of the buyout. Since the cleansing in 2004, all the remaining cows, and the majority of heifers entering the herd had not been exposed to the JD control program Two years later, in 2006, cows with JD had matured and were more likely to exhibit clinical signs. The decline in 2007 was likely due to a couple of different factors. First management had improved, with more attention being paid to the cows, so JD test positive cows were being removed before they had a chance to develop clinical signs. Second, the heifers entering the herd had been born and raised with the JD control program in place. Thus, it 170 was simply a reflection of the decreasing prevalence of JD in the herd. The mortality rate was fairly consistent throughout the study period with the exception of 2007 when it doubled. The reason for that increase was not reported. JD Prevalence The within herd JD prevalence trend is outlined in Table 4.7. The culling of all JD test positive cows occurred a couple of months prior to the 2004 test, which explains the decline in JD prevalence from 2003 to 2004. The subsequent rebound in prevalence is most likely the result of infected heifers not exposed to the JD control program maturing, entering the milking herd, and testing positive over the course of their first and second lactations. Table 4.7: Johne’s disease prevalence trends 2003—2007 for Herd 2 Year Apparent JD prevalence (ELISA &/or F C positive) 2002 12.1% 2003 9.5% 2004 4.1% 2005 5.0% 2006 9.4% 2007 4.2% Cost of the JD control program 2003-2007 The costs of the JD control program observed over the five-year period of 2003- 2007 are summarized in Table 4.8. As part of the Michigan Johne’s Disease Control 171 Table 4.8: Cost of Johne’s disease control program 2003-2007 for Herd 2 ($/cow) No. _ Capital Total Year Supplies Management Labor Total , Cows Investments (+ testrng) 2003 103 $9.68 $4.12 $7.96 $1.80 $23.57 $52.57 2004 121 $1.49 $4.30 $7.44 $1.53 $14.77 $43.77 2005 134 $3.28 $3.98 $6.72 $1.39 $15.37 $44.37 2006 132 $5.55 $3.33 $6.82 $1.41 $17.10 $46.10 2007 137 -$ 1 .82 $2.31 $5.84 $1.35 $7.68 $36.68 Ave. 125 $3.64 $3.61 $6.95 $1.50 $15.70 $44.70 Demonstration Project, the herd did not have to pay for any JD testing beyond the cost of labor to collect samples on the day of the annual test. It is unlikely the herd would have done the extensive testing that was performed if it had to pay for the testing; yet management decisions were made based on those test results, so for the sake of completeness, testing costs are included in the last column of the table. The supplies category included the purchase of blue ear tags to identify calves from JD test positive cows, milk replacer to feed all heifer calves, and colostrum supplement. It also included increased cost of straw due to more fi'equent cleaning and bedding of the maternity pen. The cost of supplies was adjusted to reflect the sale of milk that was previously fed to calves, and this adjustment explains the “negative” supply cost in 2007. Management costs reflect the time spent on record keeping, making management decisions regarding JD test positive cows, and employee education. Labor costs reflect the increased time spent cleaning the maternity pen, mixing milk replacer, fresh cow and calf care, as well as time spent assisting in the collection of samples during JD testing. The capital investment made by this farm was a second skid-steer bucket 172 purchased in January 2003 so one bucket could be dedicated to feed and one to manure handling. Economic losses due to JD 2003-2007 The annual estimated economic losses due to JD, along with the calculated, or assumed, benefits of the control program for Herd 2 are summarized in Table 4.9. The mass culling of JD test positive cows occurred in 2004, which likely explains the increase in losses seen in that year, and the subsequent decrease in losses in 2005. The increase in total losses beginning in 2006 reflects MAP infected cows maturing (and the disease progressing), but needing to be retained in the milking string to help meet cash flow needs. With the farm management and finances finally stabilizing in 2007, the JD losses once again began to decline. Table 4.9: Economic losses due to Johne’s disease and assumed benefits of Johne’s disease control program for Herd 2 — 2003-2007 ($/cow) Year No' Milk RPO Cull Value Total Assumed Cows Value Benefits 2003 103 $0.00 $0.00 $15.85 $15.85 N/A 2004 121 $0.00 $20.49 $28.19 $48.67 $16.41 2005 134 $13.33 $2.90 $1.82 $18.05 $14.21 2006 132 $53.60 $15.80 $12.73 $82.13 $49.37 2007 137 $42.07 $2.12 $21.97 $66.17 453391 Ave. 125 $21.80 $8.26 $16.11 $46.17 N/A N/A Not applicable 173 NP V calculation The NPV was calculated for four different scenarios: (1) assuming a linear decline in losses following year five and disease eradication by year 20 with the control program; (2) assuming losses and JD prevalence remain constant at year five levels while continuing the control program; (3) assuming linear increase in losses at a rate equal to that in scenario 1 with no control program; and (4) assuming losses remain constant at same level as the beginning of the study with no control plan implemented. The results are shown in Table 4.10. The reason the NPV’s for both scenarios 1 and 2 were negative was because the baseline value for the losses due to JD was lower than losses in the subsequent years of the study. Thus, mathematically, it appeared the control program had “negative” benefits that never exceeded zero when summed over the projected study period. The NPV’s for scenarios 3 and 4 remain negative over the entire 20-year projected study period because, in the absence of a control program, there was nothing to offset the losses caused by JD. Results fl sensitivig analysis The break-even cost for the JD control program is an estimate of the amount of money a farm can invest in the control program and still “break-even”, or have a NPV equal to zero. The break-even costs were -$ 1 .81 and -$30.27 for scenarios 1 and 2 respectively. Thus, based on the assumptions used to calculate the NPV’s in this study, the JD control program on this farm would not break-even, no matter how little money was invested in it. 174 Table 4.10: NPV of four scenarios for Johne’s disease (JD) control on Herd 2 Year of Scenario“ Program 1 2 3 4 5 -$144.54 -$147.35 -$175.81 -$128.81 10 -$236.11 -$282. 15 -$351.65 -$216.47 20 —$209.42 -$559.51 -$928.57 -$396.85 Payback Year N/A N/A N/A N/A N/A Not applicable Payback Year: Year of control program when NPV became positive * Scenario 1: Assuming linear decline in losses caused by JD after year 4 due to declining prevalence and eventual eradication after implementation of JD control program Scenario 2: Assuming losses caused by JD remain constant after year 4 while still investing in JD control program Scenario 3: Assuming losses caused by JD increase at a rate equal to the decline in Scenario 1 in the absence of JD control program Scenario 4: Assuming losses caused by JD remain constant at baseline level in absence of JD control program The NPV was most sensitive to the discount rate (r) used. After the discount rate, milk price, followed by the cost of the JD control program had the greatest effect on the NPV. When the individual categories within the JD control program were evaluated, the results were ranked depending on which category represented the highest proportion of the control program For this herd, changing the input costs of labor had the greatest impact on NPV, followed by supplies, management, and capital investments respectfully. Including testing in the calculations increased the input cost of the JD control program, and hence decreased the NPV. For scenario 1, the NPV decreased from -$209 175 to -$283 when ELISA testing only was included, and to -$554 when both ELISA and fecal culture was included. For scenario 2, the NPV decreased from -$550 to -$633 and -$904 respectfully, for including ELISA testing only or ELISA and fecal culture together. Producer perception of the JD control program At the conclusion of this study, the producer stated he was satisfied with the JD control program and planned to continue investing in it. He noted he had not necessarily seen any increased revenue as a result of the control program, but hoped that would change as herd health and production improved. C. Herd 3 Farm background This is a third generation farm in the very early stages of transitioning to the next generation. At the beginning of the study, the herd was milking approximately 190 cows (218 total adult cows) with a RHA of 21,865 pounds (3.5% FCM). The long term goals for this farm were: 1. Expand herd size to 450 cows milking (~500 cows total) 2. Transition ownership and management to next generation 176 In order to meet these goals, the farm built a new free stall barn and remodeled existing facilities in 2003. The herd was open. In preparation for the herd expansion, three smaller herds, including all young stock (approximately 300 cattle total), were purchased and consolidated in 2004. Additionally, six to eight bulls are purchased each year for breeding purposes, and are not screened for JD. Johne’s disease was first diagnosed in the herd in 2001, in a 3.5 year old cow raised on the farm. In 2002, approximately 12 cows (5%) were culled exhibiting clinical signs of JD. Aside from JD, the herd was experiencing other “expansion pains” including periparturient metabolic problems and hairy heal warts. Labor management was also a concern as the herd transitioned from primarily family labor to hired Hispanic labor. JD risk assessment Prior to the implementation of the JD control program, the areas at greatest risk for disease transmission on this farm were the maternity and pre-weaned calf areas. The close-up dry cows were housed on a manure pack that also served as the maternity pen. Cows suspected of having JD were not segregated in any way. Manure build up in the maternity area was occasionally an issue, and cows in the area were moderately dirty. Calves were generally removed from dams within 4-6 hours of birth. Calves were fed pooled colostrum, then unpasteurized whole milk until weaned. Weaned calves, until five months of age, were housed adjacent to the adult cow area and fed feed refusal from those cows. 177 JD control plan Upon enrolling in the Michigan Johne’s Disease Control Demonstration Project, annual testing of all adult cows for JD with serum ELISA and fecal culture began. All cows testing positive on either test are visually identified with red cable ties placed through their ear tags, as were as any calves born to these cows. Cows testing positive for JD ware kept in the herd until they developed clinical signs or their production fell below some breakeven point set by the farm. As all breeding was done by natural service, test positive cows were often bred back. An effort was made to improve the sanitation of the maternity pen by more frequent cleaning and/or bedding. Calves were removed as soon as possible after birth, generally within one hour. Feeding pooled colostrum was discontinued. Calves were fed colostrum only from test negative cows. The farm also switched from unpasteurized whole milk to milk replacer to feed calves. As a result of the herd expansion, 3 heifer grower was contracted to raise heifers from the age of six months until they are returned to the farm as springing heifers to freshen. The heifer grower only raised heifers for this farm. A new bam was built for pre-weaned calves, and an existing barn remodeled for calves from the time of weaning until they were sent to the heifer grower. This removed calves from direct contact with adult cows, but they were still occasionally fed feed refusal from the adult herd. Descrgrtive statistics Descriptive statistics for herd 3 are summarized in Table 4.1 1. 178 Table 4.11: Descriptive Statistics for Herd 3 Year Herd size RHA Cull Rate Culled due to (adult cows) (lbs 3.5% FCM) (%) clinical JD (%) 2003 218 21,865 21.6 3.2 2004 369 25,643 24.1 3 .0 2005 412 26,028 27.7 1.2 2006 432 26,329 32.9 3.7 2007 458 26,210 29.7 2.6 Herd size more than doubled over the study period due to the purchase of cattle as already discussed. No additional replacement cattle have been purchased since 2004. As of 2007, the plan was for the herd to expand internally to reach the desired herd size of 500 adult cows. As a result of trying to increase herd size, the overall cull rate was relatively low throughout the study period. To cash flow the new free stall barn, cows were needed in every stall. Cows that would have been culled previous to the expansion were kept to fill stalls. Once the facility reaches capacity, it is likely the cull rate will go up as space becomes a limiting factor, and cows will need to be culled to make room for more productive heifers. The number of cows culled due to clinical JD has been fairly consistent throughout the study period. As of the conclusion of the study, there had not been much culling pressure on JD test positive cows. Test positive cows were managed only in so far as to prevent disease transmission to calves through colostrum or milk. Otherwise, they were managed as any other cow in the herd. The mortality rates are not reported due to insufficient data. 179 JD prevalence The within herd JD prevalence is outlined in Table 4.12. Johne’s disease prevalence remained relatively unchanged over the course of this study. This might lead one to believe that the JD control program implemented on this farm was ineffective in preventing JD transmission. However, it must be remembered that this herd doubled in size, mainly through the purchase of cattle, including young stock, from herds of unknown JD status. As late as 2006, heifers were still entering the herd that had not been exposed to the JD control program implemented by this herd. In 2007, it was estimated over half the milking herd consisted of purchased cattle. Unfortunately, insufficient records prevented differentiating the JD prevalence in purchased cows vs. cows raised on the farm after implementation of the JD control program Table 4.12: Johne’s disease prevalence trends 2003-2007 for Herd 3 Year Apparent JD prevalence (ELISA &/or F C positive) 2003 10.5% 2004 7.7% 2005 12.9% 2006 11.8% 2007 16.0% Cost of JD control program 2003-2007 The costs of the JD control program observed over the five years of this study are summarized in Table 4.13. As part of the Michigan Johne’s Disease Control Demonstration Project, the herd did not have to pay for any JD testing beyond the cost of 180 labor to collect samples on the day of the annual test. It is unlikely the herd would have done the extensive testing that was performed if it had to pay for the testing, yet management decisions were made based on those test results, so for the sake of completeness, testing costs are included in the last column of the table. Table 4.13: Cost of Johne’s disease control program 2003-2007 for Herd 3 ($/cow) Year No. Supplies Management Labor Capltal Total Total Cows Investments (+ testrng) 2003 218 $1.76 $21.83 $34.02 $0.62 $58.22 $87.22 2004 369 -$5.39 $16.34 $30.15 $0.36 $41.46 $70.46 2005 412 -$2.66 $15.00 $27.01 $0.33 $39.68 $68.68 2006 432 -$0.60 $14.88 $27.19 $0.31 $41.77 $70.77 2007 458 -$8.79 $14.41 $25.64 $0.29 $31.55 $60.55 Ave. 378 -$3. 14 $16.49 $28.80 $0.38 $42.54 $71.54 The supplies category included the purchase of milk replacer to feed calves and colostrum supplement, as well as cable ties to identify JD test positive cows. It also included increased cost for straw due to more frequent cleaning and bedding of the maternity pen. The cost of supplies was also adjusted to reflect the sale of milk that would have previously been fed to calves. This adjustment explains the negative supply costs for 2004-2007. Over the course of this study, the farm was in the process of transitioning from family labor to hired labor. The herd manager was still doing many things that labor would do on other farms. Management costs for this herd reflect time spent on JD testing, record keeping, making management decisions regarding JD test positive cows, buying and selling cows, and capital purchases. It also includes time spent 181 handling cattle and colostrum, making sure that only colostrum from test negative cows is fed to calves. Labor costs are due primarily to increased time spent cleaning the maternity pen, mixing milk replacer, and flesh cow and calf care. The capital investment made by this farm was a mixing vat to mix large amounts of milk replacer purchased in the surmner of 2003. Economic losses due to JD 2003-2007 The annual estimated economic losses due to JD, along with the calculated, or assumed, benefits of the control program for Herd 3 are summarized in Table 4.14. Table 4.14: Economic losses due to Johne’s disease and assumed benefits of Johne’s disease control program for Herd 3 — 2003-2007 ($/cow) Year NO. Mllk RPO Cull Value Total Assumed Cows Value Benefits 2003 218 $10.06 $0.87 $6.94 $17.86 N/A 2004 369 $16.82 $0.00 $3.72 $20.54 -3134 2005 412 $44.59 $8.50 $2.66 $55.75 $36.55 2006 432 $97.41 $15.80 $5.08 $118.29 $99.09 2007 458 $53.13 $8.67 $4.74 $66.54 $47.34 Ave. 378 $44.40 $6.77 $4.63 $55.80 N/A N/A Not applicable 182 NPV calcglgtion The NPV was calculated for four different scenarios: (1) assuming a linear decline in losses following year five and disease eradication by year 20 with the control program; (2) assuming losses and JD prevalence remain constant at year five levels while continuing the control program; (3) assuming linear increase in losses at rate equal to that in scenario 1 with no control program; and (4) assuming losses remain constant at same level as the beginning of the study with no control plan implemented. The results are shown in Table 4.15. The reason the NPV’s for both scenarios 1 and 2 were negative was because the baseline value for the losses due to JD was lower than losses in the subsequent years of the study. Thus, mathematically, it appeared the control program had “negative” benefits that never exceeded zero when summed over the projected study period. The NPV’s for scenarios 3 and 4 remain negative over the entire 20-year projected study period because, in the absence of a control program, there was nothing to offset the losses caused by JD. Results Qfsensitivitv analysis The break-even cost for the JD control program is an estimate of the amount of Money a farm can invest in the control program and still “break-even”, or have a NPV equal to zero. The break-even costs were -$l7.85 and -$46.46 for scenarios 1 and 2 respectively. Thus, based on the assumptions used to calculate the NPV’s in this study, the JD control program on this farm would not break-even, no matter how little money was invested in it. 183 Table 4.15: NPV of four scenarios for Johne’s disease (JD) control on Herd 3 Year of Scenario* Program 1 2 3 4 5 -$333. 