L _ 1. — . .: surfing“: .3»;me r: .. «flow. tr. .. in.“ . . . a ‘ Ar. :9. .3 2:: :27 3'1. 7 LIN .3 .‘1 maimnnfiz 4.1.1"; 3.4.”? .5. ‘3 1 my}: _ l . ‘1 ski. .. 1: exit «M. 3% uxuifl 1.. a . A. t: > T .. 9mg“. «.5 .Q. .L . i 4 I}? Jams... . r. 9.5 S ‘ itxi. 1. $3.3 1 {a ..: urnmrxn. v i}: .. , .. sags: :: A... ::..Am...;. i. é : ; :- jag r. . n. . a, fifié manure In. m‘ 6% mm‘fivs .4 y u 1 ..‘ (I. 2 II 5.. P1, .1‘ I|C:.(. . 13.11: 5.». 3n » THUNS "I '9'?" n7 l This is to certify that the dissertation entitled The Effect of Bovine Somatotropin on Disease, Reproduction, Longevity, and Milk Production Economics on Four Michigan Farms presented by Lawrence J. Judge has been accepted towards fulfillment of the requirements for Ph.D. Large Animal Clinical Sciences WW Major professor degree in r/xv a, MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771 Luann—av Michlgan State Unlvorslty 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. QiiifEtDGEgOWI DATE DUE DATE DUE 1004 07 OCT09200 ”$10333? 6101 chlRC/DateDuepes-sz THE EFFECT OF BOVINE SOMATOTROPIN ON DISEASE, REPRODUCTION, LONGEVITY AND MILK PRODUCTION ECONOMICS ON FOUR MICHIGAN FARMS By Lawrence James Judge A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Large Animal Clinical Sciences 2001 ABSTRACT THE EFFECT OF BOVINE SOMATOTROPIN ON DISEASE, REPRODUCTION, LONGEVITY AND MILK PRODUCTION ECONOMICS ON FOUR MICHIGAN FARMS By Lawrence James Judge Holstein cows .(n = 555) from four Michigan dairy farms were sequentially assigned to receive bovine somatotropin (bST) or remain as untreated controls. Bovine somatotropin (500 mg) was administered every 14 days beginning at 63 to 69 days of lactation, continuing until approximately 21 days prior to end of lactation or until the cow was removed from the herd. Objectives were to determine the effect of bST on disease incidence, reproductive function, cow longevity, Dairy Herd Improvement Association production estimates and profitability of dairy cows. A total of 127 (22.9% lactation incidence) cases of. clinical mastitis occurred during lactation with 42 (33.1%) cases in the pretrial (< 63 days of lactation (prior to bST treatment» with 57.1% occurring in controls and 42.9% in bST-treated cows and 85 (66.9%) cases occurring during the remainder of lactation (during bST treatment) with 47.1% in occurring in controls and 52.9% in bST-treated cows. The logistic regression odds ratio for bST-treatment on clinical mastitis was 1.06 (95% CI: 0.62 to 1.81). The number of days milk was discarded following therapy for mastitis did not differ between the study treatment groups. A total of 229 cows (42.57 % lactation incidence) comprised of 112 (48.91%) controls and 117 (51.09%) bST-treated cows were considered reproductive failures by 150 days of lactation and 92 cows (17.10% lactation incidence) comprised of 45 (48.91%) controls and 47 (51.09%) bST-treated cows at the end of lactation. The bST-treated cows had more nonspecific off feed conditions during the study lactation. Twin conceptions occurring during the study lactation were significantly associated with bST treatment. Logistic regression odds ratios for bST treatment effect on reproductive failure at 150 days and the end of lactation were 0.95 (95% CI: 0.67 to 1.36) and 0.95 (95% CI: 0.59 to 1.54), respectively. Correlation was high (r > 0.97) between actual milk produced and Dairy Herd Improvement Association milk production projections at 305 days of lactation for both Study treatment groups. Production response to bST did not maximize until the middle of the 238-day bST treatment period. Both amount of, and amount of days to, peak milk production were different between the study treatment groups and bST- treated cows were more persistent in lactation. Projections of milk production by Dairy Herd Improvement Association underestimated actual 305-day production of bST- treated cows at the first and second monthly tests after bST initiation by greater amounts compared to control cows, but this difference was not significant by the third test date nor by 305 days of lactation. Life tables analyzed by log-rank testing found no difference in survival time between the study treatment groups. More control cows were culled for reason of low milk production during the study lactation and for reason of injury during the first 60 days of the next lactation. The hazard ratio for bST treatment for the Study and the first 60 days of the subsequent lactation was 0.93 (95 % CI: 0.68 to 1.27) and 0.93 (95% CI: 0.53 to 1.62), respectively. This work is dedicated to my family, especially my children, Ann Marie and Laura Elizabeth. They didn’t often complain when Daddy frequently told them he had “to go to school.” 