12 -$335.95 -$123.92 -$76.66 10 -$533.87 -$580.17 -$264.77 -$128.83 20 -$730.64 -$1,082.68 -$770.86 -$236. 19 Payback Year N/A N/A N/A N/A N/A Not applicable Payback Year: Year of control program when NPV became positive * Scenario 1: Assuming linear decline in losses caused by JD after year 4 due to declining prevalence and eventual eradication after implementation of JD control program Scenario 2: Assuming losses caused by JD remain constant after year 4 while still investing in JD control program Scenario 3: Assuming losses caused by JD increase at a rate equal to the decline in Scenario 1 in the absence of JD control program Scenario 4: Assuming losses caused by JD remain constant at baseline level in absence of JD control program The NPV was most sensitive to the discount rate (r) used. After the discount rate, the cost of the JD control program, followed by milk price had the greatest effect on the NPV. When the individual categories within the JD control program were evaluated, the results were ranked depending on which category represented the highest proportion of the control program For this herd, changing the input costs of labor had the greatest impact on NPV, followed by management, capital investments, and supplies respectfully. Including testing in the calculations increased the input costs of the JD control program, and hence decreased the NPV. For scenario 1, the NPV decreased from -$731 to -$864 when ELISA testing only was included, and to -$1075 when both ELISA and fecal culture was included. For scenario 2, the NPV decreased from -$1083 to -$1156 184 and -$1427 respectfully, for including ELISA testing only or ELISA and fecal culture together. ProduceLperception of the JD control program At the conclusion of this study, the producer was satisfied with the JD control program and planned to continue investing in it. Subjectively, herd health had improved over the course of this study as had production which led to increased revenues. D. Herd 4 F arm background This is a rotational grazing Jersey herd, although the cows are confined to free stalls during the winter months. The owners sold the herd and got out of the dairy business in 1994. In 1995, they changed their minds and reassembled a herd with the purchase of 17 cows from multiple sources. The last outside cattle added to the herd occurred in 1997 with the purchase of ten cows and two springing heifers from a single herd in North Carolina. In 2003, they were milking approximately 70 cows (75 total adult cows) with a RHA of 12,446 pounds, and were in the final stages of becoming a certified organic dairy farm. The herd’s goals were as follows: 185 1. Become a certified organic dairy farm 2. Decrease herd size to milk ~50 cows 3. Maintain a low input operation (rotational grazing) 4. Sell dairy replacements The herd gained organic certification in the spring of 2005, at which time they contracted the price they receive for milk at $34/cwt; more than double the price the other herds in this study were receiving. J ohne’s disease was first diagnosed in the herd in the winter of 2002, in a three- year old cow that was raised on the farm, but whose dam was purchased. A total of four cows were culled in 2002, two of which were believed to have JD based on clinical signs. JD risk assessment Prior to the implementation of a JD control program, this herd was at high risk for disease transmission in almost every area on the farm. The maternity pen doubled as the sick cow pen. It was a manure pack that was cleaned infrequently. Calves were often left to nurse the dam, or surrogate dam, for one week, up to one month. Otherwise calves, were fed unpastuerized whole milk, and housed in a pen adjacent to the maternity pen with direct contact with adult cows. After weaning, heifers were housed in a pen adjacent to the barnyard where they had nose-to-nose contact with adult cows. Bred heifers were housed with the adult herd two months prior to calving, and grazed with the lactating herd during the summer months. Finally, the same loader bucket was used for feed and manure handling. 186 JD control plcm Upon enrolling in the Michigan Johne’s Disease Control Demonstration Project, annual testing of all adult cows for JD with serum ELISA and fecal culture began. Cows positive on fecal culture and exhibiting clinical signs were culled immediately. Cows testing positive on fecal culture and/or ELISA and not showing clinical signs were not rebred, and were culled as soon as they began showing clinical signs, or their milk production fell below a break-even point defined by the farm. The most immediate management changes focused on the maternity and calf areas. The maternity area was cleaned weekly with lime put down under fresh straw bedding. Calves were removed as soon as possible from the dams, generally within two hours. An existing barn was renovated for pre-weaned calves to remove them fiom contact with adult cows. Colostrum was not pooled, and only colostrum fi'om JD test negative cows was fed to calves and frozen to be used as needed. Instead of whole milk, calves were fed milk replacer until weaned. Weaned heifers were still housed next to the barn yard, but the feeding area was moved to minimize the potential for feed contamination by manure fi'om the lactating herd. Bred heifers were still grazed with lactating cows during the summer, and springing heifers were housed with the lactating herd two months prior to calving. In 2005, a front-end loading tractor was purchased, mainly for the purpose of the JD control program. This, in addition to their old tractor, allowed one tractor to be used exclusively for handling feed and the other for handling manure. 187 Descriptive stgtistics Descriptive statistics for Herd 4 are summarized in Table 4.16. Aside from becoming a certified organic dairy farm in 2005, this herd is unique from the other herds in this study for several reasons; but the main one is the personal attachment the owner has with the cows. The owner recognizes each cow by name, without the aid of any other identification such as ear tags or neck chains. Culling decisions are difficult for the owner, and sentiment plays a much larger role than on other farms. This is evident in the relatively low cull rates up until 2006. In 2006, a concentrated effort was made to decrease herd size, and the majority of cows culled were sold to other dairy farms rather than to slaughter. lrregardless of sentiment, cows with a positive JD test (ELISA and/or fecal culture) were culled as soon as possible. As within herd JD prevalence declined, so did the number of cows being culled for JD. The mortality rate remained low throughout the study. The increase in 2006 was due to a total of three cows dying for various reasons, as compared to only one cow in each of the other study years. Table 4.16: Descriptive Statistics for Herd 4 Year (Hdeurld size (13% Cull Rate CIIlled due to Mortality Rate a t cows) FCM) (%) clrmcal JD (%) (%) 2003 75 12,446 28.0 9.3 0.0 2004 74 12,149 28.4 12.2 1.4 2005 77 12,307 24.7 1.3 1.3 2006 72 12,578 50.0 2.8 4.2 2007 68 13,429 44.0 2.9 1.5 188 JD prevalence The within herd JD prevalence trend is outlined in Table 4.17. Prevalence decreased steadily until 2005, and then rebounded in 2006 and 2007. This has been somewhat disconcerting for the owner, as the number of cows exhibiting clinical signs has declined. In fact, in 2007, there were no cows reported with clinical signs. Table 4.17: Johne’s disease prevalence trends 2003-2007 for Herd 4 Year Apparent JD prevalence (ELISA &/or PC positive) 2003 11.7% 2004 9.0% 2005 3.8% 2006 7.9% 2007 6.6% Cost of the JD control program 2003-2007 The costs of the JD control program observed over the five years of this study are summarized in Table 4.18. As part of the Michigan Johne’s Disease Control Demonstration Project, the herd did not have to pay for any JD testing beyond the cost of labor to collect samples on the day of the annual test. It is unlikely the herd would have done the extensive testing that was performed if it had to pay for the testing, yet management decisions were made based on those test results, so for the sake of completeness, testing costs are included in the last column of the table. The supplies category included costs associated with switching to milk replacer to feed calves and the purchase of colostrum replacer to be used when real colostrum from 189 JD test negative cows was unavailable. The reason for the relatively large “negative” supply costs 2005-2007 was because of the adjustment made for the sale of milk that was previously fed to calves. As a certified organic dairy farm, the cost of the milk replacer the farm was allowed to use tended to be higher than that used by the other herds in this study. However, upon becoming certified, the price the farm was receiving for their milk more than compensated for the switch to milk replacer. Table 4.18: Cost of Johne’s disease control program 2003-2007 for Herd 4 ($/cow) No. . Capital Total Year Supplies Management Labor Total _ Cows Investments (+ testrng) 2003 75 $7.40 $18.00 $17.78 $0.00 $43.18 $72.18 2004 74 $2.18 $30.01 $25.14 $0.00 $57.33 $86.33 2005 77 -$41.44 $29.57 $24.16 $55.70 $67.99 $96.99 2006 72 -$39.85 $32.99 $25.83 $56.59 $75.56 $104.56 2007 68 -$38.00 $35.87 $26.65 $56.55 $81.07 $110.07 Ave. 73 -$21.94 $29.29 $23.91 $33.77 $65.02 $94.02 The management costs for this herd seemed extremely high when compared to the other, although larger, herds in the study. Initially, the producer reported spending a half hour each day managing JD and valued this time at $50 per hour. Even if this time was limited to weekdays, that equates to $85 per cow per year. As that seemed like an unrealistic number, further conference with the producer resulted in the above estimate, which was adjusted to reflect ten minutes per weekday spent managing the JD control program. While this still seems very high in comparison to the other herds, only the 190 producer knows what is happening on the farm. This demonstrates one of the issues with the method employed to determine the costs of the JD control programs in this study. The questionnaire used to collect information, particularly in regards to the management and labor sections, was dependent upon the producer recalling their daily routine and allotting time to a particular enterprise. The labor category reflects costs associated with annual herd testing and additional fresh cow and calf care. Again, it is subject to the producer’s recall and allotment of time to the JD control program. The capital investment made by this farm was the purchase of a new loader tractor in January 2005. This purchase was financed, and the costs reflect the annuity value for the tractor as well as the annual interest paid. While the producer stated the tractor was purchased as a direct result of the JD control program; it should be noted that, from a JD control standpoint, the farm could have achieved the same outcome by purchasing a second bucket for their existing tractor at a much lower cost. Economic losses fidue to JD 2003-2007 The annual estimated economic losses due to JD, along with the calculated, or assumed, benefits of the control program for Herd 4 are summarized in Table 4.19. The loss due to subOptimal culling (RPO) could not be calculated for this herd because individual cow production data was not available. Thus, the total losses due to JD are underestimated for this herd. 191 Table 4.19: Economic losses due to Johne’s disease and assumed benefits of Johne’s disease control program for Herd 4 — 2003-2007 ($/cow) Year No' Milk RPO Cull Value Total Assumed Cows Value Benefits 2003 75 $58.94 $0.00 $1 1.03 $69.97 N/A 2004 74 $68.87 $0.00 $8.31 $77.18 -$3.60 2005 77 $32.06 $0.00 $0.84 $32.90 $40.67 2006 72 $70.09 $0.00 $7.62 $77.71 -$4. 14 2007 68 $56.40 $0.00 $3.53 $59.93 $13.64 Ave. 73 $57.27 $0.00 $6.27 $63.54 N/A N/A Not applicable NP V calculation The NPV was calculated for four different scenarios: (1) assuming a linear decline in losses following year five and disease eradication by year 20 with the control program; (2) assuming losses and JD prevalence remain constant at year five levels while continuing the control program; (3) assuming linear increase in losses at rate equal to that in scenario 1 with no control program; and (4) assuming losses remain constant at same level as the beginning of the study with no control plan implemented. The results are shown in Table 4.20. The reason the NPV’s for both scenarios 1 and 2 were negative was because, on average, the costs of the control program exceeded the potential benefits. Thus, mathematically, it appeared the control program had “negative” benefits that never exceeded zero when summed over the projected study period. However, when interpreting these figures, it should be remembered that the economic losses, and hence, 192 Table 4.20: NPV of four scenarios for J ohne’s disease (JD) control on Herd 4 Year of Scenario“ Program 1 2 3 4 5 -$225.08 -$227.63 -409.90 -$367.33 10 -$325.55 -$367.25 -$689.69 -$567.26 20 -$337.46 -$654.53 -$1,460.20 -$978.62 Payback Year N/A N/A N/A N/A N/A Not applicable Payback Year: Year of control program when NPV became positive * Scenario 1: Assuming linear decline in losses caused by JD after year 4 due to declining prevalence and eventual eradication after implementation of JD control program Scenario 2: Assuming losses caused by JD remain constant afier year 4 while still investing in JD control program Scenario 3: Assuming losses caused by JD increase at a rate equal to the decline in Scenario 1 in the absence of JD control program Scenario 4: Assuming losses caused by JD remain constant at baseline level in absence of JD control program the potential benefits, were underestimated because the RPO could not be calculated Also, the JD control program was footing the entire cost of the new loading tractor. The tractor was likely being used for other enterprises on the farm, even though the producer stated the tractor would not have been purchased if not for the JD control program. As a result, the costs of the JD control program for this herd were robust. A more conservative estimate of costs could have been achieved by assigning the cost of the bucket for the tractor to the JD control program, and the remaining cost of the tractor to some other farm enterprise. The net result of these issues on the benefits and costs was to lower the NPV. The NPV’s for scenarios 3 and 4 remain negative over the entire 20-year 193 projected study period because, in the absence of a control program, there was nothing to offset the losses caused by JD. Results ofsensitivitv angbgsg The break-even cost for the JD control program is an estimate of the amount of Money a farm can invest in the control program and still “break-even”, or have a NPV equal to zero. The break-even costs were $38.86 for scenario 1 and $13.08 for scenario 2. Thus, had this herd been able to cut the cost of the JD control program by 40% and 80% respectively for scenarios 1 and 2, the NPV would have equaled zero; and cutting costs even more would have resulted in a positive NPV. The NPV was most sensitive to the discount rate (I) used. After the discount rate, the cost of the JD control program, followed by milk price had the greatest effect on the NPV. When the individual categories within the JD control program were evaluated, the results were ranked depending on which category represented the highest proportion of the control program. For this herd, changing the input costs of capital investment had the greatest impact on NPV, followed by management, labor, and supplies respectfully. Including testing in the calculations increased the input costs of the JD control program, and hence decreased the NPV. For scenario 1, the NPV decreased from -$337 to -$411 when ELISA testing only was included, and to -$682 when both ELISA and fecal culture was included. For scenario 2, the NPV decreased from -$655 to -$728 and -$999 respectfully, for including ELISA testing only or ELISA and fecal culture together. 