'I hope that they grow to view the process of education as I do: A challenging and demanding, yet joyful process by which knowledge and understanding of the world we live in is gained. iv ACKNOWLEDGMENTS While it is impossible to mention everyone who. helped in the completion of this work, I would like to take this opportunity to recognize the major contributors whose time and efforts are greatly appreciated. First of all, my graduate committee, consisting of Drs. Paul Bartlett (major professor), Ron Erskine (acting chair, 1996- 1997), Jim Lloyd, Phil Sears and Roy Fogwell, deserve special recognition. They all took the time to help me understand things that I did not know before, often with very little personal. reward. They did so, I’m sure, because they love to teach and cared that I learned tobe a scientist; I am lucky to be the beneficiary of their efforts. Dr. Pam Ruegg deserves specific mention for both assistance in securing the original grant that began this project as well as her willingness to serve as my external examiner. Also, I would like to thank Chris Phipps, Therese Burns, and Mark Ormel for both their assistance with on-farm data collection as well as entering and processing data. Jason Vanlente and Brian Prescott spent many hours processing milk weights and entering other data. Nichol Zarb assisted with the processing of calving data. Without the help of these good people, I would still be staring at a computer screen, processing the over half a million individual milk weights, working with the over three thousand milk samples that were collected, or processing other data that was so vital to this project. I am also appreciative of the financial assistance that I received while working on this project. The monetary support of Drs. Bartlett and Erskine, the Department of V Large Animal Clinical Sciences, the Michigan Agricultural Initiative, the Sterner Grant Fund and the Robert Schirmer Grant Fund were greatly appreciated and timely. Without this support, this project simply would not have been possible. Finally, I would like to acknowledge my family’s contribution to this work. My daughters, Ann and Laura, both sacrificed and assisted more than they will ever know with “Daddy’s school work.” My parents assisted in my ways, from working on the remodeling of my house, to moving my household and in other ways to numerous to mention. My brothers and sisters all offered encouragement as I made this change of direction for my career. I hope that all who offered support and assistance to me of any kind though this eight-year period understand the deep gratitude I have for their unselfish efforts and I hope that they feel their efforts were both gratifying and worthwhile. vi TABLE OF CONTENTS LIST OF TABLES xii LIST OF FIGURES xvii LIST OF ABBREVIATIONS ‘ xix CHAPTERI: Introduction, the Problem and General Methods Introduction 1 Pr'oblem statement 2 Study objectives 7 Study benefits 7 General methods 8 StUdy design 8 Selection of farms 8 Selection of cows 9 Management of data 10 References 12 vii CHAPTER 2: Literature Review Clinical mastitis Reproductive effects Longevity DHIA lactation projections Economic impact References CHAPTER 3: Recombinant Bovine Somatotropin and Clinical Mastitis: Incidence and Days of Milk Discarded due to Therapy Abstract Introduction Specific methods Results Mastitis Discarded milk Discussion Conclusion References viii 13 15 20 21 22 25 31 32 33 36 4O 41 45 46 CHAPTER 4: Recombinant Bovine Somatotropin: Association with Reproductive Performance in Dairy Cows Abstract Introduction Specific methods Results Discussion Conclusion References CHAPTER 5: Survival Analysis of the Effect of Bovine Somatotropin on Death and Culling of Dairy Cows Abstract Introduction Specific methods Results Bivariate analysis of removal from the herd Survival measured by lifetable analysis Reasons for removal from the herd Survival analyzed by proportional hazards multivariate modeling Discussion Conclusion References 49 51 53 57 62 67 69 75 77 79 82 84 91 97 102 103 CHAPTER 6: The Effect of Recombinant Bovine Somatotropin on Milk Production and Lactation Milk Estimates for Cows Milked Three Times per Day Abstract 106 Introduction 108 Specific methods - 110 Results Pretrial milk production 112 Patterns of milk production 113 Correlation and accuracy of milk production measures 121 Discussion 124 Conclusion 130 References 131 CHAPTER 7: The Impact of Recombinant Bovine Somatotropin on Net Farm Income: Analysis of Four Michigan Dairy Farms Abstract 134 Introduction 136 Specific methods . 138 Results Pretrial milk production 143 Disease and reproductive measures 142 Milk production measures 144 Financial measures 146 Discussion 1 153 Conclusion . 161 References I , 163 SUMMARY 167 CONCLUSIONS - . 170 APPENDIX A 177 BIBLIOGRAPHY 179 xi LIST OF «TABLES TABLE PAGE TABLE 3.1: Incidence of clinical mastitis (cases per 100 cow-months) 37 in dairy cows administered 500 mg of bST (PosilacR; Monsanto Company, St. Louis, M0) at 14-day intervals and in untreated controls. TABLE 3.2: Milk culture results (duplicate quarter samples) at 30 to 60 days 38 of lactation (pretrial period) in dairy cows later administered 500 mg of bST (PosilacR; Monsanto Company, St. Louis, M0) at l4-day intervals and in untreated controls. TABLE 3.3: Milk culture results (single quarter samples) from clinical 39 mastitis cases in dairy cows administered 500 mg of bST (PosilacR; Monsanto Company, St. Louis, M0) at 14-day intervals and in untreated controls between approximately 63 days of lactation and the end of lactation. TABLE 3.4: Logistic regression analysis results of clinical mastitis in 40 dairy cows administered 500 mg of bST (PosilacR; Monsanto Company, St. Louis, M0) at 14-day intervals and in untreated controls. TABLE 4.1: Reproductive parameters for cows administered 500 mg 58 of bST at 14-day intervals from 63 days of lactation until approximately 21 days before the end of lactation and in untreated controls. TABLE 4.2: Cows with a disease occurring at anytime during lactation 60 when administered 500 mg of bST at 14-day intervals from 63 days of lactation until approximately 21 days before the end of lactation and in untreated controls. TABLE 4.3: Logistic regression analysis of reproductive failure at 61 150 days of lactation for cows administered 500 mg of bST at l4-day intervals from 63 days of lactation until approximately 21 days before the end of lactation and in untreated controls. Table 4.4: Logistic regression analysis of reproductive failure by the 61 end of lactation for cows administered 500 mg of bST at 14-day intervals from 63 days of lactation until approximately 21 days before the end of lactation and in untreated controls. xii TABLE Table 5.1: Analysis of culling, deaths and removal from the herd (culling plus deaths = total) for primiparous, multiparous and all study cows from calving until the end of lactation for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls (number of cows removed (percent of treatment group removed)). TABLE 5.2: Analysis of culling, deaths and removal from the herd (culling plus deaths = total) for primiparous, multiparous and all study cows during the first 63 days of the subsequent lactation for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at l4-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls (number of cows removed (percent of treatment group removed)). TABLE 5.3: Life table analysis of removal from the herd for primiparous cows from calving until the end of lactation for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co._. St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls. - TABLE 5.4: Life table analysis of removal from the herd for multiparous cows from calving until the end of lactation for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at l4-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls. . TABLE 5.5: Life table analysis of removal from the herd for all study cows from calving until the end of lactation for all study cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at 14—day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls. TABLE 5.6: Reasons for removal from the herd for all study cows from calving until the end of lactation for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls (number of cows removed (percent of total cows removed)). xiii PAGE 82 83 85 86 9O TABLE TABLE 5.7: Reasons for removal from the herd for all study cows during the first 63 days of the subsequent lactation for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls (number of cows removed (percent of total cows removed)). Table 5.8: Cox proportional hazards modeling of removal from the herd for all study cows from calving until the end of lactation for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls. TABLE 5.9: Cox proportional hazard modeling of removal from the herd for all study cows during the first 63 days of the subsequent lactation for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls. TABLE 6.1: Pretreatment (from 4 to 63 days of lactation) milk production estimates for actual milk produced for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at l4-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows (kg / first 63 days of lactation). TABLE 6.2: Peak milk production and time to peak milk production for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co. , St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end. of lactation compared to untreated control cows. TABLE 6.3: Correlation (Pearson’s) of the first three (after bST treatment initiation) and final (305-day) DHIA milk production estimates with actual milk produced (daily milk weight measurements) for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows. xiv PAGE 91 112 119 120 TABLE PAGE TABLE 6.4: Differences in milk production between the first three (after 122 bST treatment initiation) and final (305-day) DHIA milk production estimates and actual milk produced (measured by daily milk weight measurements) for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co. , St. Louis, M0) at l4-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows. TABLE 7.1: Feed prices (during 1994-95) used to calculate costs for 139 cows supplemented with 500 mg of bST (PosilacR, Monsanto Co. , St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows. TABLE 7.2: Average daily milk production of cows supplemented with 144 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at 14—day intervals beginning at approximately 63 days of lactation until 362 days of lactation compared to untreated control