194 Producer perception of the JD control program Overall, the producer was pleased with the results of the JD control program, and planned to continue investing in it after the end of the study. Subjectively, herd health improved over the course of the study which resulted in increased production and increased revenues. An additional value resulting from the JD control program on this herd was seen in the marketing of herd replacements. Other producers were willing to pay more for cattle raised on this farm because of the JD control program the herd had implemented and the amount of diagnostic testing that was performed. E. Herd 5 Farm background This farm has been in existence at the present location for approximately 50 years. In 2004, when the herd enrolled in the Michigan J ohne’s Disease Control Demonstration Project, it was milking approximately 440 Holstein cows (484 total adult cows) with a RHA of 26,839 pounds. The farm was expanding with the following goals: 1. Milk 600 cows by 2006 2. Milk 1200 cows within 10 years 3. Build new facilities, including a new dry cow and maternity barn 195 As part of the expansion, the herd was purchasing cattle on a routine basis. Often entire herds were purchased, although they also contracted with a cattle broker to purchase cattle in their stead. As of 2004, approximately 25% of the adult herd had been purchased. The JD status of the cows or herds that were purchased was not considered. The first case of clinical JD was diagnosed in this herd in a two year-old purchased cow in 1999. Since that time, the number of cows diagnosed and being culled for JD had increased; and by 2004, JD had become a concern for the producer. Aside fiom JD, the herd was experiencing the normal “expansion” woes, but no one particular problem seemed to stand out. Bulk tank SCC averaged around 200,000, and most of the mastitis problems were environmental in nature. When the herd expansion began around 2000, calves were moved off-site to a heifer grower’s facility, where they are raised and bred This grower raises heifers only for this farm. The heifers are returned to the home farm when they were 6-7 months pregnant. JD risk assessment Prior to the implementation of a JD control program, the area at greatest risk for disease transmission on this farm was the calving area. The farm had a group maternity pen that housed approximately 10-20 cows at all times. It was cleaned every 2-3 months, but was bedded “as needed,” and the cows were not always clean. As it was a bedded pack near the parlor, sick cows with poor mobility were often kept in this area. There was no attempt to segregate cows suspected of having JD, and all cows calved in this pen. Calves were often left with the dam in the group pen for 12-24 hours, and allowed to nurse, before being moved to a remote calf barn with individual pens. The calves were 196 fed pooled colostrum and unpasteurized waste or whole milk In January 2004, a pasteurizer was purchased, and calves were then fed pasteurized milk. Upon weaning, all calves were moved off-site to the heifer grower’s facility, and returned to the home farm when they were 6-7 months pregnant. The springing heifers were housed in a barn separate from the lactating herd, but fed feed refusal from the adult cows. JD control program Upon enrolling in the Michigan Johne’s Disease Control Demonstration Project, annual testing of all adult cows for JD with serum ELISA began. All cows testing positive were flagged in the computer for management. Cows with a positive ELISA had “one strike” against them in terms of culling. However, these cows were often bred back, and kept in the herd until they developed clinical signs of JD, some other problem warranting culling, or their production fell below some break-even point determined by the farm. The use of pooled colostnnn was discontinued. Only colostrum from test negative cows was fed to heifer calves and banked for use as needed. A comer in the maternity pen was gated off to serve as a holding area for newborn calves until they could be moved to the calf barn. In 2007, a new maternity barn was built with individual calving pens. Cows were kept in the pens for the minimum time necessary, and one pen is still reserved as a holding area for calves. Calves were moved to this holding area as soon as possible after birth. This was achieved by offering employees a monetary incentive for moving the calves. This worked well, as the maternity area was located adjacent to the milking parlor, with employees passing by several times a day as they 197 moved cows for milking. Calves were still fed whole milk until weaning, but beginning in January 2004, all milk fed to calves was pasteurized. Descriptive statistics Descriptive statistics for Herd 5 are summarized in Table 4.21. Herd size increased by approximately one-third over the course of the study. This was accomplished mainly through the purchase of cattle. Herd size was fairly stable for the first three years of the study and milk production gradually increased In 2007, a bunch of purchased cattle were added to the herd, and the result was a 5% production decrease. The cull data was incomplete for this herd, and the overall cull rate and mortality rate was not reported. The number of cows culled due to clinical signs of JD was reported, and decreased over the course of the study; although it remained consistently low. As long as cattle continue to be purchased and added to the herd with no regard to their JD status, it is likely this number will remain fairly consistent, or even increase in the future. Table 4.21: Descriptive Statistics for Herd 5 Herd size RHA Culled due to Year _ . (adult cows) (lbs 3.5% FCM) Cllnlcal JD (%) 2004 484 26,839 2.5% 2005 500 28,470 2.6% 2006 497 29,668 0.6% 2007 641 28,232 0.5% 198 JD prevalence The within herd JD prevalence trend, as determined by serum ELISA, is outlined in Table 4.22. For all practical purposes, the JD ELISA prevalence remained unchanged in this herd over the course of the study. This was not unexpected, given that a significant proportion of the herd was purchased fi'om multiple sources with unknown JD status. In fact, the proportion of test positive cows that were purchased ranged from 26- 50% per year. As long as cattle from herds with unknown JD status continue to be added to the herd, the best this herd can hope for is to maintain the JD prevalence at the current level. Table 4.22: Johne’s disease prevalence trends 2003-2007 for Herd 5 Year Apparent JD prevalence (serum ELISA) 2003 NT 2004 5.6% 2005 5.4% 2006 6.4% 2007 4.0% NT — Not Tested Cost of the JD control program 2004-2007 The costs of the JD control program observed over the five years of this study are summarized in Table 4.23. As part of the Michigan Johne’s Disease Control Demonstration Project, the herd did not have to pay for any JD testing beyond the cost of labor to collect samples on the day of the annual test. It is unlikely the herd would have 199 Table 4.23: Cost of Johne’s disease control program 2004-2007 for Herd 5 ($/cow) No. Capital Total Year Supplies Management Labor Total _ Cows Investments (+ testrng) 2004 484 $6.1 1 $1.36 $4.73 $2.65 $14.85 $43.85 2005 500 $5.91 $0.74 $6.00 $2.56 $15.22 $44.22 2006 497 $5.95 $0.39 $6.04 $2.58 $14.95 $43.95 2007 641 $5.79 $0.31 $4.78 $2.00 $12.87 $41.87 Ave. 531 $5.94 $0.70 $5.39 $2.45 $14.47 $43 .47 done the extensive testing that was performed if it had to pay for the testing, yet management decisions were made based on those test results, so for the sake of completeness, testing costs are included in the last column of the table. The supplies category included the increased costs of operating the pasteurizer, such as electricity and sanitation supplies. Management included time the herd manager spent aiding in annual JD testing, inputting test results into the computer, and making buying and selling decisions. Labor included time spent aiding in JD testing, removing calves from dams as soon as possible after birth, pasteurizing milk, and sanitizing the pasteurizer. The capital investments made by this farm included the pasteurizer and a bulk tank to hold the milk after pasteurization. Economic losses due to JD 2004-2007 The annual estimated economic losses due to JD, along with the calculated, or assumed, benefits of the control program for Herd 5 are summarized in Table 4.24. 200 Table 4.24: Economic losses due to Johne’s disease and assumed benefits of Johne’s disease control program for Herd 5 — 2004-2007 ($/cow) Year NO' Mllk RPO Cull Value Total Assumed Cows Value Benefits 2004 484 $23.61 $1.98 $3.99 $29.58 N/A 2005 500 $41.54 $2.82 $3.55 $47.91 -39. 16 2006 497 $27.48 $6.07 $2.00 $35.55 $3.20 2007 641 $17.51 $1.02 $0.71 $19.24 $19.51 Ave. 531 $27.54 $2.97 $2.56 $33.07 N/ A N/A Not applicable NP V calculation The NPV was calculated for four different scenarios: (1) assuming a linear decline in losses following year five and disease eradication by year 20 with the control program; (2) assuming losses and JD prevalence remain constant at year five levels while continuing the control program; (3) assuming linear increase in losses at rate equal to that in scenario 1 with no control program; and (4) assuming losses remain constant at same level as the beginning of the study with no control plan implemented. The results are shown in Table 4.25. The NPV for scenario I became positive in year nine, and that for scenario 2 became positive in year 14. The NPV’s for scenarios 3 and 4 remain negative over the entire 20-year projected study period because, in the absence of a control program, there was nothing to offset the losses caused by JD. 201 Table 4.25: NPV of four scenarios for J ohne’s disease (JD) control on Herd 5 Year of Scenario“ Program 1 2 3 4 5 -$17.89 -$20.26 -$l67.57 -$154.71 10 $10.70 -$6.57 -$296.98 -$259.99 20 $128.39 $21.59 -$622. 13 -$476.64 Payback Year 9 14 N/A N/A N/A Not applicable Payback Year: Year of control program when NPV became positive * Scenario 1: Assuming linear decline in losses caused by JD after year 4 due to declining prevalence and eventual eradication after implementation of JD control program Scenario 2: Assuming losses caused by JD remain constant after year 4 while still investing in JD control program Scenario 3: Assuming losses caused by JD increase at a rate equal to the decline in Scenario 1 in the absence of JD control program Scenario 4: Assuming losses caused by JD remain constant at baseline level in absence of JD control program Results of sensitivity analvsis The break-even cost for the JD control program is an estimate of the amount of money a farm can invest in the control program and still “break-even”, or have a NPV equal to zero. The break-even cost calculated for scenario 1 was $24.90/cow/year, and for scenario 2, $16.21/cow/year. The NPV was most sensitive to the discount rate (r) used. After the discount rate, milk price, followed by cost of the JD control program had the greatest effect on the NPV. When the individual categories within the JD control program were evaluated, the results were ranked depending on which category represented the highest proportion of 202 the control program. For this herd, changing the input costs of supplies had the greatest impact on the NPV, followed by labor, capital investments, and management respectfully. Including testing in the calculations increased the input costs of the JD control program, and hence decreased the NPV. For scenario 1, the NPV decreased fi'om $128 to $55 when ELISA testing only was included, and to -$216 when both ELISA and fecal culture was included For scenario 2, the NPV decreased from $22 to -$52 and -$323 respectfully, for including ELISA testing only or ELISA and fecal culture together. Producetmerception of the JD control program At the conclusion of this study, the producer was satisfied with the JD control program and planned to continue investing in it. Subj ectively, herd health had improved over the course of this study, as had production, which led to increased revenues. F. Herd 6 Farm background This herd has been in existence at the current location since the early 1970’s. In 2003, the herd was milking around 130 adult Holstein cows (145 total cows) with a RHA of 26,875 pounds (3.5% FCM). The herd had been completely closed for 30 years. The goals for this farm were: 203 1. Continue dairy farming 2. Transfer farm to next generation 3. Maintain a 28,000 pound RHA 4. Minimize JD in the herd 5. Expand An ongoing threat to this farm was bovine tuberculosis (TB). This farm was located in the middle of the bovine TB zone in the northeastern lower peninsula of Michigan. Bovine TB had been diagnosed on an adjacent farm, and that herd had been depopulated twice since 1995. On this farm, a son had come of age, and wanted to continue dairying, but major improvements to, or replacement of, the existing facilities was necessary. While the farm was doing everything it could to protect the herd fi'om becoming infected with TB; knowing it was nearby, and the uncertainty over the state’s long term plan for eradicating or controlling TB (depopulation vs. test and cull), made the producer reluctant to invest a large amount of money into new facilities. So, over the course of this study, expanding was put on hold, and the existing facilities were repaired and remodeled to meet the herd’s needs. The first case of clinical JD was diagnosed in the herd in 2002, in a second lactation cow. In the year following, a total of six (4%) cows were culled due to clinical signs of JD. Given that both bovine TB and JD are caused by Mycobacterium, and there is potential for cross-reactivity between screening tests for the two diseases, controlling JD became a high priority for this herd 204 Aside from JD, this herd reported occasional problems with metabolic diseases such as displaced abomasums, ketosis, fatty livers, and acidosis. Mastitis was generally not a problem, and the bulk tank SCC ranged between 200,000-300,000. JD risk assessment In terms of controlling JD, this herd was doing everything wrong. Cows were calved on a bedded pack in a group maternity pen. Manure often built up before fi'esh bedding was added, resulting in soiled udders and legs. All cows calved in this pen, regardless of their JD status, or suspected status. Also, sick and/or treated cows were housed in this pen, and milked in adjacent stanchions with a bucket milker until their milk was good to go into the bulk tank. Then they were moved back to the main herd. Once a cow calved she was moved with her calf to an individual pen where they would stay for 3-5 days until the cow’s milk was okay to go into the bulk tank During this time the calf was allowed to nurse the cow. Colostrum and waste milk was pooled and fed to all calves in the maternity barn. Once the darn was moved to the main herd, the calf remained in the maternity barn as long as there was waste milk available to feed it. Once waste milk was no longer available, the calf was moved to an individual pen in a calf barn and switched to milk replacer. Once calves were weaned they were housed in a super hutch or group pen for approximately 1-2 months before being moved to a heifer barn that housed all replacements and the far-off dry cow group. Regardless of group, a common skid-steer bucket was used for feed and manure handling. 205 JD control program Upon enrolling in the Michigan Johne’s Disease Control Demonstration Project, annual testing of all adult cows for JD with serum ELISA and fecal culture began All cows testing positive were visually identified with colored ear tags and/or neck chains. Cows positive on fecal culture and exhibiting clinical signs were culled immediately. Fecal culture positive cows not showing clinical signs were not bred back, and were culled when they either began to exhibit clinical signs, or their milk production fell to below a break-even point defined by the farm Cows positive on serum ELISA had one strike against them, and were evaluated on a case by case basis. In the winter of 2003, the maternity barn was remodeled. Sand-bedded free stalls took the place of the bedded pack to house the close-up dry cow group. Adjacent to the free-stalls were individual calving pens. Cows were moved to individual pens when they began to calve. An attempt was made to calve all JD test positive cows in a pen separate from where JD test negative cows calved. Pooled colostrum was no longer fed Only colostrum from JD test negative cows was fed to heifer calves and frozen for use as needed. Heifer calves were removed fi'om JD test positive dams as soon as possible. Otherwise, bull calves or calves born to JD test negative cows were allowed to stay with the dam until her milk was saleable and she was moved to the main parlor. Despite recommendations against this practice, it was the producer’s belief that fresh cows transitioned better if they were allowed to stay with their calf for a couple of days. In 2004, a second skid-steer was purchased, allowing one to be dedicated to feed handling and one to manure handling. 