cows (mean kg per cow per day). TABLE 7.3: Partial budget analysis for cows supplemented with 500 mg 146 of bST (PosilacR, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until approximately 301 days of lactation compared to untreated control cows (5 per supplemented period per cow). TABLE 7.4: Partial budget analysis for cows supplemented with 500 mg 147 of bST (PosilacR, Monsanto Co., St. Louis, M0) at l4-day intervals beginning at approximately 63 days of lactation until approximately 301 days of lactation compared to untreated control cows ($ per supplemented period per cow). TABLE 7.5: Break-even analysis for cows supplemented with 500 mg 148 of bST (PosilacR, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until 301 days of lactation compared to untreated control cows assuming costs of $0.084 per Mcal for additional feed, $5.86 per dose of bST and a $28.09 per 100 kg milk price (per bST-treated cow). XV TABLE PAGE TABLE 7.6: Results of sensitivity analysis for several variables for cows 150 supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until 301 days of lactation compared to untreated control cows assuming $0.084 per Mcal for additional feed and $5.86 per dose of bST. TABLE 7.7: Financial analysis of twin calvings for cows supplemented 152 with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, M0) at l4-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows xvi LIST OF FIGURES FIGURE PAGE FIGURE 5.1: Survivorship curve for primiparous cows from calving until 87 the end of lactation for cows supplemented with 500 mg of bST b) (PosilacR, Monsanto Co., St. Louis, M0) at l4-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls (—). Vertical lines on x-axis denote median survival times. FIGURE 5.2: Survivorship curve for multiparous cows from calving until 88 the end of lactation for cows supplemented with 500 mg of bST (—) (PosilacR, Monsanto Co., St. Louis, M0) at l4-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls (—-). Vertical lines on x-axis denote median survival times. FIGURE 5 .3: Survivorship curve for all study cows from calving until 89 the end of lactation for cows supplemented with 500 mg of bST (—) (PosilacR, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls (-—). Vertical lines on x-axis denote median survival times. FIGURE 6.1: Variability (standard deviation) in daily milk production 113 within each of 17 injection cycles for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, MO) (A) at l4-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows ( 0). Star indicates standard deviation of milk production differed between the study groups (P < 0.05). FIGURE 6.2: Daily milk production during the first three of 17 injection 114 cycles for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co. , St. Louis, MO) (A) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows ( O) . Vertical lines indicate day of bST administration. FIGURE 6.3: Milk production on the first day of each of 1? injection cycles 115 for cows supplemented with 500 mg of bST (PosilacR, Monsanto Co. , St. Louis, MO) (A) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows ( O). xvii FIGURE PAGE FIGURE 6.4: Actual (five-day interval average) milk production for 1 l6 primiparous cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, MO) (A) at l4-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows. FIGURE 6.5: Actual (five-day interval average) milk production for 117 multiparous cows supplemented with 500 mg of bST (PosilacR, Monsanto Co., St. Louis, MO) (A) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows ( O) . Vertical line indicates initiation of bST treatment. - FIGURE 6.6: Actual (five-day interval average) milk production for all cows 118 supplemented with 500 mg of bST (Posilack, Monsanto Co., St. Louis, MO) (A) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows ( 0). Vertical line indicates initiation of bST treatment. xviii LIST OF ABBREVIATIONS bST = bovine somatotropin CI = confidence interval CM = clinical mastitis DHIA = Dairy Herd Improvement Association FDA = Food and Drug Administration FSH = follicle stimulating hormone IGF-l = insulin-like growth factor one IMI = intra-mammary infection LH = luteinizing hormone NAI-IMS = National Animal Health Monitoring System NFI = Net Farm Income SA = survival analysis SCC = somatic cell count TMR = total mixed ration xix CHAPTER ONE INTRODUCTION, THE PROBLEM AND GENERAL METHODS INTRODUCTION In 1937, Asimov and Krouze found enhanced milk production in cows that were injected with extracts of bovine pituitary glands. In subsequent years, other workers discovered that the hormone responsible for this effect was bovine growth hormone, also called bovine somatotropin (bST). However, this finding was of little practical significance because of limited hormone-supplies. Then, in the 1970's, recombinant DNA (deoxyribose nucleic acid) technology made possible mass production of bST. In the 1980's, several pharmaceutical companies began studies in an attempt to gain governmental approval to manufacture and market a bST product. In late 1993 , the Monsanto Company was granted approval by the FDA to market their bST product, PosilacR', which is currently the only approved bST product in the US. PosilacR is designated for use in lactating dairy cows as a galactopoetic agent. AS a result of being the first product produced for use in animal agriculture using biotechnology, bST has undergone intense scientific and public scrutiny. Concern has been raised over the potential for bST to increase the incidence of disease, particularly mastitis, which would consequently increase antibiotic use for therapy of affected cows. If such effects exist, aside from increasing the risk of chemical (e.g. , antibiotic) residues in food products, they could also reduce the reported positive effect bST has on farm profits. Also at issue are potential negative effects which bST may 1 have on reproductive performance and cow longevity. Consequently, despite FDA approval, many questions require more complete answers in order to more accurately determine the impact of this technology on dairy cows, the food supply and farm economics. PROBLEM STATEMENT Despite the intensive preapproval study of bST, several questions remain to be answered. Preapproval clinical studies required by the FDA are designed primarily to address animal safety and efficacy (Code of Federal Regulations, 21 CFR 514, Part 514.50(d)(2 and. 3)). AS such, these studies are required to demonstrate that a new product does what the company claims it will do, and at the same time, does no harm to the animals. "The FDA's primary mission is to assure that approved products are safe for both the animals they are used on as well as humans consuming products originating from animals that have been treated with the product. In this regard, most preapproval clinical studies conducted with bST were primarily designed to evaluate the effect of different dosages on efficacy (milk production enhancement) and cow safety (clinical toxicologic signs). The intensive nature of such studies imposed practical limits on the number of cows that could be utilized. Therefore, many preapproval studies utilized a randomized block deSign to increase the probability of detecting significant treatment effects in small numbers of cows. While this design increased the efficiency of the studies, it reduced the ability of the study to detect any statistically significant change in the incidence of relatively rare diseases (Pellet al. , 1992) or reproductive efficiency 2 (Cole et al., 1992) resulting from bST treatment. For example, a minimum of 500 cows (with 250 in the treatment group and 250 in the control group) is the required sample size to detect a 15 % difference in the rate of clinical mastitis between study treatment groups, given a clinical mastitis incidence of 30% in the study population (significance level of 95% and a study power of 80%). Few preapproval studies were able to involved similar numbers of cows. However, any disease effects associated with bST are important in determining if bST will be a successful product. In an attempt to increase statistical power, reviewers of pre-approval studies have conducted meta—analysis to examine the effects of bST on clinical mastitis (Canadian Expert Panel, 1998; White et al., 1994; Willeberg, 1993). However, the combining the results of several studies may result in the introduction of several biases. Examples of such biases include publication bias (the results of positive studies are published more frequently than negative ones) and selection bias (the selection of study subjects (either cows or herds) is different between the studies analyzed) (LeLorier et al. , 1997). Blocking was used in many pre-approval studies in an attempt to control for “nuisance” variables (e.g. , parity .and level of milk production at treatment initiation) which were related to the study endpoints (outcomes). However, by using this technique, the effects of the blocking variables on the outcomes were not able to be determined. Furthermore, in studies where each treatment level is represented only once in each block, it is impossible to determine any interaction which may have existed between the treatment and the blocking variables. It is important determine 3 how these variables effect the study outcomes before an accurate assessment of bST effects can be made. To expand the review of bST beyond the primary objectives of efficacy and cow safety, a Post Approval Monitoring Program was conducted by the Monsanto Company. This program was a clinical study comprised of 1,213 cows located in 28 herds distributed across the US, and was the first study of its kind to be required by the FDA. With this single large data base gathered under commercial conditions, there were sufficient numbers of cows and disease data to allow the accurate determination of disease and reproductive effects. However, adequate data for the analysis of the economic impact (such as accurate milk yield data) of the observed effects was not collected. The data required for accurate determination of the economic impact of bST includes daily, individual-cow milk