206 Descriptive statistics Descriptive statistics for Herd 6 are summarized in Table 4.26. Herd size increased by approximately 20 cows, however it remained within the capacity of existing facilities. The overall cull rate was consistently lower than the state average of 37.7% (Hadley, et al 2006), and varied by less than 6 % throughout the study. The number of cows culled due to clinical JD increased and then decreased This was consistent with the pattern set by JD prevalence. The mortality rate increased over the course of this study. The reason for this is unknown, as the data simply reported which cows died; it did not detail why they died. Table 4.26: Descriptive Statistics for Herd 6 Year (Hdelrld size) (lbs 3.5% Cull/Rate (13.11110: 31:63) Mortality a t cows FCM) ° 0 c 1n1c %) Rate (%) 2003 145 26,875 ID ID ID 2004 143 27,987 30.1 2.8 5.6 2005 153 27,326 29.4 10.5 5.9 2006 169 26,593 32.5 5.3 7.1 2007 167 26,899 26.9 4.2 7.8 ID Insufficient data JD prevalence The within herd JD prevalence trend is outlined in Table 4.27. There was a dramatic increase in JD prevalence between 2003 and 2004. The reason for this increase was unknown, and samples were retested to rule out laboratory error. Test results 207 confinned estimated prevalence, and coincided with the producer observing an increase in the number of cows exhibiting signs of weight loss and diarrhea. This explains the sharp increase in the number of cows culled due to JD, with over one third of the cows culled in 2005 being culled due to clinical signs of JD. Subsequently, the declining prevalence of JD was associated with a decreasing incidence of cows developing clinical signs of the disease. Table 4.27: Johne’s disease prevalence trends 2003-2007 for Herd 6 Year Apparent JD prevalence (ELISA &lor F C positive) 2003 14.7% 2004 43.7% 2005 19.6% 2006 13.9% 2007 4.7% Cost of the JD control program 2003-2007 The costs of the JD control program observed over the five years of this study are summarized in Table 4.28. As part of the Michigan Johne’s Disease Control Demonstration Project, the herd did not have to pay for any JD testing beyond the cost of labor to collect samples on the day of the annual test. It is unlikely the herd would have done the extensive testing that was performed. if it had to pay for the testing, yet management decisions were made based on those test results, so for the sake of completeness, testing costs are included in the last column of the table. 208 Table 4.28: Cost of Johne’s disease control program 2003-2007 for Herd 6 ($/cow) No. Capital Total Year Supplies Management Labor Total . Cows Investments (+ testrng) 2003 145 $0.34 $4.83 $0.62 $0.00 $5.79 $34.79 2004 143 $0.00 $7.58 $1.86 $27.88 $37.33 $66.33 2005 153 $0.00 $7.27 $1.74 $34.30 $43.30 $72.30 2006 169 $0.00 $6.86 $1.64 $29.70 $38.20 $67.20 2007 167 $0.00 $3.74 $1.66 $28.62 $34.02 $63.02 Ave. 155 $0.07 $6.06 $1.50 $24.10 $31.73 $60.73 The costs for the supplies category were minor, and only consisted of the purchase of colored neck strings to identify JD test positive cows. Management included extra time spent aiding annual JD testing, record keeping, and making capital investment decisions. Labor included extra time spent on fresh cow and calf care. The capital purchase made by this farm was second skid-steer bought in the summer of 2004. It was financed, so interest paid is also included in the costs for capital investments. Remodeling the maternity barn was necessary and planned prior to the JD control program, so none of its costs are included in the cost of the control program. Similar to the purchase of a new tractor in Herd 4, assigning the entire cost of the new skid steer to the JD control program resulted in a robust estimate for the costs to this herd as compared to the other study herds. From a JD control standpoint, the objective of separate equipment to handle feed and manure could have been achieved with the purchase of a second bucket for the existing skid-steer, at less cost; even after adjusting labor costs to reflect the extra time required for changing the buckets between feeding 209 and cleaning pens. However, this study sought to estimate the actual costs of the JD control program as reported by the producer, and it was the producer’s choice to buy the new skid-steer and assign that cost to the JD control program. Economic losses due to JD 2003-2007 The annual estimated economic losses due to JD, along with the calculated, or assumed, benefits of the control program for Herd 6 are summarized in Table 4.29. Economic losses due to JD were calculated based on the calendar year. This herd was enrolled in the study in the fall of 2003, and annual testing occurred every fall thereafter. As a result, JD fecal culture results were generally not available until the following year, which is when management decisions based on those results were made. Thus, there often was a lag period between when JD prevalence was estimated and when the losses associated with that prevalence occurred. In other words, the estimated losses due to JD in 2004 were more reflective of the JD prevalence in 2003, and so on throughout the rest of the study. Also, availability of the herd data needed for this study was sketchy prior to study enrollment, which is why the estimated losses due to JD were so low in 2003. 210 Table 4.29: Economic losses due to Johne’s disease and assumed benefits of Johne’s disease control program for Herd 6 — 2003-2007 ($/cow) Year NO. Mllk RPO Cull Value Total Assumed Cows Value Benefits 2003 145 $0.00 $13.39 $4.99 $18.38 N/A 2004 143 $112.07 $43.03 $16.09 $171.19 -3275] 2005 153 $182.96 $0.00 $43.24 $226.20 $44.79 2006 169 $212.81 $15.85 $14.82 $243.48 -5203 5 2007 167 $194.71 $3.24 $21.59 $219.54 -3713 Ave. 155 $140.51 $15.10 $20.14 $175.76 N/A N/A Not applicable NPV cglculation The NPV was calculated for four different scenarios: (1) assuming a linear decline in losses following year five and disease eradication by year 20 with the control program; (2) assuming losses and JD prevalence remain constant at year five levels while continuing the control program; (3) assuming linear increase in losses at rate equal to that in scenario 1 with no control program; and (4) assuming losses remain constant at same level as the beginning of the study with no control plan implemented. The results are shown in Table 4.30. The estimated losses due to JD dictated the rate of increasing benefits projected over the remaining 20 years. The high estimated losses in 2007 for this herd resulted in rapidly increasing benefits in scenario 1, and led to a positive NPV by year 18. However, in scenario 2, the estimated losses beyond the observed study period were held constant at the rate equal to the losses in 2007. As the 2007 losses for this herd 211 exceeded its estimated baseline JD losses, the control program appeared to have “negative” benefits projected across the remaining 16 years of the calculation, resulting in a negative NPV. The NPV’s for scenarios 3 and 4 remain negative over the entire 20- year projected study period because, in the absence of a control program, there was nothing to offset the losses caused by JD. Table 4.30: NPV of four scenarios for Johne’s disease (JD) control on Herd 6 Year of Scenario“ Program 1 2 3 4 5 -$202.46 -$21 1.79 -$949.26 -$793.32 10 -$201.91 -$354.67 -$1,78l.75 -$1,333.24 20 $512.91 -$648.64 -$4,208.35 -$2,444.18 Payback Year 18 N/A N/A N/A N/A Not applicable Payback Year: Year of control program when NPV became positive * Scenario 1: Assuming linear decline in losses caused by JD after year 4 due to declining prevalence and eventual eradication after implementation of JD control program Scenario 2: Assuming losses caused by JD remain constant after year 4 while still investing in JD control program Scenario 3: Assuming losses caused by JD increase at a rate equal to the decline in Scenario 1 in the absence of JD control program Scenario 4: Assuming losses caused by JD remain constant at baseline level in absence of JD control program Results of sensitivity analysis The break-even cost for the JD control program is an estimate of the amount of money a farm can invest in the control program and still “break-even”, or have a NPV 212 equal to zero. The break-even cost calculated for scenario 1 was $74.20/cow/year, and for scenario 2, -$l9.22/cow/year. The NPV was most sensitive to the discount rate (r) used. After the discount rate, milk price, followed by cost of the JD control program had the greatest effect on the NPV. When the individual categories within the JD control program were evaluated, the results were ranked depending on which category represented the highest proportion of the control program. For this herd, changing the input costs of capital investments had the greatest impact on the NPV, followed by management, labor, and supplies respectfully. Including testing in the calculations increased the input costs of the JD control program, and hence decreased the NPV. For scenario 1, the NPV decreased from $513 to $439 when ELISA testing only was included, and to $168 when both ELISA and fecal culture was included. For scenario 2, the NPV decreased from -$649 to -$722 and -$993 respectfully, for including ELISA testing only or ELISA and fecal culture together. Producer perception of the JD control program At the conclusion of this study, the producer was satisfied with the JD control program and planned to continue investing in it. Subj ectively, herd health had improved over the course of this study, as had production, which led to increased revenues. Summary results 213 A. Farms and JD prevalence Table 4.31 outlines each study herd in terms of herd size, breed, and housing management, while the within herd JD prevalence is shown in Table 4.32. Over the five- year course of this study, several changes occurred in the management of these farms, beyond the simple implementation of a JD control program that may have had some impact on the JD status of the herd. For example, in 2004, herd 2, went through the dissolution of a partnership, placing the farm in a tenuous financial position accompanied by management instability and labor issues as it transitioned to a sole proprietorship. More importantly, from a JD prevalence standpoint, all cows that had ever tested positive for JD were culled from the herd as part of the partnership buyout. Herds 3 and 5 underwent significant expansion during the course of this study. Both expansions occurred through the purchase of a substantial number of cattle, and neither farm gave any consideration to the JD status of the individual cows purchased, or their herds of origin. Herd 3 increased herd size by purchasing and consolidating three small herds, including all young stock, in a period of about four months in 2004; and then did not buy any more cows, although they did continue to purchase breeding-age bulls. Herd 5 also purchased a couple of smaller herds, but continued to buy cows through a cattle broker throughout the study. As a result of their respective expansions, both herds decided to contract their replacements to heifer growers. Herd 3 keeps heifers through weaning to four months of age before sending them to the grower, while Herd 5 sends heifers to the grower at weaning. Both growers only raise heifers fOr the respective farms. Herd 4 became certified organic in the spring of 2005, contracting their milk at more than double the price on the commercial market. Herd 6 had been an entirely closed herd for 30 years 214 prior to the start of the study and remained so throughout the study. However, there was an unexplained increase in JD prevalence fi'om 15% to 44% between 2003 and 2004. Samples were retested to rule out lab error and verify the estimate of prevalence. Table 4.31: Herd size, breed and housing management of study herds Herd Ave. herd Herd size Breed Housing size Start End 1 191 170 215 Holstein Confinement/ grazing 2 125 103 137 Holstein Total confinement 3 3 78 2 l 8 458 Holstein Confinement] grazing 4 73 75 68 Jersey Rotational grazing (organic) 5 531 484 641 Holstein Total confinement 6 155 145 167 Holstein Total confinement Table 4.32: J ohne’s Disease test prevalence (fecal culture and/or ELISA positive) for herds during observed study period 2003-2007 Herd Year 1 2 3 4 5 6 2003 12% 10% 11% 12% NT 15% 2004 24% 4% 8% 9% 6% 44% 2005 22% 5% 13% 4% 5% 20% 2006 10% 9% 12% 8% 6% 14% 2007 15% 4% 16% 7% 4% 5% NT: Not tested 215 B. C 0st of the JD control program The annual costs of the JD control program for each herd are summarized in Table 4.33. The range was $5.79 - $81.07/cow/year with an average cost of $30.33/cow/year and median cost of $23.71/cow/year. Figure 4.1 shows the average annual costs of the JD control program broken down by category. While how much each herd spent varied greatly by category, many of the herds spent the greatest proportion of their money on labor and management. The negative supply cost for herds 3 and 4 is due to the adjustment to costs for additional milk sold by switching from whole milk to milk replacer to feed calves. In other words, the cost of milk replacer was less than the market value of an equivalent amount of whole milk. Of course, herd 4 was the organic Jersey farm, so they were receiving, on average, over twice as much for their milk as compared to the other herds. Table 4.33: Annual Cost of Johne’s disease control programs implemented by study herds 2003-2007. All costs represent $/cow in herd. Herd Year 1 2 3 4 5 6 2003 $16.24 $23.57 $58.22 $43.18 NC $5.79 2004 $16.29 $14.77 $41.46 $57.33 $14.85 $37.33 2005 $10.88 $15.37 $39.68 $67.99 $15.22 $43.30 2006 $10.05 $17.10 $41.77 $75.56 $14.95 $38.20 2007 $9.27 $7.68 $31.55 $81.07 $12.87 $34.02 Average $12.54 $ 15.70 $42.54 $65.02 $14.47 $31.73 NC: Not calculated 216 Figure 4.1: Average costs of Johne’s disease control program 2003-2007 broken down by category. Values are shown in Skew/year. Capital lnvestrnent Labor .1 Management # I Herd 1 [11 Herd 2 \\\\\ E Herd 3 I Herd 4 Supplies 9 Herd 5 GHerd 6 ’ ml , a .L -$30.00 -$20.00 -$10.00 $0.00 $10.00 $20.00 $30.00 $40.00 US$lcowaear C. Economic losses due to JD The annual losses due to JD for each herd are summarized in Table 4.34. The range of the losses was $15.85 - $243.48/cow/year with an average loss of $79.31/cow/year and a median loss of $66.17. Figure 4.2 shows the average annual losses due to JD broken down by category. As with the costs of the JD control program, the losses due to JD also varied greatly across herds. The one consistent thing in all herds was that the highest proportion (49 - 90%) of the JD economic losses was due to lost income from decreased milk production. 217 Table 4.34: Annual losses due to Johne’s disease for study herds 2003-2007. All values represent $/cow in herd. Herd Year 1 2 3 4 5 6 2003 $98.95 $15.85 $17.86 $69.97 NC $18.38 2004 $138.53 $48.67 $20.54 $77.18 $29.58 $171.19 2005 $79.97 $18.05 $55.75 $32.90 $47.91 $226.20 2006 $83.57 $82.13 $118.29 $77.71 $35.55 $243.48 2007 $60.23 $66.17 $66.54 $59.93 $19.24 $219.54 Average $92.25 $46.17 $55.80 $63.54 $33.07 $169.85 NC: Not calculated Figure 4.2: Average losses due to Johne’s disease 2003-2007 broken down by category. Values are shown in Skew/year. IHerd1 IDHerd 2 Cull Income 5 Herd 3 I Herd 4 I Herd 5 flHerd 6 RPO-value Milk Income $0.00 $50.00 $100.00 $150.00 $200.00 USS/cowlyear 218 D. NPV results The NPV of the JD control program with a 20-year projection horizon for each of four different scenarios for each herd is summarized in Table 4.35. For scenario 1, in which JD control program was assumed to result in disease eradication after 20 years, only three of the six herds (herds l, 5, and 6) had a positive NPV. The average NPV for all herds under scenario 1 was $24 per cow with a median of -$41 per cow. For scenario 2, in which it was assumed that, in spite of continued investing in the JD control program, the losses caused by the disease would remain at the level observed in 2007 for the remainder of the 20-year projection period, only 2 herds (herds l and 5) had a positive NPV. The average NPV for all herds under scenario 2 was -$411 per cow and a median of -$604 per cow. The NPV’s for scenarios 3 and 4 were negative for all herds, as there were no benefits because no control program was implemented. These two scenarios simply estimated the potential economic losses should the farm do nothing for JD control. E. Results of sensitivity analysis The break-even cost for the JD control program is an estimate of the amount of money a farm can invest in the control program that will “break-even”, or result in a NPV equal to zero. The break-even costs were calculated for scenarios 1 and 2 for each herd and are shown in Table 4.36. All the break-even costs for herds 2 and 3 are negative, suggesting that, given the assumed benefits outlined in each scenario, no amount of money invested into a JD control program will yield a zero NPV. On the other hand, if herd 4 could decrease the cost of its JD control program by 40% and 80% for scenarios 1 219 Eamoa 15:8 9. do 8:88 E .26. 8:88 a: 258:8 :EEE DH .3 88:8 v.88— w:_E:8< ”v 2.5:on 68on 3:28 0:: 8:88 :_ _ orgoom :_ 85 2 3:8 38 a: 88.85 9. .3 @838 888. w:_E:8< ”m 23:on :28on 35:8 0.. E w::8>:_ :8 233 v :8» 8c: E388 55:8 9. ,3 88:8 882 w:_E:8< ”N 288$ 88on 35:8 n: .8 :o_§:oEo_n_E_ 8c: 8:868 3388 :8 8:288: w:_:=8: 8 on: v :8» 8c: 0.. .3 :838 882 5 8:8: 28:: m5€=8< H. 2,3:on ... 216$.Nm- 8.: _ ma- 5.05;? wodmmm- 8.34m- mmfitm. E- v NENchom- 5.08%- 3.5%. ~ m- mmHRm- 3.53m- admormm- m 8.30%. on. _ mm mmémom- wodwo. _ m- Edwwm- 362% N 3.2mm afiwN—m ovsmmm- 3.3.5. 9.3%. MWwEm _ c m w m N _ "$8:ch so: :8: :_ 38% 8: 82m.» =< .8888 ::80&€ 58 :8 E8.» on San Emcmoa 35:8 53 8:85 8.230.. do 2:? “:88: 82 ”mm: 28... 220 .8885 35:8 0.. E @5885 :8 8:53 4 :8» 85a 588:8 :88: A: .3 888 888. 8358.4 “m 288% Ema—moi 35:8 0.. .3 85585255 853 858380 3388 :8 8838:: 83:8: 8 on: v :8.» 853 9. 3 58:8 888— :_ 8:8: 58:: w:_E:8< ”_ 28:on ... mad—m- 5N3; mod _ m 3.9%- 5.9%. 3.2% N Omaha . omémm owwmm 3.2m- Ema- $65 _ o m w m N _ v_aom8:8m :8: 583388 E 88 88> =< .N :8 _ 8.888 3.: Samoa 35:8 8883 8.85.33: 88 3.5 :53: 88-585, 30¢ 28:. 