weights. This is because the available bST product is a 14—day sustained release preparation which causes variability in the milk production response through the l4-day injection cycle. Milk production following bST treatment gradually rises and peaks on approximately day 8 to 9 post-injection. Production then returns to near pre-treatment levels by approximately day 12 to 13 post-treatment. Therefore, the day on which milk production is measured, during the 14—day injection cycle, may result in a different level of production enhancement being found. Because few farms have daily, individual-cow milk weight technology, most farms currently estimate production enhancement from bST using a single, monthly milk production measurement (by use of DHIA milk measurements). This may result in a significant 4 amount of error in the estimation of the effect that bST has on the economics of milk production. Economic feasibility is not a requirement for FDA approval of pharmaceutical products. The economic aspects of a new product are determined by the market itself. Therefore, if producers perceive that a product will have a positive effect on profitability, then the product will be adopted. This "trial. and error" system is not optimal. However, without research aimed at addressing the economic efficiency of new products, this is the only way that such technologies will be "tested. " In the absence of such research, the economic estimates made using the data generated in the preapproval studies often becomes the basis for the reported economic viability of the product (Shoeffling et al., 1991; Elbehri and Yonkers, 1995). However, these estimates may not be accurate because they are based on data that were not designed to address economic issues nor were the data collected derived from commercial-type dairy Situations. Also, current estimates of enhanced profitability from bST use only account for three costs: 1) increased milk production, 2) increased feed consumption, and 3) labor to administer the product. If there are other costs of using bST, such as reductions in reproductive efficiency or increased disease rates, these costs would need to be subtracted from any gains to accurately reveal changes in net farm income (profitability) from bST use. While the issue of cow safety is addressed adequately in preapproval studies, the related issue of cow longevity is not. The close relationship of these two issues 5 (longevity and safety) should make them inseparable issues. However, preapproval studies require that cows be removed from study only if they are either determined to be a health risk to herdmates or die themselves. Because culling is not an issue under this format, any determination of cow longevity as it would occur in commercial dairy herds cannot be made from such data. Therefore, it is still unclear as to whether bST significantly affects cow longevity. An estimate of cow longevity is needed because of the importance of this factor in determining profitability of cows (Beaudeau et al. , 1995). Galton et al. (1997) have suggested that bST may have a positive effect on cow longevity because fewer culls because of low milk production likely occur in bST- treated cows. However, bST may increase the incidence of some diseases or reduce reproductive performance, thereby increasing the rates of culling in affected cows. These two factors may balance each other so that no difference would be observed between bST-treated and untreated cows. However, even small differences in culling rates could dramatically impact the profitability of bST use because of the large farm expense incurred from raising or purchasing replacement animals. In summary, despite an unprecedented amount of scrutiny, the effect that bST may have on disease incidence, reproduction, longevity and milk production economics in commercial dairy herds remains unclear. Through the conduct of this research, knowledge in these areas will hopefully be expanded and clarified. Study objectives 1) To determine the effect of bST on the incidence of clinical mastitis and days of milk productionwithheld resulting from therapy of these cases. 2) To investigate whether bST is associated with diseases affecting reproduction and if reproductive performance is affect by use of bST. 3) To measure the effect of bST on cow longevity (culling). 