221 and 2 respectively, the expected NPV would equal zero; and cutting costs even more would result in a positive NPV. Across all herds, the NPV was most sensitive to the discount rate (r) used After the discount rate, milk price, followed by cost of the JD control program had the greatest effect on the NPV for all herds except herds 3 and 4. For those two herds the reverse was true, aside from the discount rate, the cost of the JD control program had the greatest effect on the NPV with milk price following. When the individual categories within the JD control program were evaluated, the results varied between herds depending on which category represented the highest proportion of the control program (Figure 4.1). The NPV calculations including JD testing are shown in Table 4.37 and can be compared to the NPV’s calculated without testing in Table 4.35. Including testing in the calculation increased the cost of the JD control program and hence decreased the NPV. Although not shown, the NPV value for running fecal culture alone would fall in between that of running the ELISA alone and running the two tests in parallel. F. Producer perceptions of JD control program At the end of the five-year study period producers were asked for their assessment of the JD control program implemented on their farms. All six producers stated that they were satisfied with the program and planned to continue it after the conclusion of the study. In support of this decision was the number of cows with clinical JD had decreased, and there was a feeling that overall herd and calf health had improved Five of the six herds reported that because of improved herd health, production had improved resulting in increased income. Herd 2 reported not seeing any increase in revenue as a 222 0:330 300“: “0..: E::wo:: 35:8 0.. E mc5w0>:_ :50 0:53 v ::0» :0c: 5:338 :_:::0: 9. »n 308:0 80:2 m:_:::mm< ”N 28:05 E::wo:: 35:8 9. .3 55888035 :33 55:03.20 33:03 3:: 00:0_:>0:: w::::00: o: 0:: v ::0» 8:: DH »5 303:0 80.8— E 0::00: ::0:= w:_E:mm< ”_ 23:05 ... NNS- moom- wa- mmmm- £5. 30%. 82m- 33%. mmom- woo»- own” 2; N omvm we; mmm Ema- :vm- mwcm- 3mm- mus—m- mwmm- 33. none 3.4% _ :8 5:: 5.5 «as: 5:5 $3: 58 <3: 5.5 <05: 5.5 $3.: 2%, MAP was cultured from the lactating cow floor and/or the manure storage area 75% of the time. Prevalence only had to increase to just over 5% before MAP started being cultured from other areas on the farm, the most common being the maternity area. In summary, MAP was widely distributed in the environment of infected dairy farms. As the within herd prevalence increased, so did the number and distribution of culture positive samples on the farm This was expected as the greater the number of MAP infected cows in a herd, the greater the potential for replenishing the environmental reservoir. 5.3. Do farm management practices, designed to limit the transmission of MAP infection, actually decrease the JD burden in a herd over time? With the environmental reservoir for MAP investigated, the next step was to study the source that replenishes it, MAP infected cattle. Farm management practices recommended for the prevention and control of JD are well documented, but there is little data fi'om real-life farm settings validating their effectiveness. Again, this is due mainly to the slowly progressive and chronic nature of JD that would require such studies be 239 conducted over a period of many years. This project is one of the few studies that has endeavored to validate the effectiveness of JD control programs implemented on naturally-infected dairy herds over a period of several years. This project was a series of intervention studies; one for each of the seven herds. A logical way of assessing the effectiveness of the intervention, or JD control program, is to compare some measure of disease before and after the implementation of the control program As most of the management practices implemented to control JD focused on minimizing transmission of MAP in young calves, the comparison was the prevalence or incidence of JD in cows born after the control program to that in cows born prior to the control program Regarding JD prevalence, the relative risk (R) was calculated for each herd. In all instances, the risk of testing positive for JD was lower for cattle born after the start of the control program than in cattle born before the control program (Table 3.11); providing summary evidence that the JD control programs were effective in preventing JD transmission. Another way to assess the effectiveness of the control programs implemented was to compare the JD incidence by lactation between cows born before and after the control program If the JD control program was working, disease burden in the herd would be reduced, as would MAP exposure, which would lower the infectious dose and delay the onset of disease. Thus, not only would the number of infected cattle be lower in cows exposed to the control program, the age at which they are detected will increase. Over the first three lactations, the incidence of JD in cows exposed to the control program as calves was consistently less than in cows not exposed to the control program. This 240 provided further evidence that the JD control programs in these herds were successful in minimizing disease burden. 5.4. What specific management practices are the most effective in decreasing the JD burden in a herd? Once it was verified that the control programs put in place on the study herds were successfiil in preventing or minimizing JD transmission, the next step was to determine which of the multiple management practices implemented were most effective in preventing disease. If one looks at the recommendations to control JD, the list is quite lengthy. This can overwhelm producers if presented in its entirety with no ranking, or prioritizing, of the practices in terms of “getting the most bang for the buck” so to speak. Convincing producers to adopt JD control programs is easier if management practices can be invested in over time, with the changes having the greatest impact on preventing MAP infection being implemented first. The specific management practices put in place to control JD, and the extent to which they were implemented, were unique to each herd and varied greatly; making it impossible to look at each practice individually. So in lieu of analyzing specific practices, the risk of JD transmission, as assessed by a standardized risk assessment, was used to determine the areas of the farm in which management changes had the greatest effect on reducing disease. Of the risks for JD transmission assessed, exposure to adult cows other than the dam at birth and feeding colostrum from one cow to multiple calves 241 had the most significant effect on cows testing positive for JD as adults (Table 3.16). In both instances, the probability of exposure of susceptible calves to MAP being shed by infected cows is increased. These results are consistent with the fact that the susceptibility of cattle becoming infected with MAP decreases with age. Also, the results seemed plausible, considering observations made on these farms. The JD prevalence in these herds varied greatly over the course of this study (0.6% to 43%, Figure 3.5); even ignoring the two herds (herds 3 and 5) that underwent significant expansion through the purchase of a large number of cows and whose JD prevalence remained relatively static. In fact, the herd with highest JD prevalence in this study was closed for over 30 years prior to the start of the program. This suggested that something in the management of these herds, beyond the purchase of cattle, was contributing to the spread of JD within these herds. When the risk assessments were compared, the risk of JD transmission was similar across the herds once calves were weaned. The areas where there was the greatest difference in management in terms of risk for JD transmission were the maternity and pre-weaned calf areas. Thus, it seems reasonable that the bulk of MAP infections occur in these areas. With these areas identified, management practices for JD control can be recommended that are designed specifically for each herd’s unique needs and capabilities. 242 5.5. Are management practices to control JD cost effective? Proving that JD control programs are effective in minimizing the disease burden in infected dairy herds, is not enough for the widespread adoption of management practices to control and prevent JD disease. The costs and benefits of investing in JD control, as well as the economic losses of JD in the absence of control, need to be quantified Producers are unlikely to invest in JD control programs if the cost of controlling the disease is greater than what the disease is costing them. To determine the cost-effectiveness of the JD control programs implemented by the herds in this study, the net present value (NPV) of the control program was projected over a 20-year period from the start of the program. The costs and benefits of the respective control programs observed over the five-year course of this study were extrapolated, and the NPV calculated based on different assumptions on the extent of disease burden in the presence, and absence, of a JD control program. Assuming JD could be eradicated after 20 years, the control program netted a positive return on investment in half the herds in this study; while resulting in a net loss for the other half of the herds. If it was assumed that JD prevalence and the economic losses remained at the same level as those observed at the end of the study for the remaining 20-year projection, the control program yielded a positive team on investment in only two herds. However, across all herds, when the potential ongoing losses due to JD in the absence of a control program were considered, investing in a control program was always a better economic choice than doing nothing (Table 4.35). 243 Moreover, irregardless of the calculated NPV of the JD control program, all the producers in this study were generally satisfied with their control programs, and planned to continue investing in them into the future. While the assessment was subjective, the producers stated that following the implementation of the control program, they saw an overall improvement in herd health (beyond a reduction in the number of cows with clinical JD) that translated into improved production and increased revenues. This suggests two things. First, producer perception of, and satisfaction with, the JD control program are just as important in the on-farm decision-making process as complicated calculations projected over an extended period of time and necessarily based on assumptions. Second, the calculations underestimated the NPV of the JD control program Certainly there was the opportunity for error to enter into the calculations. The questionnaire used to gather data on the economic costs of the JD control program, particularly regarding management and labor, relied on the producers’ ability to recall time spent on JD control. Also, the economic losses, from which the benefits of the control program were estimated, were only those directly attributed to JD. It did not take into consideration any ancillary effects the control program might have. Many of the management practices implemented to control JD are also recommended for the control of a multitude of other infectious diseases such as Salmonella, E. coli, Mycoplasma, bovine leukosis, etc. Unfortunately, baseline records for cow and calf health aside from JD were not available for the herds in this study. Future studies on the cost-effectiveness of JD control programs should include some measure for collateral improvement in overall herd health. This measure is potentially significant based on the subjective 244 accounts of the producers in this study. As a result, the estimates of the cost- effectiveness of JD control programs reported here are probably conservative. 5.6. Conclusion In conclusion, the four questions posed at the beginning of this research project have been answered. Serial culturing of the environment of infected dairy farms found that MAP is widely distributed; with the areas most commonly contaminated being those where manure from adult cows is concentrated. As JD prevalence in the herds declined so too did the number of culture positive environmental samples and the areas from which those samples originated. The JD control programs implemented on these farms were successful in minimizing disease transmission as evidenced by a reduced prevalence and incidence of JD in cattle exposed to the control program as calves compared to those not exposed to the program. Evidence also suggested that the majority of MAP infected cows acquired the infection at, or shortly after, birth. Therefore, management practices to minimize exposure of calves to MAP in the maternity and pre-weaned calf areas should take priority in any JD control program. Finally, JD control programs can be cost- effective. Investing in a JD control program was always a better economic decision than doing nothing. Moreover, the producers in this study were generally satisfied with the JD control programs they had implemented. They reported a decrease in clinical cases, an improvement in overall herd health resulting in better production and increased revenues, and they planned to continue investing in the JD control program into the firture. 245 APPENDICES 246 Risk Assessment 188888 : Johne’s D' Appendix A . 3358:: $2: :9: Ho> :9: 88232 3%. 30.— Ho> ”m9: @528 05 5 0.053.. @5085: :2 :6: 39:. 05 03::sz 00.0.: 00:60: 2 5:3: :0 8:29:05 9. :o 5:95 05 :0228 .ll 0. 0:80 20; So» .0: u 0:80 525:8: w::202: Emu 00:5: 002:0 7.50:: an .... 0055:: 8L Emu £3, »::m 002:0 0E2. $332.18; 80:: 2:8 :050 5 EB 002:0 :mazzlasa 0mg : 202:: :03: 05::2 wfizzlaizq 0:00:03: : 0.850 E. :0 08000:: 4832:1002: 0:58 :20 :o: :00: 8.: :02 a; 5% is: 50.: 8:000; :00 :0: :0: a: 2:5 05::2 vial civiro'co'rx'ad :8 82,8 02:00.18 m%@ 00:: .25: 0.352 Es: :2: a: a: 2: 85:8 208.: :82 :5 83:3 E08022: 05 :o 59: 05 0: x8 05 5 x :: 8:5: 22%: :2: $2 9:38 .< M01 1 Ham '9 Mi 'A 'r 50:: .2, E055 : 00:02 train it 'or marrow '9 5:050:22: :0 £058.20: 8.: 5 85:85 0>_::_0: 0080:: a 0:388 :05 220.8809: 9.0:: 0:80 5:23.? .w 00:09:: »__:0_:_ 50:: 50:: :05: 08228 E08858 E05: 39: 0:02 35008.5 :82: 0:020 :02: .860. 05 32.0.: 83:5 2:050:22: 50:0 :0: 520083 0.5:: 5 :8 :o 6:05: 00:00:: 53:52:00.8 0:: _::o:_:00: :o: 0.0:: 05 0::E_:mm 30:8 :0 8:03:25 2: 88:93:: 6:0 :0 83:20:30 :83 co 000:3 00:80 2:050:32: 0.0.”. 247 10> :00: .0> 20:50 : 00:02 F...: 0:20:00: 00.0.: 00:00: 0: 5.0: :0 0000.05: 9. :0 6:05. 05 :00.0:00 . :8: 2:: s x s 00:. 30.. 30.. 0> ”0:020: 08:03.0: 5 0.05.2. 0:600:00 :0: 0.0.: 0.00... 0:: 0::E=0m 0. 0:000 20.: 50> .00 u 0:000 0:05.08: 7.02000 2:8 5030:“. ...: 00500 :8: 0002. 0.0000 :0 €080.30 0030.00 050:: .038 E ”:0: :0 :oszEmEoo 95:9: 33:90: :0 53:00 300 .025 . $00500 x005 20:00:". ...: 02:00 :E: 0002. 0.000: :0 £080.30 .:0:::.:0 0.5:: .0300 :3 ”:06; :0 000: :0 00.60.5208 05:00: 0.0.0000 . :000500 >000: E0008“. I: 00:000 :6: 0002: 0.000: :0 05:: .0__0:0:0 .:00>:::. :: 0....:: :0 82:00.8 :0 08:55:28 0500:: 0.0.0000. .038 :0: a: .... 028 $0.00: 00 0...E 00.000 00:50:08.5 00:. H038 0.0.: 50.: .l- 02:000.: 0:. :0 552. 002:0 .::0>00 0: :58 .3220: 50:: 52:00.00 00: cxico'v: :038 0.0.: 501.: ..l 02:00: a., :0 552. 82:00.00 00.00: 00:. . ufim 'A 'Ol Ham '9 amapow '9 M01 2 WI 'A “t 0:200: 0.0.: 5.0: 80:03.20 248 ...: 0> ...—.— .0:.0.: 00000: 0: 0:...0: :0 0000.:>0:: Q. :0 :0005. 05 00.0000 . :00: .0> 50:50 : 00:02 :8: 2:: s x =0 :02: 0:05:02 :3 30.. 0> ”0:000: 00:83.50: 0. 0.0.0.0.. 05:00:00 :0: 0.0.: 0.0:... 0:: 055.3 0. 0:80 0:2. 50> .00 u 0:000 505.08: 35008:... ...: :0>02. 000000 0500 00:00:00.. : 0000:: 000:0: :0 000::0 0500.2 ...: 0.50002: ..I 0.02. .038 5.; 05:00: 0:000 .0 500000: I... 5.62. .2000: :0 =05: .:0::0.00 050:: .038 .3 :0 5.; 00:0..0 00:03 .5300 :0 00.3000 :0 000055058 :0: _:.50:00 .0 3:00:00: ...: :0>0z. .tocs: :0 20000 0000.00 0.00:: .038 5050500 .000: 00:0:0 .00.::: 38 00050: ”000: :0 00.50.5058 05:05 0.0.0000 .N E032 .... 0:02. 05:00: .038 5.; 000055058 :0: :0 :0058 8 50:5 .: «Em 'A 'l. amapow 'v l M01 'A ' 230w"— me :33: uozuogémon— .0 249 00.... 00> L0... _= 0'00; 00.0.: 00000: 0: 3.....0 00 0000.0>0:0 n... :0 8000:. 00: 00.0000 . :000 .0> 50:50 : 00:02 0:... 2.. s x .0 :0:: 30... 30.:. 50> ”0:000; 00:0 0. 0.00000 00.00000 :0: 0.0.0 30:... 00: 0.00:..0m 0. 0:000 0:00 50> ...:N n 0:000 05.0.08: 50800:: 1:002. 000000 00:00 00:00:00.. : 0000:: 000:0: 00 000:00 0500.). I15 $00002“. ..... 0:02: .038 5.3 05.030000 50008:“. I: 052. 0.0000 :0 00.5: 0000.00 00.00 .0300 .3 :0 0:.3 00:000 ”000500 :0:03 :0 8000.80.08 :0: .0.:00:00 3.000000 II. 0:02. .::000: :0 0.002.: 0000.00 050:: .0300 0000:0500 .000: 00:00 .0000: 38 00080: ”000: :0 02.00.5508 05000: 0.0.0000 0:02.: ..I 082. 05000: .038 5.3 00.00.5508 000 :0 80:08 38 80:5 . "an 'A '9 amapow 'c M01 ‘A 1 0:28: 0.0.0 3:0: 00:0 . 250 :8... 2:: 8 x s... 00:: 30.. 3 ._ 0> ”038 00080 0.00000 00.00000 :0: 0.0.0 3.0:... 00: 0:05.:0m. 00.0.: 00:00: 0: 3:30 00 0000.0>0:0 n... :0 8000:. 00: 00.0000 .III 0. 0:000 0:00 50> .0: n 0:000 50058.2 :000 .0> :00:50 : 00:02 3:00:02”. ...I 052. 000000 0000 00:00:00.. : 00005 000:0: 00 000:00 05005. .... 3:00:02“. ...: 0:624 0500...: 00:20 :0 00:0.0E0000 0: 000000 60:5 .m 3:00:02: l. 0020 0.0000 :0 000:: 0000.00 0.00:: .038 3: :0:03 :0 00.:00.0:0:08 0500:: 0.0.0000 .N 0%:8000 :1905 0.0000 :0 0000: 0000.00 0.00:: ..000:0.000 :3 00:20 :0 00: 0003 ”000: :0 00.:00.E0:08 0500:: 300 0.0.0000 .: ufim 'V M01 'l 0:28: :00: :30 .0 251 Appendix B: Economic Questionnaire Name Date Cost and revenue changes due to Johne’s Disease Control Program Please read through the following questionnaire that relates to biosecurity and Johne’s disease control changes and investments. Check the appropriate box ([1) next to any change(s) you have made or plan on making that was directly related to the Johne’s control program on your farm. Also provide quantities and dollar values where appropriate. If you have a change that is not listed please fill it in under “other.” Use the space provided including the back of pages as necessary. This survey covers the period from the start of the Johne’s disease Control Program (2003) until the current time A. Operations management changes because of Johne’s Control Program 1. Has any part of the farm enterprise changed as a direct result at the Johne’s Control Program (ie. Outsourced replacements to custom heifer raiser or eliminated steer enterprise)? Yes No If yes, what is the approximate cost of that change in terms of either additional expense or lost revenue? 