4) To determine the accuracy of DHIA lactation production estimates in bST- treated cows. 5) To assess the farm-level economic impact of bST use in dairy cows. Study benefits This study will allow dairy producers and industry professionals to have a better estimate of the effect that bST has on the dairy cows and farms with which they work. While it is acknowledged that‘the results of this study will not apply to all farms, it is hoped that the methods used in the generation of these results will be used to make similar estimates in other herds and studies. This will result in more accurate estimates of the impact that bST has on: 1) diseases which affect milk quality (e.g., mastitis), 2) reproductive efficiency, and 3) cow profitability. Through the use of daily, individual-cow milk weight technology, the milk 7 production data collected will allow the accurate measurement of milk production enhancement from bST use. Application of the knowledge gained from this research will allow producers to make better decisions and industry advisors to offer improved advise regarding use of bST. GENERAL METHODS Study design This study was designed as a clinical trial (experiment). As such, cows were assigned by a random sequential method to a study treatment group. No blocking or matching was performed for any variable in this study. Cows in the treatment group received 500 mg bovine somatotropin in a sustained release preparation ((Posilac“; Monsanto Company, St. Louis, MO) injected subcutaneously at 14—day intervals. Cows that were assigned to the control group that did not receive any injections. Selection of farms Three commercial dairy herds and one university demonstration herd in Michigan participated in the study. Herd sizes were approximately 120, 250, 250, and 600 milking cows. Herds were selected on the following criteria: 1) mean bulk tank SCC < 300,000 cells/ml for the 12 months before the trial, 2) practice of a mastitis control program that included pre- and postmilking teat dipping along with dry cow treatment of all cows, 3) establishment of a Standardized course of therapy for cases of CM, 4) willingness to comply with the experimental protocol, 5) availability of a 8 computerized system for recording events and daily individual milk weights from the cows, and 6) enrolhnent in the Michigan DHIA. All farms housed lactating cows in free stalls, fed cows a TMR that was balanced to meet the nutritional requirements of their current milk yield, and milked all cows three times per day. Veterinary care at all farms consisted of routine herd health visits on a monthly basis and clinical disease treatments when necessary. Herdsmen diagnosed, treated and reported cases of disease when they were knowledgeable and confident of the underlying disease process and capable of administering appropriate treatment; otherwise, diseases were diagnosed and treated by a veterinary practitioner. Selection of cows All cows in this study calved during an approximately 6-month period on all farms (between June, 1994 and January, 1995). Therefore, the study population consisted of approximately 50% of the cows present in each herd. In the ninth week of lactation, cows in each herd were assigned to either the bST treated or control study group. For three farms, an alternate week assignment was used; all cows calving during a given week were assigned to the treated group, and all cows calving the next week were assigned to the control group. The remaining farm assigned treatments on an alternate cow basis. Injections of bST were administered by herd personnel; control cows received no injections. Cows that were assigned to bST treatment and that were determined by herd managers to be in poor health at the time to initiate treatment and did not receive bST (per label directions). Similarly, cows that were assigned as 9 controls were excluded when health conditions would have prevented them from receiving bST had they been assigned to the treatment group. As Specified by bST labeled usage, cows were excluded from the Study if they were in poor health, as assessed by the herd managers at ninth week of lactation. This exclusion criterion was applied equally to both the control and bST treatment groups. Cowstwere considered in poor health if they had a clinical disease or severe loss of body condition. Of 580 cows that were initially selected, 25 cows (10 selected as controls and 15 selected as bST—treated cows) were excluded from the study. Twenty- two of these cows were excluded because of poor health, while three cows were excluded because of clerical errors unrelated to their .. health status. Of the 22 cows excluded because of poor health, 10 were controls and 12 were from the to be bST- treated grOUp. Specifically, four control and five to be bST-treated cows were excluded due to poor udder health; five control and five to be bST-treated cows were excluded due to metritis, ketosis, displaced abomasum or lameness; and one control and two treated cows were excluded due to other health conditions which resulted in low body condition. Therefore, 555 cows were utilized in the final analysis. Management of data All farms used the dairy herd management software DairyComp 305R (Valley Agricultural Software, 1990) for daily herd activities and the spreadsheet program ExcelR (Microsoft Corporation, 1993) was used by the investigators to manage the data. Recorded data included calving dates, artificial inseminations, results of veterinary 10 reproductive examinations, daily, individual-cow milk weights, disease events as well as disease and reproductive treatments. Some of these data were captured electronically while various data were recorded on paper records. Data were collected for each case of disease and farm personnel recorded the drugs that were used for therapy of these conditions by documenting drug name, dose, route