252 2. Changes in supplies _DUE T O JOHNE ’S CONTROL PROGRAM ‘5’ g Change in 3 g” m or g quantity Cost = on W of Supply .5. 3 (specify unit) (S/unit) o 5 dung” Milk replacer I} D 3 C} Colostrum . C! D D E] replacer/supplement Ear tags/Animal :1 D 3 D Identification ' Sanitation supplies 2' D 3 3 Bedding :3 El 3 Z] Other (specify): Cl Cl C] :1 Other (specify): :1 C] [J 3 Other (specify): :1 D 3 CJ Comments: 253 3. Changes in managerial time and responsibilities DUE T O JOHNE ’S CONTROL PROGRAM Enterprise Increase Decrease Change (hours/month) Describe Once Ongoing Date or Exam of change Johne’s diagnosis/testing LJ [:1 Record-keeping Buying/ Selling decisions Animal logistics (where to calve cows, house calves, etc.) Capital Investment decision making Other (specify): Other (specify): Other (specify): L] What would it cost the operation to hire equivalent management services (gross or amount/hr)? What could the person making these management decisions earn doing something else (gross or amount/hr)? 254 4. Changes in labor and custom hired services DUE T O JOHNE ’S CONTROL PROGRAM g 2:; Change 2 E anllfnog; of 0 ‘- 0 00 Labor use E: g (hours 61w) Describe o 5. change Johne’s _ , — diagnosis/testing 3 3 C! “J Record-keeping Z] I} I} 3 Sanitation/cleaning 3 I} I] I! Calf care I} :1 3 I] Cow care 3 3 3 3 Animal handling 3 3 :1 C] Other (specify): 3 I] I] 3 Other (specify): 3 3 3 3 What is the labor cost for the reporting period? (e. g., $/hour)? Include all costs of hired labor such as benefits. 255 B. Capital investments (or anticipated investments) DUE TO JOHNE’S CONTROL PRERAM. Do NOT include capital investments that were made irrespective of Johne’s Disease Control FinancingI Total 0/ % Y Specific Cost % equity debt Cost- Category Item share Manure storage facilities 3 Feed storage facilities I] Feed handling equipment C] Cattle feeding facilities '3 Livestock housing Cl Maternity pens 3 Manure equipment Spreader U Skid-steer 3 Other (specify): :2 Pasteurizer :3 Other machinery j Improvements 3 Fencing 3 Other (specify): :1 Other (specify): 7‘} Other (specify): :3 1 Equity includes farm or personal resources such as savings. Debt refers to borrowed money. Cost share refers to money obtained through programs such as EQIP. 256 C. Producer Perceptions 1. Since beginning the Johne’s control program, have you seen any increases in revenues (e.g. higher price replacement heifers)? Explain. 2. Since beginning the Johne’s control program, have any other changes happened (even changes not related to J ohne’s disease at the farm? Explain. 3. Since beginning the Johne’s control program, do you think the farm is doing better (or worse) financially? Explain. 4. Since beginning the Johne’s control program, do you think the farm is doing better (or worse) produgtign-wise? Explain. 5. Since beginning the Johne’s control program, do you think herd-health is better (or worse)? Explain 6. On average, how much lower is the slaughter value of Johne’s disease clinical animals? Also, do Johne’s disease test-positive animals have a lower slaughter value? Explain. 7. Are you glad that you joined the J ohne’s disease demo-herd program? Do you think it was a good financial decision? Explain 8. Do you plan to continue taking steps to manage and reduce Johne’s disease in your herd after the end of the J ohne’s disease demo-herd program? Explain. 257 Appendix C: Farm input data for OptiCowTM Model 258 US Dollars Farm Input Source 2003 2004 2005 2006 2007 . OptiCowTM Herd average nnlk calculation from yield/yr (lbs/cow/yr) 311““th 00W ata Ave. milk price Ave. MMPA ($/cwt) basefarmprice $14.34 $18.07 $15.92 $14.02 $18.39 3:11;: replacement HRS-USDA $1572 $1871 $2021 $1878 $2004 Calf value HRS-USDA $108 $130 $145 $140 $154 . . OpIiCowTM Weight @ birth default 90 lbs - - ()ptiCowTM Mature hve weight default 1350 lbs fife/1th “amass ERS-USDA $0.53 $0.58 $0.58 $0.50 $0.50 V _ Michigan Dairy etennary COStS Farm Business - ($/cow/year) Analysis $100.69 $106.30 $106.43 $114.67 $119.83 Summary Financial losses at OptiCowTM S 50 disposal g $/case) default Insemination OpIiCowTM S 1 2 ($/insemination) default Heat detection rate SPHCOWTM 40% efault - OpiiCowTM o Conceptlon rate default 40 /o VWP Herd DHIA data Age at first calving Herd DHIA data REFERENCES 259 REFERENCES Adaska, J .M., Munoz-Zanzi, C.A., .Hietala, S.K., 2002. Evaluation of result variability with a commercial J ohne’s disease enzyme-linked immunosorbent assay kit and repeat testing of samples. J Vet Diagn Invest. 14, 423-426. Adaska, J .M., Anderson, RJ., 2003. Seroprevalence of Johne’s disease infection in dairy cattle in California, USA. Prev Vet Med. 60, 255-261. Aly, S.S., Thurmond, M.C., 2005. Evaluation of Mycobacterium avium subsp paratuberculosis infection of dairy cows attributable to infection status of the dam. J Am Vet Med Assoc. 227, 450-454. Antognoli, M.C., Hirst, H.L., Garry, F.B., Salman, M.D., 2007. Immune response to and faecal shedding of Mycobacterium avium ssp paratuberculosis in young dairy calves, and the association between test results in the calves and the infection status of their dams. Zoonoses Public Health. 54, 152—159. Ayele, W.Y., Bartos, M., Svastova, P., Pavlik, I., 2004. Distribution of Mycobacterium avium subsp. paratuberculosis in organs of naturally infected bull-calves and breeding bulls. Vet Microbiol. 103, 209-217. Barrington, G.M., Gay, J .M., Eriks, I.S., Davis, W.C., Evermann, J.F., Emerson, C., O’Rourke, J .L., Hamilton, M.J., Bradway, D.S., 2003. Temporal patterns of diagnostic results in serial samples from cattle with advanced paratuberculosis infections. J Vet Diagn Invest. 15, 195-200. Beard, P.M., Daniels, M.J., Henderson, D., Pirie, A.A., Rudge, K, Buxton, D., Rhind, S., Greig, A., Huthings, M.R., McKendrick, 1., Stevenson, K., Sharp, J .M., 2001. Paratuberculosis infection in nonruminant wildlife in Scotland. J Clin Microbiol. 39,1517-1521. Beaudeau, F., Belliard, M., Joly, A., Seegers, H., 2007. Reduction in milk yield associated with Mycobacterium avium subspecies paratuberculosis (Map) infection in dairy cows. Vet Res. 38, 625-634. Belknap, E.B., Getzy, D.M., Johnson, L.W., Ellis, R.P., Thompson, G.L., Shulaw, W.P., 1994. Mycobacterium paratuberculosis infection in two llamas. J Am Vet Med Assoc. 204, 1805-1808. Belloli, A., Carli, S., Pravettoni, D., Proverbio, D., Giangaspero M., Lauzi, S., Bonizzi, L., 2001. Case study — Attempted treatment of a cow with Mycobacterium avium enterocolitis. Bov Prac. 35, 56-60. 260 Benedictus, G., Dijkhuizen, A.A., Stelwagen, J ., 1987. Economic losses due to paratuberculosis in dairy cattle. Vet Rec. 121, 142-146. Benedictus, G., Kalis, C.J.H., 2003. Paratuberculosis: eradication, control and diagnostic methods. Act Vet Scand. 44, 231-241. Berghaus, RD, Lombard, J E Gardner, I.A., Farver, T.B., 2005. Factor analysis of a Johne’s disease risk assessment questionnaire with evaluation of factor scores and a subset of original questions as predictors of observed clinical paratuberculosis. Prev Vet Med. 72, 291-309. Berghaus, R.D., Farver, T.B., Anderson, R1, Jaravata, C.C., Gardner, LA, 2006. Environmental sampling for detection of Mycobacterium avium ssp. paratuberculosis on large California dairies. J Dairy Sci. 89, 963-970. Bielanski, A., Algire, J ., Randall, G.C.B., Surujballi, 0., 2006. Risk of transmission of Mycobacterium avium ssp. paratuberculosis by embryo transfer in vivo and in vitro fertilized bovine embryos. Theriogenology. 66, 260-266. Bolton, M.W., Grooms, D.L., Kaneene, J .B., 2005. Fecal shedding of Mycobacterium avium subsp. paratuberculosis in calves: implications for infection control and management. Proc 8th Int Coll Paratuberculosis. 596-600. Brumbaugh, G.W., Edwards J.F., Roussel, A.J., Thomson, TD, 2000. Effect of monensin sodium on histological lesions of naturally occurring bovine paratuberculosis. J Comp Path. 123, 22-28. Buergelt, C.D., Duncan, J .R., 1978. Age and milk production data of cattle culled fi'om a dairy herd with paratuberculosis. J Am Vet Med Assoc. 173, 478-480. Buergelt, C.D., Ginn, PE, 2000. The histopathologic diagnosis of subclinical Johne’s disease in North American Bison (Bison bison). Vet Microbiol. 77, 325-331. Buza, J .J ., Hikono, H., Mori, Y., Nagata, R., Hirayama, S., Bari, A.M., Aodon-geril, Shu, Y., Tsuji, N.M., Momotani, E., 2004. Neutralization of interleukin-10 significantly enhances gamma interferon expression in peripheral blood by stimulation with J ohnin purified protein derivative and by infection with Mycobacterium avium subsp. paramberculosis in experimentally infected cattle with paratuberculosis. Infect Immun. 72, 2425-2428. Carpenter, T.B., Gardner, [A., Collins, MT, Whitlock, RH., 2004. Effects of prevalence and testing by enzyme-linked immunosorbent assay and fecal culture on the risk of introduction of Mycobacterium avium subsp paratuberculosis- infected cows into dairy herds. J Vet Diagn Invest. 16, 31-38. 261 C etinkaya, B., Erdogan, H.M., Morgan, KL, 1997. Relationships between the presence of Johne’s disease and farm and management factors in dairy cattle in England. Prev Vet Med. 32, 253-266. Chi, J ., VanLeeuwen, J.A., Weersink, A., Keefe, G.P., 2002a. Direct production losses and treatment costs from bovine viral diarrhoea virus, bovine leukosis virus, Mycobacterium avium subspecies paratuberculosis, and Neospora caninum. Prev Vet Med. 55, 137-153. Chi, J ., VanLeeuwen, J .A., Weersink, A., Keefe, G.P., 2002b. Management factors related to seroprevalences to bovine viral-darr'hoea virus, bovine-leukosis virus, Mycobacteri um avium subspecies paratuberculosis, and Neospora caninum in dairy herds in the Canadian Maritirnes. Prev Vet Med. 55, 57-68. Chiodini, RJ., Van Kruiningen, H.J., 1983. Eastern white-tailed deer as a reservoir of ruminant paratuberculosis. J Am Vet Med Assoc. 182, 168-169. Chiodini, R.J., Van Kruiningen, H.J., Merkal, RS, 1984. Rurninant paratuberculosis (Johne’s disease): The current status and future prospects. Cornell Vet. 74, 217- 262. Chiodini, RJ, 1996. Immunology: Resistance to paratuberculosis. Vet Clin North Am Food Anim Pract. 12, 313-343. Chiodini, R1,, Rossiter, C.A., 1996. Paratuberculosis: a potential zoonosis. Vet Clin North Am Food Anim Pract. 12, 457-467. Collins, MT, Sockett, D.C., Ridge, S., Cox, J.C., 1991. Evaluation of a commercial enzyme-linked immrmoassay for Johne’s disease. J Clin Microbiol. 29, 272-276. Collins, MT, Morgan, J.R., 1992. Simulation model of paratuberculosis control in a dairy herd. Prev Vet Med. 14, 21-32. Collins, MT, 1994. Clinical approach to control bovine paratuberculosis. J Am Vet Med Assoc. 204, 208-210. Collins, MT, Sockett, D.C., Goodger, W.J., Conrad, T.A., Thomas, C.B., Carr, DJ, 1994. Herd prevalence and geographic distribution of, and risk factors for, bovine paratuberculosis in Wisconsin, J Am Vet Med Assoc. 204, 636—641. Collins, MT, 1996. Diagnosis of paratuberculosis. Vet Clin North Am Food Anim Pract. 12, 357-371. Collins, MT, 2003. Paratuberculosis: Review of present knowledge. Acta Vet Scand. 44, 217-221. 262 Collins, M.T., Wells, S.J., Petrini, KR, Collins, J E Schultz, R.D., Whitlock, RH, 2005. Evaluation of five antibody detection tests for diagnosis of bovine paratuberculosis. Clin Diagn Lab Immunol. 12, 685-692. Collins, M.T., Gardner, LA, Garry, F.B., Roussel, A.J., Wells, S.J., 2006. Consensus recommendations on diagnostic testing for the detection of paratuberculosis in cattle in the United States. J Am Vet Med Assoc. 229, 1912-1919. Corn, J .L., Manning, E.J.B., Sreevatsan, S., Fischer, J.R., 2005. Isolation of Mycobacterium avium subsp. paratuberculosis from free-ranging birds and mammals on livestock premises. Appl Env Microbiol 71, 6963-6967. Coussens, P.M., 2004. Model for immune responses to Mycobacterium avium subspecies paratuberculosis in cattle. Infect Immun. 72, 3089-3096. Cramwell, MP, 1993. Control of Johne’s disease in a flock of sheep by vaccination. Vet Rec. 133, 219-220. Daniels, M.J., Ball, N., Hutchings, M.R, Greig, A., 2001. The grazing response of cattle to pasture contaminated with rabbit faeces and the implications for the transmission of paratuberculosis. Vet J. 161, 306-313. Daniels, M.J., Hutchings, M.R., Allcroft, D.J., McKendrick, I.J., Greig, A., 2002. Risk factors for Johne’s disease in Scotland — the results of a survey of farmers. Vet Rec. 150, 135-139. Daniels, M.J., Henderson, D., Greig, A., Stevenson, K, Sharp, J.M., Hutchings, M.R., 2003a. The potential role of wild rabbits Oryctolagus cuniculus in the epidemiology of paratuberculosis in domestic ruminants. Epidemiol Infect. 130, 553-559. Daniels, M.J., Hutchings, M.R., Greig, A., 2003b. The risk of disease transmission to livestock posed by contamination of farm stored feed by wildlife excreta. Epidemiol Infect. 130, 561-568. Dargatz, D.A., Byrum, B.A., Barber, L.K, Sweeney, RW., Whitlock, R.H., Shulaw, W.P., Jacobson, R.H., Stabel, J .R., 2001. Evaluation of a commercial ELISA for diagnosis of paratuberculosis in cattle. J Am Vet Med Assoc. 218, 1163-1166. Davidson, W.R., Manning, E.J.B., Nettles, V.F., 2004. Culture and Serologic sm'vey for Mycobacterium avium subsp. paratuberculosis infection among Southeastern white-tailed deer (Odocoileus virginianus). J Wildl Dis. 40, 301-306. 263 de Lisle, G.W., Yates, G.F., Cavaignac, S.M., Collins, D.M., Paterson, B.M., Montgomery, RH, 2002. Mycobacterium paratuberculosis in wild ferrets - a potential wildlife reservoir of Johne’s disease. Proc 7‘h Int Coll Paratuberculosis. pp. 90. Dieguez, F.J., Gonzalez, A.M., Menendez, S., Vilar, M.J., Sanjuan, M.L., Yus, E., Amaiz, 1., 2008. Evaluation of four commercial serum ELISAs for detection of Mycobacterium avium subsp. paratuberculosis infection in dairy cows. Vet J. in press. Dorshorst, N.C., Collins, M.T., Lombard, J E 2006. Decision analysis model for paratuberculosis control in commercial dairy herds. Prev Vet Med 75, 92-122. Ellingson, J.L.E., Anderson, J .L., Koziczkowski, J .J., Radcliff, R.P., Sloan, S.J., Allen, S.E., Sullivan, N.M., 2005. Detection of viable Mycobacterium avium subsp. paratuberculosis in retail pasteurized whole milk by two culture methods and PCR. J Food Prot. 68, 966-972. Ellner, J .J., Wallis, RS, 1989. Immunologic aspects of mycobacterial infections. Rev Infect Dis. Suppl 2, $455-$459. FDA, Center for Veterinary Medicine, 2004. FDA Approves Rumensin® for increased milk production efficiency in dairy cows. Press release. Available from: http://www.fda.gov/cvm/CVM_Updates/monensinup.htm. Last accessed July 27, 2008. Featherstone, C., 1997. M cells: portals to the mucosal immune system. Lancet. 350, 1230. Feller, M., Huwiler, K, Stephan, R., Altpeter, E., Ahang, A., Furrer, H., Pfyffer, G.E., Jemmi, T., Baumgartner, A., Egger, M., 2007. Mycobacterium avium subspecies paratuberculosis and Crohn’s disease: a systematic review and meta-analysis. Lancet. 7, 607-613. Fischer, C.A., Matlova, L., Dvorska, L., Svastova, P., Bartl, J ., Melicharek, I., Weston, R.T., Pavlik, I., 2001. Diptera as vectors of mycobacterial infections in cattle and pigs. Med Vet Entomol. 15, 208-211. Fischer, O.A., Matlova, L., Dvorska, L., Svastova, P., Bartos, M., Weston, R.T., Kopecna, M., Trcka, 1., Pavlik, 1., 2005. Potential risk of Mycobacterium avium subspecies paratuberculosis spread by syrphid flies in infected cattle farms. Med Vet Entomol. 4, 360-366. Fredriksen, B., Djonne, B., Sigurdardottir, 0., Tharaldsen, J ., Nyberg, 0., Jarp, J ., 2004. Factors affecting the herd level of antibodies against Mycobacterium avium subspecies paratuberculosis in dairy cattle. Vet Rec. 154, 522-526. 264 Frehel, C., deChastellier, C., Lang, T., Rastogi, N., 1986. Evidence for inhibition of fusion of lysosomal and prelysosomal compartments with phagosomes in macrophages infected with pathogenic Mycobacterium avium. Infect Immun. 52, 252-262. F rehel, C., Thorel, M.F., Rastogi, N., 1989. Evidence that host-recycling of Mycobacterium avium preserves its ability to hinder macrophage killing functions. Acta Leprol. 7 Suppl. 160-163. Fyock, B.A., Smith, J ., Hoving, E., Schukken, Y., Whitlock, RH, 2005. Whole herd fecal pools to assess herd MAP bioburden compared to composite environmental samples. Proc 8th Int Coll Paratuberculosis. 258. Gilrnour, N.J.L., Nisbet, D.I., Brotherston, J .G., 1965. Experimental oral infection of calves with Mycobacteriumjohnei. J Comp Path. 75, 281-286. Godden, S.M., F etrow, J.F., F eirtag, J.M., Green, L.R., Wells, S.J., 2005. Economic analysis of feeding pasteurized nonsaleable milk versus conventional milk replacer to dairy calves. J Am Vet Med Assoc. 226, 1547-1554. Gollnick, N.S., Mitchell, R.M., Baumgart, M., Janagama, H.K Sreevatsan, S., Schukken, Y.H., 2007. Survival of Mycobacterium avium subsp. paratuberculosis in bovine monocyte-derived macrophages is not affected by host infection status but depends on the infecting bacterial genotype. Vet Irnmunol Immunopathol. 120, 93-105. Gonda, M.G., Chang, Y.M., Shook, G.E., Collins, M.T., Kirkpatrick, B.W., 2007. Effect of Mycobacterium paratuberculosis infection on production, reproduction, and health traits in US Holsteins. Prev Vet Med. 80, 103-119. Goodger, W.J., Collins, M.T., Nordlund, KV., Eisele, C., Pelletier, J ., Thomas, C.B., Sockett, DC, 1996. Epidemiologic study of on-farm management practices associated with prevalence of Mycobacterium paratuberculosis infections in dairy cattle. J Am Vet Med Assoc. 208, 1877-1881. Grant, I.R., Ball, H.J., Neill, S.D., Rowe, M.T., 1996. Inactivation of Mycobacterium paratuberculosis in cow’s milk at pasteurisation temperatures. Appl Environ Microbiol. 62, 631-636. Grant, I.R., Ball, H.J., Rowe, M.T., 1998. Effect of high temperature, short time pasteurisation on milk containing low numbers of Mycobacterium paratuberculosis. Lett Appl Microbiol. 26, 166-1 70. Grant, LR, 2003. Mycobacterium paratuberculosis and milk. Act Vet Scand. 44, 261- 266. 265 Greig, A., 2005. Mycobacterium avium subsp. paratuberculosis — its plans for survival. Proc 8th Int Coll Paratuberculosis. pp. 583-586. Grewal, S.K., Rajeev, S., Sreevatsan, 3., Michel, FC., 2006. Persistence of Mycobacterium avium subsp. paratuberculosis and other zoonotic pathogens during simulated composting, manure packing, and liquid storage of dairy manure. Appl Environ Microbiol. 72, 565-574. Groenendaal, H., Galligan, D.T., 1999. Economic consequences of Johne’s disease control programs. Center of Animal Health and Productivity, School of Veterinary Medicine, University of Pennsylvania, Animal Health Economics Staff Paper. Available from: http://cahpwww.vetupenn.edu/field/johne_report.pdf. Last accessed 10/31/2008. Groenendaal, H., Nielen, M., Jalvingh, A.W., Horst, S.H., Galligan, D.T., Hesselink, J .W., 2002. A simulation of Johne’s disease control. Prev Vet Med 54, 225- 245. Groenendaal, H., Galligan, D.T., 2003. Economic consequences of control programs for paratuberculosis in midsize dairy farms in the United States. J Am Vet Med Assoc. 223, 1757-1763. Grooms, D.L., Durst, P., Kaneene, J .B., 2006. Low Mycobacterium avium subspecies paratuberculosis ELISA specificity in a dairy herd. Proc 39th Ann Conv Am Assoc Bov Prac. pp. 323-324. Gwozdz, J.M., Thompson, K.G., Manktelow, B.W., Murray, A., West, D.M., 2000. Vaccination against paratuberculosis of lambs already infected experimentally with Mycobacterium avium subspecies paratuberculosis. Aust Vet J. 78, 560- 566. Hadley, G.L., Wolf, C.A., Harsh, SB, 2006. Dairy cattle culling patterns, explanations, and implications. J Dairy Sci. 89, 2286-2296. Hagan, W.A., 1938. Age as a factor in susceptibility to Johne’s disease. Cornell Vet. 38, 34. Hansen, D., Rossiter, C.A., 1999. Tools to use against Johne’s disease in cattle herds. Bov Prac. 33, 188-191. Hardin, L.E., Thorne, J .G., 1996. Comparison of milk with serum ELISA for the detection of paratuberculosis in dairy cows. J Am Vet Med Assoc. 209, 120-122. Harris, N.B., Barletta, R.G., 2001. Mycobacterium avium subsp. paratuberculosis in veterinary medicine. Clin Microbiol Rev. 4, 489-512. 266 Hasonova, L., Pavlik, 1., 2006. Economic impact of paratuberculosis in dairy cattle herds: a review. Veterinarni Medicina, 51, 193-211. Hendrick, S.H., Duflield, T., Leslie, K, Lissemore, K, Archarnbault, M., Kelton, D., 2005a. The prevalence of milk and serum antibodies to Mycobacterium avium subspecies paratuberculosis in dairy herds in Ontario. Can Vet J. 46, 1126-1129. Hendrick, S.H., Duflield, T.F., Kelton, D.F., Leslie, KE., Lissemore, K.D., Archarnbault, M., 2005b. Evaluation of enzyme-linked immunosorbent assays performed on milk and serum samples for detection of paratuberculosis in lactating dairy cows. J Am Vet Med Assoc. 226, 424-428. Hendrick, S.H., Kelton, D.F., Leslie, KE., Lissemore, K.D., Archarnbault, M., Duffield, T.F., 2005c. Effect of paratuberculosis on culling, milk production, and milk quality in dairy herds. J Am Vet Med Assoc. 227, 1302-1308. Hendrick, S.H., Duffield, T.F., Leslie, KE., Lissmore, K.D., Archambault, M., Bagg, R., Dick, P., Kelton, D.F., 2006a. Monensin might protect Ontario, Canada dairy cows from paratuberculosis milk-ELISA positivity. Prev Vet Med. 76, 237-248. Hendrick, S.H.,, Kelton, D.F., Leslie, KE., Lissemore, K.D., Archambault, M., Bagg, R., Dick, P., Duffield, T.F., 2006b. Efficacy of monensin sodium for the reduction of fecal shedding of Mycobacterium avium subsp. Paratuberculosis in infected dairy cattle. Prev Vet Med. 75, 206-220. Hill, B.B., West, M., Brock, KV., 2003. An estimated prevalence of Johne’s disease in a subpopulation of Alabama beef cattle. J Vet Diagn Invest. 15, 21-25. Hirst, H.L., Garry, F .B., Salman, M.D., 2002. Assessment of test results when using a commercial enzyme-linked immunosorbent assay for diagnosis of paratuberculosis in repeated samples collected from adult dairy cattle. J Am Vet Med Assoc. 220, 1685-1689. Hirst, H.L., Garry, G.B., Morley, P.S., Salmon, M.D., Dinsmore, R.P., Wagner, B.A., McSweeney, KD., Goodell, G.M., 2004. Seroprevalence of Mycobacterium avium subsp paratuberculosis infection among dairy cows in Colorado and herd- level risk factors for seropositivity. J Am Vet Med Assoc. 225, 97-101. Hoe, F .G.H., Ruegg, P.L., 2006. Opinions and practices of Wisconsin dairy producers about biosecurity and animal well-being. J Dairy Sci. 89, 2297-2308. Holmes, I.R.L., Jubb, T.F., Callinan, A.P.L., 2004. Infection rates in reactors to an absorbed ELISA used in a test and cull program for bovine Johne’s disease. Aust Vet J. 82, 233-235. 267 Horton, N.J., Bebchuk, J .D., Jones, C.L., Lipsitz, S.R., Catalano, P.J., Zahner, G.E.P., Fitzmaurice, G.M., 1999. Goodness-of-fit for GEE: An example with mental health service utilization. Statist Med. 18, 213-222. Hugh-Jones, M.E., Hubbert, W.T., Hagstad, H.V. (Eds), 1995. Zoonoses, Recognition, Control and Prevention. Iowa State University Press, Ames, Iowa. Hutjens, M.F., 2005. Dairy efficiency and dry matter intake. Proc. 7th Western Dairy Management Conf. pp. 71-76. Jansen, J ., 1948. Paratuberculosis. J Am Vet Med Assoc. 112, 52-54. Johnson-Ifearulundu,Y., Kaneene, J .B., 1997. Relationship between soil type and Mycobacterium paratuberculosis. J Am Vet Med Assoc. 210, 1735-1740. Johnson-Ifearulundu,Y., Kaneene, J .B., 1998. Management-related risk factors for M. paratuberculosis infection in Michigan, USA dairy herds. Prev Vet Med 37, 41- 54. Johnson-Ifearulundu, Y., Kaneene, J.B., Lloyd, J .W., 1999. Herd-level economic analysis of the impact of paratuberculosis on dairy herds. J Am Vet Med Assoc. 214, 822-825. Johnson-Ifearulundu, Y., Kannene, J.B., Sprecher, D.J., Gardiner, J .C., Lloyd, J .W., 2000. The effect of subclinical Mycobacterium avium paratuberculosis infection on days open in Michigan, USA dairy cows. Prev Vet Med 46, 171-181. Johnson, Y.J., Kaneene, J.B., Gardiner, J.C., Lloyd, J .W., Sprecher, D.J., Coe, RH, 2001. The effect of subclinical Mycobacterium paratuberculosis infection on milk production in Michigan dairy cows. J Dairy Sci. 84, 2188-2194. Jubb, T.F., Galvin, J .W., 2004. Effect of a test and control program for bovine Johne’s disease in Victorian dairy herds 1992-2002. Aust. Vet J. 82, 228-232. Judge, J ., Kyriazakis, 1., Greig, A., Hutchings, MR, 2005. Quantification of the faecal- oral route of Mycobacterium avium subspecies paratuberculosis transmission from rabbits to livestock. Proc 8th Int Coll Paratuberculosis. pp. 632. Jungersen, G., Huda, A., Hansen, J .J ., Lind, P., 2002. Interpretation of the Gamma Interferon test for diagnosis of subclinical paratuberculosis in cattle. Clin Diagn Lab Immunol. 9, 453-460. Juste, R.A., Garcia Marin, J.F., Peris, B., Saez de Ocariz, C., Badiola, J.J., 1994. Experimental infection of vaccinated and non-vaccinated lambs with Mycobacterium paratuberculosis. J Comp Pathol. 110, 185-194. 268 Kalis, C.H.J., Hesselink, J.W., Russchen, B.W., Barkema, H.W., Collins, M.T., Visser, I.J.R., 1999. Factors influencing the isolation of Mycobacterium avium subsp. paratuberculosis from bovine fecal samples. J Vet Diagn Invest. 11, 345-351. Kalis, C .H.J., Collins, M.T., Hesselink, J .W., Barkema, H.W., 2003. Specificity of two tests for the early diagnosis of bovine paratuberculosis based on cell-mediated immunity: the Johnin skin test and the gamma interferon assay. Vet Microbiol. 97, 73-86. Kalis, C.HJ., Collins, M.T., Barkema, H.W., Hesselink, J.W., 2004. Certification of herds as free of Mycobacterium paratuberculosis infection: actual pooled faecal results versus certification model predictions. Prev Vet Med. 65, 189-204. Keller, L.L., Harrell, C.D., Loerzel, S.M., Rae, DO, 2004. Johne’s disease: seroprevalence of Mycobacterium avium subspecies paratuberculosis in Florida beefand dairy cattle. Bov Prac. 38, 135-141. Khalifeh, M.S., Stabel, J .R., 2004. Effects of gamma interferon, inteleukin-IO, and transforming growth factor-B on the survival of Mycobacterium avium subsp. paratuberculosis in monocyte-derived macrophages from naturally infected cattle. Infect Immun 72, 1974-1982. Kim, S.G., Kim, E..,H Lafi'erty, C.J., Miller, L.J., Koo, H.J., Stehman, S.M., Shin, S.J., 2004. Use of conventional and real-time polymerase chain reaction for confirmation of Mycobacterium avium subsp. paratuberculosis in a broth-based culture system ESP [1. J Vet Diagn Invest. 16, 448-453. Klijn, N., Harrewegh, A.A.P.M., deJong, P., 2001. Heat inactivation data for Mycobacterium avium subsp. paratuberculosis: implications for interpretation. J Appl Microbiol. 91, 697-704. Kobayashi, S., Tsutsui, T., Yamamoto, T., Nishicuchi, A., 2007. Epidemiologic indicators associated with within-farm spread of J ohne’s disease in dairy farms in Japan. Bacteriology. 69, 1255-1258. Koets, A.P., Adugna, G., Janss, L.L.G., van Weering, H.J., Kalis, C.H.J., Wentink, G.H., Rutten, V.P.M.G., Schukken, Y.H., 2000. Genetic variation of susceptibility to Mycobacterium avium subsp paratuberculosis infection in dairy cattle. J Dairy Sci. 83, 2702-2708. Kopecky, KE., 1977. Distribution of bovine paratuberculosis in the US. J Am Vet Med Assoc. 170, 320-324. Korrnendy, B., Kopal, T., Balint, T., Szilagyi, M., Beki, L., 1989. Economic losses caused by paratuberculosis in a dairy herd: case report. Acta Vet Hung. 37, 54- 53. 269 Korrnendy, B., 1992. Paratuberculosis vaccine in a large dairy herd. Acta Vet Hung. 40, 171-184. Korrnendy, B., 1994. The effect of vaccination on the prevalence of paraturculosis in large dairy herds. Vet Microbiol. 41, 117-125. Kovich, D.A., Wells, S.J., Friendshuh, K, 2006. Evaluation of the voluntary Johne’s disease herd status program as a source of replacement cattle. J Dairy Sci. 89, 3466-3470. Kruip, T.A.M., Muskens, J ., vanRoermund, H.J.W., Bakker, D., Stockhofe-Zurwieden, N., 2003. Lack of association of Mycobacterium avium subsp. paratuberculosis with oocytes and embryos from moderate shedders of the pathogen. Theriogenology. 59, 165 l-1660. Kudahl, A.B., Ostergaard, S., Sorensen, J .T., Nielsen, SS, 2007. A stochastic model simulating paratuberculosis in a dairy herd. Prev Vet Med 78, 97-117. Lambrecht, R.S., Carriere, J.F., Collins, M.T., 1988. A model for analyzing growth kinetics of a slowly growing Mycobacterium sp. Appl Environ Microbiol. 54, 910-916. Larsen, A.B., Merkal, R.S., Cutlip, RC, 1975. Age of cattle as related to resistance to infection with Mycobacterium paratuberculosis. Am J Vet Res. 36, 255-257. Lawrence, W.E., 1956. Congenital infection with Mycobacteriumjohnei in cattle. Vet Rec. 68, 312—312. Lin, D.Y., Wei, L.J., Ying, Z., 2002. Model-checking techniques based on cumulative residuals. Biometrics. 58, 1-12. Linnabary, RD, Meerdink, G.L, Collins, M.T., Stabel, J.R., Sweeney, RW., Washington, M.K, Wells, S.J., 2001. Johne’s disease in cattle. Council for Agricultural Science and Technology. Issue Paper. 17, 1-10. Lombard, J .E., Garry, F.B., McCluskey, B.J., Wagner, BA, 2005. Risk of removal and effects on milk production associated with paratuberculosis status in dairy cows. J Am Vet Med Assoc. 227, 1975-1981. Lombard, J E Wagner, B.A., Smith, R.L., McCluskey, B.J., Harris, B.N., Payeur, J .B., Garry, F .B., Salman, M.D., 2006. Evaluation of environmental sampling and culture to determine Mycobacterium avium subspecies paratuberculosis distribution and herd infection status on US dairy Operations. J Dairy Sci. 89, 4162-4171. 270 Manning, E.J.B., Augenstein, M., Collins, M.T., Nelson, KM., 2003. Case report - Johne’s disease: the recipient risk. Bov Prac. 37 , 20-22. Marce, C., Beaudeau, F ., Bareille, N., Seegers, H., Fourichon, C., 2007. Effects of infection by Mycobacterium avium paratuberculosis on fertility of dairy cows. Proc 7th Int Coll Paratuberculosis. pp. 114. McDonald, W.L., O’Riley, KJ., Schroen, C.J., Condron, RJ., 2005. Heat inactivation of Mycobacterium avium subsp. paratuberculosis in milk. Appl Environ Microbiol. 71, 1785-1789. McKenna, S.L.B., Keefe, G.P., Tiwari, A., VanLeeuwen, J., Barkema, H.W., 2006. Johne’s disease in Canada part H: Disease impacts, risk factors, and control programs for dairy producers. Can Vet J. 47, 1089-1099. McNab, W.B., Meek, A.H., Martin, S.W., Duncan, J.R., 1991. Associations between dairy production indices and lipoarabinomannan enzyme-immunoassay results for partuberculosis. Can J Vet Res. 55, 356-361. McQueen, D. S., Russel, E. G., 1979. Culture of Mycobacterium paratuberculosis from bovine fetuses. Aust Vet J. 55, 203- 204. Merck Veterinary Manual, 7th Ed. Merck & Co., Inc, New Jersey, 1991. pp. 1474, 1540. Merkal, R.S., Larsen, A.B., Booth, GD, 1975. Analysis of the effect of inapparent bovine paratuberculosis. Am J Vet Res. 36, 837-838. Merkal, RS., Miller, J .M., Hintz, A.M., Bryner, J .H., 1982. Intrauterine inoculation of Mycobacterium paratuberculosis into guinea pigs and cattle. Am J Vet Res. 43, 676-678. Meyer, A.L., Hall, H.H., 1994. Economic analysis of the impact of paratuberculosis on the Kentucky cattle industry. Department of Agricultural Economics, University of Kentucky. Staff paper #343. Milmer, E., Daroogheh, K, Mehta, P.K, Cirillo, S.L.G., Cirillo, J .D., Bermudez, LB, 2005. Identification of Mycobacterium avium genes that affect invasion of the intestinal epithelium. Infect Immun. 73, 4214-4221. Momotani, E., Whipple, D.L, Thierrnann, A.B., Cheville, NR, 1988. Role of M cells and macrophages in the entrance of Mycobacterium paratuberculosis into domes of ileal Peyer’s patches in calves. Vet Pathol. 25, 131-137. 271 Motiwala, A.S., Strother, M., Theus, N.E., Stich, RW., Byrum, B., Shulaw, W.P., Kapur, V., Sreevatsan, S., 2005. Rapid detection and typing of strains of Mycobacterium avium subsp. paratuberculosis fi'om broth cultures. J Clin Microbiol. 43, 2111— 2117. Muskens, J ., Elbers, A.R.W. vanWeering H.J., Noordhuizen, J.P.T.M., 2003a. Herd management practices associated with paratuberculosis seroprevalence in Dutch dairy herds. J. Vet Med. 50, 372-377. Muskens, J., Mars, M.H., Elbers, A.R.W., van Mannen, K, Bakker, D., 2003b. The results of using faecal culture as confirmation test of paratuberculosis-seropositive dairy cattle. J Vet Med. 50, 231-234. Nielsen, S.S., Gronbaek, C., Agger, J .F ., Houe, H., 20023. Maximum-likelihood estimation of sensitivity and specificity of ELISAs and faecal culture for diagnosis of paratuberculosis. Prev Vet Med 53, 191-204. Nielsen, S.S., Grohn, Y.T., Quaas, R.L., Agger, J .F., 2002b. Paratuberculosis in dairy cattle: variation of the antibody response in offspring attributable to the dam. J Dairy Sci. 85, 406-412. Nielsen, S.S., Toft, N., 2006. Age-Specific characteristics of ELISA and fecal culture for purpose-specific testing for paratuberculosis. J Dairy Sci. 89, 569-579. Nielsen, S.S., Toft, N., 2007. Assessment of management-related risk factors for paratuberculosis in Danish dairy herds using Bayesian mixture models. Prev Vet Med 81, 306-317. Nordlund, KV., Goodger, W.J., Pelletier, J ., Collins, M.T., 1996. Associations between subclinical paratuberculosis and milk production, milk components, and somatic cell counts in dairy herds. J Am Vet Med Assoc. 208, 1872-1876. Obasanjo, I.O., Grohn, Y.T., Mohammed, HQ, 1997. Farm factors associated with the presence of Mycobacterium paratuberculosis infection in dairy herds on the New York State paratuberculosis control program. Prev Vet Med 32, 243-251. Olsen, 1., Sigurdardottir, O.G., Djonne, B., 2002. Paratuberculosis with special reference to cattle: a review. Vet Quarterly. 32, 12-28. Olsen, K.D., 2004. Farm Management Principles and Strategies. Iowa State Press, Ames, Iowa. Ott, S.L., Wells, S.J., Wagner, B.A., 1999. Herd-level economic losses associated with Johne’s disease on US dairy operations. Prev Vet Med. 40, 179-192. 272 Palmer, M.V., Stoffregen, W.C., Carpenter, J.G., Stabel, J .R., 2005. Isolation of Mycobacterium avium subsp paratuberculosis (Map) from feral cats on a dairy farm with Map-infected cattle. J Wildlife D. 41, 629-635. Patterson, C.J., LaVenture, M., Hurley, S.S., Davis, JP, 1988. Accidental self- inoculation with Mycobacterium paramberculosis bacterin (Johne’s bacterin) by veterinarians in Wisconsin. J Am Vet Med Assoc. 192, 1197-1199. Pavlik, 1., Yayo Ayele, W., Fischer O.A., 2002. Role of the external environment, plants, and non-vertebrates for the spread of Mycobacterium avium subsp. paratuberculosis. Proc 7th Int Coll Paratuberculosis. pp. 90. Pearson, J .KL., McClelland, T.G., 1955. Uterine infections and congenital Johne’s disease in cattle. Vet Rec. 67, 615-617. Pearson, L., 1908. A note on the occurrence in America of chronic bacterial dysentery of cattle. Am Vet Rev. 32, 602-605. Pence, M., Baldwin, C., Carter-Black, C., 2003. The seroprevalence of Johne’s disease in Georgia beef and dairy cull cattle. J Vet Diagn Invest. 15, 475-477. Pence, M., Thomson, J .U., Cole, D. 2004. Important elements of the HACCP process to control Johne’s disease, BVDV, and BLV. Bov Prac. 38, 126-132. Perez, V., Tellechea, J ., Badiola, J.J., Gutierrez, M., Garcia Marin, J .F., 1997. Relation between serologic response and pathologic findings in sheep with naturally acquired paratuberculosis. Am J Vet Res. 58, 799-803. Pinedo, P.J., Rae, D.O., Williams, J E Donovan, G.A., Melendez, P., Buergelt, CD, 2008. Association among results of serum ELISA, faecal culture and nested PCR on milk, blood and faeces for the detection of paratuberculosis in dairy cows. Tranbound Emerg Dis. 55, 125-133. Prescott, J .F ., Baggot, J .D., Walder, RD. (Eds) 2000. Antimicrobial Therapy in Veterinary Medicine. 3rd ed. Iowa State University Press, Ames, Iowa. Raizman, E.A., Wells, S.J., Godden, S.M., Bey, R.F., Oakes, M.J., Bentley, D.C., Olsen, KE., 2004. The distribution of Mycobacterium avium ssp. paratuberculosis in the environment surrounding Minnesota dairy farms. J Dairy Sci. 87, 2959-2966. Raizman, B.A., Wells, S.J., Jordan, P.A., DelGiudice, G.D., Bey, RR, 2005. Mycobacterium avium subsp. paratuberculosis fiom free-ranging deer and rabbits surrounding Minnesota dairy herds. Can J Vet Res. 69, 32-38. 273 Raizman, E.A., Wells, S.J., Godden, S.M., Fetrow, J ., Friendshuh, K, Oakes, J.M., 2006. Characterization of Minnesota dairy herds participating in a Johne’s disease control program and evaluation of the program risk assessment tool. Prev Vet Med 75, 22-33. Raizman, B.A., Fetrow, J., Wells, S.J., Godden, S.M., Oakes, M.J., Varquez, G., 2007. The association between Mycobacterium avium subsp. paratuberculosis fecal shedding or clinical Johne’s disease and lactation performance on two Minnesota, USA dairy farms. Prev Vet Med. 78, 179-195. Rajeev, S., Shulaw, W., Berghaus, R, Zhang, Y., Byrum, B., 2006. A testing scheme for . the detection of Mycobacterium avium subsp. paratuberculosis in bovine feces utilizing the ESP para-JEM liquid culture system. J Vet Diagn Invest. 18, 529- 535. Rankin, JD., 1961. The experimental infection of cattle with Mycobacterium johnei III. Calves maintained in an infectious environment. J Comp Pathol. 71, 10-15. Rankin, JD., 1962. The experimental infection of cattle with Mycobacterium johnei IV. Adult cattle maintained in an infectious environment. J Comp Pathol. 72, 113- 1 1 7. Rast, L., Whittington, RJ., 2005. Longitudinal study of the spread of ovine Johne’s disease in a sheep flock in southeastern New South Wales. Aust. Vet J. 83, 227- 232. Rastogi, N., David H.L., 1988. Mechanisms of pathogenicity in mycobacteria. Biochimie. 70, 1101-1120. Reddacliff, L.A., Vadali, A., Whittington, RJ., 2003. The effect of decontamination protocols on the numbers of sheep strain Mycobacterium avium subsp. paratuberculosis isolated from tissues and faeces. Vet Microbiol. 95, 271-282. Reddacliff, L., Eppleston, J ., Windsor, P., Whittington, R., Jones. 8., 2006. Efficacy of 3 killed vaccine for the control of paratuberculosis in Australian sheep flocks. Vet Microbiol. 15, 77-90. Richards, W.D., Thoen, CO, 1977. Effect of freezing on the viability of Mycobacterium paratuberculosis in bovine feces. J Clin Microbiol. 6, 392-395. Richards, W.D., 1989. Environmental acidity may be the missing piece in the Johne’s disease puzzle. In: Millner, A.R, Wood, P.R. (Eds), Johne’s Disease: Current Trends in Research Diagnosis and Management. CSIRO, Victoria, Australia, pp. 90-103. 274 Ridge, S.E., Baker, 1.M., Hannah, M., 2005. Effect of compliance with recommended calf-rearing practices on control of bovine J ohne’s disease. Aust Vet J. 83, 85- 90. Rossiter, C.A., Burhans, W.S., 1996. Farm-specific approach to paratuberculosis (Johne’s disease) control. Vet Clin North Am Food Anim Pract. 12, 383-415. Roussel, A.J., Libal, M.C., Whitlock, R.H., Hairgrove, T.B., Barling, KS., Thompson, J .A., 2005. Prevalence of and risk factors for paratuberculosis in purebred beef cattle. J Am Vet Med Assoc. 226, 773-778. Roussel, A.J., Fosgate, G.T., Manning, E.J.B., Collins, M.T., 2007. Association of fecal shedding of mycobacteria with high ELISA-determined seroprevalence for paratuberculosis in beef herds. J Am Vet Med Assoc. 230, 890-901. Rowe, M.T., Grant, LR, 2006. Mycobacterium avium ssp. paratuberculosis and its potential survival tactics. Lett Appl Microbiol. 42, 305-311. Ryter, A, Frehel, C., Rastogi, N., David, H.L., 1984. Macrophage interaction with mycobacteria including M. leprae. Acta Leprol. 2, 211-226. Seitz, S.E, Heider, L.E., Heuston, W.D., Bech-Nielsen, S., Rings, D.M., Spangler, L., 1989. Bovine fetal infection with Mycobacterium paratuberculosis. J Am Vet Med Assoc. 194, 1423-1426. Smith, RD., 1995. Veterinary Clinical Epidemiology: A Problem-Oriented Approach. CRC Press, Inc., Boca Raton, Florida. Sockett, D.C., Carr, D.J., Collins, M.T., 1992. Evaluation of conventional and radiometric fecal culture and a commercial DNA probe for diagnosis of Mycobacterium paratuberculosis infections in cattle. Can J Vet Res. 56, 148- 153. Spangler, E., Heider, L. E. ,Bech-Nielsen, S. Dom, C. R, 1991. Serologic enzyme-linked immunosorbent assay responses of calves vaccinated with a killed Mycobacterium paratuberculosis vaccine. Am J Vet Res. 52, 1197-1200. Spangler, E., Bech-Nielsen, S., Heider, LE, 1992. Diagnostic performance of two serologic tests and fecal culture for subclinical paratuberculosis, and associations with production. Prev Vet Med. 13, 185-195. St. Jean, G., 1996. Treatment of clinical paratuberculosis in cattle. Vet Clin North Am Food Anim Pract. 12, 417-430. 275 Stabel, J .R, 1997. An improved method for cultivation of Mycobacterium paratuberculosis from bovine fecal samples and comparison to three other methods. J Vet Diagn Invest. 9, 375-380. Stabel, J.R., 1998. Johne’s disease: A hidden threat. J. Dairy Sci. 81, 283-288. Stabel, J .R, 2000. Transitions in immune responses to Mycobacterium paratuberculosis. Vet Microbiol. 77, 465-473. Stabel, J .R., Whitlock, RH., 2001. An evaluation of a modified interferon-y assay for the detection of paratuberculosis in dairy herds. Vet Immunol Immunopathol. 79, 69-81. Stabel, J .R., Wells, S.J., Wagner, B.A., 2002. Relationships between fecal culture, ELISA, and bulk tank milk test results for Johne’s disease in US dairy herds. J Dairy Sci. 85, 525-531. Stabel, J .R, Hurd, S., Calvente, L., Rosenbusch, RF., 2004. Destruction of Mycobacterium paratuberculosis, Salmonella spp., and Mycoplasma spp. in raw milk by a commercial on-farm hi gh-temperature, short-time pasteurizer. J Dairy Sci. 87, 2177-2183. Stehman, S.M., 1996. Paratuberculosis in small ruminants, deer, and South American camelids. Vet Clin North Am Food Anim Pract. 12, 441-455. Step, D.L., Streeter, RN., Kirkpatrick, J .G., 2000. Johne’s disease update. Bov Prac. 34, 6-12. Streeter, RN., Hoffsis, G.F., Bech-Nielsen, S., Shulaw, W.P., Rings, D.M., 1995. Isolation of Mycobacterium paratuberculosis fi'om colostrum and milk of subclinically infected cows. Am J Vet Res. 56, 1322-1324. Storset, A.K, Hasvold, H.J., Valheim, M., Brun-Hanse, H., Bemtsen, G., Whist, S.K., Djonne, 3., Press, C.M., Holstad, G., Larsen, H.J., 2001. Subclinical paratuberculosis in goats following experimental infection. An immunological and microbiological study. Vet Immunol Immunopathol. 80, 271-287. Storset, A.K, 2003. Immunology of Mycobacterium avium subsp. paratuberculosis infection - immunodiagnostic methods. Acta Vet Scand. 44, 223-229. Stott, A.W., Jones, G.M., Humphry, RW., Gunn, G.J., 2005. Financial incentive to control paratuberculosis (J ohne’s disease) on dairy farms in the United Kingdom. Vet Rec. 156, 825-831. 276 Sweeney, RW., Whitlock, R.H., Rosenber,A.E., 19923. Mycobacterium paratuberculosis isolated fiom fetuses of infected cows not manifesting signs of the disease. Am J Vet Res. 53, 477-480. Sweeney, RW., Whitlock, R.H., Rosenberger, A.E., 1992b. Mycobacterium paratuberculosis cultured from milk and supramammary lymph nodes of infected asymptomatic cows. J Clin Microbiol. 30, 166-171. Sweeney, RW., Whitlock, R.H., Rosenberger, A.E., 1992c. Isolation of Mycobacterium paratuberculosis after oral inoculation of uninfected cattle. Am J Vet Res. 53, 1312-1314. Sweeney, RW., Whitlock, R.H., Buckley, C.L., Spencer, RA, 1995. Evaluation of a commercial enzyme-linked immunosorbent assay for the diagnosis of paratuberculosis in dairy cattle. J Vet Diagn Invest. 7, 488-493. Sweeney, RW., 1996. Transmission of paratuberculosis. Vet Clin North Am Food Anim Pract. 12, 305-312. Sweeney, RW., Whitlock, R.H., McAdams, S., Fyock, T., 20063. Longitudinal study of ELISA seroreactivity to Mycobacterium avium subsp. paratuberculosis in infected cattle and culture-negative herd nurtes. J Vet Diagn Invest. 18, 2-6. Sweeney, RW., Uzonna, J ., Whitlock, R.H., Habecker, P.L., Chilton, P., Scott, R, 2006b. Tissue predilection sites and effect of dose on Mycobacterium avium subsp. paratuberculosis organism recovery in a short-term bovine experimental oral infection model, Res Vet Med. 80, 253-259. Tavompanich, S., Gardner, I.A., Anderson, RJ., Shin, S., Whitlock, R.H., Fyock, T., Adaska, J .M., Walker, R.L., Hietala, S.K., 2004. Evaluation of microbial culture of pooled fecal samples for detection of Mycobacterium avium subsp. paratuberculosis in large dairy herds. Am J Vet Res. 65, 1061-1070. Tavompanich, S., Munoz-Zanzi, C.A., Wells, S.J., Raizman, B.A., Carpenter, T.B., Johnson, W.O., Gardner, I.A., 2008. Simulation model for evaluation of testing strategies for detection of paratuberculosis in Midwestern US dairy hreds. Prev Vet Med. 83, 65-82. Taylor, T.K, Wilks, C.R, McQueen, D.S., 1981. Isolation of Mycobacterium paratuberculosis from the milk of a cow with J ohne’s disease. Vet Rec. 109, 532-533. Thoen, C.O., Jamagin, J .L., Richards, W.D., 1975. Isolation and identification of mycobacteria fiom porcine tissues: a three-year study. Am J Vet Res. 36, 1383- 1386. 277 Thoen, C.O., Moore, LA, 1989. Control of Johne’s disease in four commercial dairy herds in Iowa. J Vet Diagn Invest. I, 223-226. Thorne, J .G., Hardin, LE, 1997. Estimated prevalence of paratuberculosis in Missouri, USA cattle. Prev Vet Med. 31, 51-57. Tiwari, A., VanLeeuwen, J .A., McKenna, S.L.B., Keefe, G.P., Barkema, H.W., 2006. Johne’s disease in Canada part I: Clinical symptoms, pathophysiology, diagnosis, and prevalence in dairy herds. Can Vet J. 47, 874-882. USDA, 1997. Johne’s Disease on US. Dairy Operations. NAHMS USDA:APHIS:VS:CEAH, Fort Collins, CO, USA, #N245.1097. USDA, 2005. Johne’s disease on US. dairy operations, 2002. USDA:APHIS:VS:CEAH, National Animal Health Monitoring System, Fort Collins, CO #N27.0205. pp. 26-27. USDA, 2005. Uniform program standards for the voluntary bovine Johne’s disease Mycobacterium avium subsp. Paratuberculosis. USDA-API-IIS 91-45-016. Available from http://www.aphis.usda.gov/animal_health/animal_diseases/johnes/ downloads/johnes-umrpdf. Last accessed 5/16/2008. USDA, 2008. Johne’s Disease on US. Dairies, 1991-2007. NAHMS USDAzAPI-IIS2VS, CEAH, Fort Collins, CO, USA, #N521.0408. vanRoermund, H.J.W., deJong, M.C.M., 2005. Horizontal transmission experiment of paratuberculosis. Proc 8th Int Coll Paratuberculosis. pp. 672-678. van Schaik, G., Kalis, C.HJ., Benedictus, G., Dijkhuizen, A.A., Huirne, RB.M., 1996. Cost-benefit analysis of vaccination against paratuberculosis in dairy cattle. Vet Rec. 139, 624-627. van Schaik, G., Rossiter, C.R., Stehman, S.M., Shin, S.J., Schukken, Y.H., 2003a. Longitudinal study to investigate variation in results of repeated ELISA and culture of fecal samples for Mycobacterium avium subsp. paratuberculosis in commercial dairy herds. Am J Vet Res. 64, 479-484. van Schaik, G., Stehman, S.M., Schukken, Y.H., Rossiter C.R., Shin, S.J., 2003b. Pooled fecal culture sampling for Mycobacterium avium subsp. paratuberculosis at different herd sizes and prevalenc. J Vet Diagn Invest. 15, 233-241. van Schaik, G., Stehman, S.M., Jacobson, R.H., Schukken, Y.H., Shin, S.J., Lein, DH, 2005. Cow-level evaluation of a kinetics ELISA with multiple cutoff values to detect fecal shedding of Mycobacterium avium subspecies paratuberculosis in New York State dairy cows. Prev Vet Med. 72, 221-236. 278 van Schaik, G., Pradenas, M.F., Mella, A.N., Kruze, J .V., 2007. Diagnostic validity and costs of pooled feal samples and individual blood or fecal samples to determine the cow- and herd-status for Mycobacterium avium subsp. paratuberculosis. Prev Vet Med. 82, 159-165. Vary, P.H., Andersen, P.R, Green, E., Hermon-Taylor, J ., McFadden, J .J ., 1990. Use of highly specific DNA probes and the polymerase chain reaction to detect Mycobacterium paratuberculosis in Johne’s disease. J Clin Microbiol. 28, 93 3- 937. Ward, M.P., Perez, A.M., 2004. Association between soil type and paratuberculosis in cattle herds. Am J Vet Res. 65, 10-14. Waters, W.R., Miller, J.M., Palmer, M.V., Stabel, J.R, Jones, D.E., Koistinen, KA., Steadham, B.M., Hamilton, M.J., Davis, W.C., Bannantine, JP, 2003. Early induction of humoral and cellular immune responses during experimental Mycobacterium avium subsp. paratuberculosis infection of calves. Infect Immun. 71, 5130-5138. Wells, S.J., Ott, S.L., Hillberg Seitzinger, A., 1998. Key health issues for dairy cattle — new and old J. Dairy Sci. 81, 3029-3035. Wells, S.J., Wagner, B.A., 2000. Herd-level risk factors for infection with Mycobacterium paratuberculosis in US dairies and association between familiarity of the herd manger with the disease or prior diagnosis of the disease in that herd and use of preventive measures. J Am Vet Med Assoc. 216, 1450- 1457. Wells, S.J., Whitlock, RH., Lindeman, C.J., Fyock, T., 2002a. Evaluation of bacteriologic culture of pooled fecal samples for detection of Mycobacterium paratuberculosis. Am J Vet Res. 63, 1207-1211. Wells, S.J., Whitlock, R.H., Wagner, B.A., Collins, J., Garry, F., Hirst, H., Lawrence, J ., Saville, W.J.A., Larew-Naugle, A.L., 2002b. Sensitivity of test strategies used in the Voluntary J ohne’s Disease Herd Status Program for detection of Mycobacterium paratuberculosis infection in dairy cattle herds. J Am Vet Med Assoc. 220, 1053-1057. Wells, SJ, Godden, S.M., Lindeman, C.J., Collins, J .E., 2003. Evaluation of bacteriologic culture of individual and pooled fecal samples for detection of Mycobacterium paratuberculosis in dairy cattle herds. J Am Vet Med Assoc. 223, 1022-1025. Wentink, G.H., Bonger, J.H. Zeeuwen, A.A., Jaartsveld F.H., 1994. Incidence of paratuberculosis after vaccination against M. paratuberculosis in two infected dairy herds. J Vet Med Ser B. 41, 517-522. 279 Whan, L., Grant, I.R, Ball, H.J., Scott, R, Rowe, M.T., 2001. Bactericidal effect of chlorine of Mycobacterium paratuberculosis in drinking water. Lett Appl Microbiol. 33, 227-231. Whipple, D.L., Kapke, P.A., Andersen, RR, 1992. Comparison of a commercial DNA probe test and three cultivation procedures for the detection of Mycobacterium paratuberculosis in bovine feces. J Vet Diagn Invest. 4, 23-27. Whitlock, RH, 1992. Diarrhea in Cattle: Johne’s disease. In: Anderson, N.V. (Ed), Veterinary Gastroenterology. 2"(1 Ed. Philadelphia, pp. 783-789. Whitlock, RH., Buergelt, C., 1996. Preclinical and clinical manifestations of paratuberculosis (including pathology). Vet Clin North Am Food Anim Pract.. 12, 345-356. Whitlock, RH., Wells, S.J., Sweeney, RW., Van Tiem, J ., 2000. ELISA and fecal culture for paratuberculosis (Johne’s disease): sensitivity and specificity of each method. Vet Microbiol. 77, 387-398. Whitlock, R.H., Sweeney, RW., Fyock, T., McAdams, S., Gardner, I.A., McClary, D.G., 2005. Johne’s disease: the effect of feeding monensin to reduce the bioburden of Mycobacterium avium subspecies paratuberculosis in neonatal calves. 2005. Proc 38tb Annu Conv Am Assoc Bov Prac. pp. 191- 192. Whiltock,.R.H., Gardiner, I.A., Mangold, B.L., Smith, J., Sweeney, RW., Schukken, Y., 2006. Johne’s disease: Mycobacterium paratuberculosis super-shedders: detection and contribution to passive shedding (false-positive fecal cultures). Proc 39th Annu Conv Am Assoc Bov Prac. 286-288. Whittington, RJ., Sergeant, E.S.G., 2001. Progress towards understanding the spread, detection and control of Mycobacterium avium subsp paratuberculosis in animal populations. Aust Vet J. 79, 267-278. Whittington, RJ., Marsh, I.B., Taylor, P.J., Marshall, D.J., Taragel, C., Reddacliff, L.A., 2003. Isolation of Mycobacterium avium subsp paratuberculosis from environmental samples collected from farms before and after destocking sheep with paratuberculosis. Aust Vet J. 81, 559-563. Whittington, RJ., Marshall, D.J., Nicholls, P.J., Marsh, 13., Reddacliff, L.A., 2004. Survival and dormancy of Mycobacterium avium subsp. paratuberculosis in the environment. Appl Environ Microbiol. 70, 2989-3004. Whittington, RJ., Windsor, P.A., 2007. In utero infection of cattle with Mycobacterium avium subsp paratuberculosis. Proc 9th Int Coll Paratuberculosis. pp. 112. 280 Williams, E.S., Snyder, S.P., Martin, KL, 1983. Experimental infection of some North American wild ruminants and domestic sheep with Mycobacterium paratuberculosis: clinical and bacteriological findings. J Wildl Dis. 19, 185- 191. Wilson, D.J., Rossiter, C., Han, H.R., Sears, P.M., 1993. Association of Mycobacterium avium paratuberculosis infection with reduced mastitis, but with decreased milk production and increased cull rate in clinically normal dairy cows. Am J Vet Res. 54, 1851-1857. Wood, P.R., Comer, L.A., Plackett, P., 1990. Development of a simple, rapid in vitro cellular assay for bovine tuberculosis based on the production of y—interferon. Res Vet Sci. 49, 46-49. Zubrick, B.G., Czuprynski, C.J., 1987. Ingestion and intracellular growth of Mycobacterium paratuberculosis within bovine blood monocytes and monocyte- derived macrophages. Infect Immun. 55, 1588-1593. Zubrick, B.G., Follett, D.M., Czuprynski, C.J., 1988. Cytokine regulation of the intracellular growth of Mycobacterium paratuberculosis in bovine monocytes. Infect Immun. 56, 1692-1697. 281 3 03062 4385 12 Amg