of administration, and days that milk was withheld from sale. The investigators assisted in updating records during routine monthly visits and contacted participating producers on a weekly basis during the experiment. Milk weight files were generated by each farm weekly and then downloaded by the investigators into Excel“. Each file was scanned for missing weights (approximately 5 % of. all milkings) which were then estimated by averaging the corresponding data (i.e., the first, second or third milking) from before and after the missing value. After all missing data were estimated, the three individual milkings for each cow were summed to produce a daily total for each cow. Approximately 20% of the cows were already in early lactation when the study began and consequently some early milk weight data were missing for these cows. This lack of data was present for three of the four study herds. These missing data were extrapolated by constructing lactation curves for each farm using the milk weight data from study cows on the same farm for which complete lactation milk weights were available. Three separate curves were constructed for each farm by dividing the existing data into high, medium and low production groups based on total production that had occurred by 63 days of lactation. ll REFERENCES Beaudeau, F., V. Ducrocq, C. Fourichon, and H. Seegers.1995. Effect of disease on length of productive life of French Holstein dairy cows assessed by survival analysis. J. Dairy Sci. 78:103-117. Canadian Veterinary Medical Association Expert Panel. 1998. Report of the Canadian Veterinary Medical Association Expert Panel on rBST. Ottawa, CA. Dairy Comp 305“, Version 3.1. 1990. Valley Agricultural Software. Tulare CA. Elbehri, A., and R. D. Yonkers. 1995. Economic analysis of the impacts of bovine somatotropin on the profitability of representative dairy farms in the northeast. Ag. and Res. Econ. Rev. 4:88-100. Excel“, Version 5.0. 1993. Microsoft Corporation, Redmond, WA. Galton, D. M., R. W. Everett, M.‘ E. Van Amburgh, D. L. Bauman, and W. A. Knoblauch. 1997. Extended calving intervals with the use of bST. Presented at the Tri-State Dairy Nutrition Conference, Fort Wayne, IN. LeLorier, J ., G. Gregoire, A. Benhaddad, J. Iapierre, and F. Derderian. 1997. Discrepancies between meta-analyses and subsequent large randomized, controlled trials. The New England Journal of Medicine. 337(8):536-542. Pell, A. N., D. S. Tsang, B. A. Howlett, M. T. Huyler, V. K. Meserole, W. A. Samuels, G. F. Hartnell, and R. L. Hintz. 1992. Effects of a prolonged-release formulation of sometribove(n-methionyl bovine somatotropin) on Jersey cows. 1 J. Dairy Sci. 75:3416-3431. PosilacR product label. 1993. Monsanto Co., St. Louis, MO. Shoeffling, J. R., R. C. Angus, D. V. Armstrong, and J. T. Huber. 1991. Economic implications of bovine somatotropin use for the Arizona dairy industry. J. Dairy Sci. 74:2347-2352. White, T. C., et. al. 1994. Clinical mastitis in cows treated with sometribove (recombinant bovine somatotropin) and its relationship to milk yield. J. Dairy Sci. 77:2249-2260. Willeberg, P. 1993. Bovine somatotrOpin and clinical mastitis: epidemiological assessment of the welfare risk. Livest Prod Sci. 36:55-66. 12 CHAPTER TWO LITERATURE REVIEW Clinical mastitis Much concern has been raised and an equally greater amount written in regard to the potential effects of bST on milk quality. The approval of bST (PosilacR; Monsanto Company, St Louis, M0) for lactating dairy cows has caused considerable scientific and public controversy over potential effects this product might have on the incidence of CM (Kronfeld, 1994). The states of Wisconsin and Vermont have several dairies companies so fearful of this issue that they have asked producers to sign contracts stating they will not use this product. Part of this fear is because any increase in the number of clinical cases that may be associated with bST use might-necessitate increased use of antibiotics for therapy of these cases (Kronfeld, 1994) . The treatment of CM is a primary reason for the use of antimicrobials in dairy cattle. Therefore, technologies which increase CM incidence may V concurrently increase antibiotic usage and the risk of milk residues. hi addition to milk quality and antibiotic residue concerns, mastitis is an economically important disease to the dairy industry. Costs associated with mastitis have been estimated at approximately $100.00 per clinical case (Hoblet et al., 1991; Miller et al. , 1993). The risk of CM has been found to increase as milk production increases (Oltenacu and Ekesbo, 1994) with a concurrent decrease in milk production 13 following disease onset (Bartlett and Van Wijk, 1991; Firat, 1993; Oltenacu and Ekesbo, 1994). Therefore, technologies that enhance milk production may be positively correlated with CM and thereby increase financial losses associated with this disease. Of nine trials in which 500- to 700-m’g doses of bST were administered at 14-day intervals, seven studies concluded that the incidence of CM did not differ between cows that were administered bST and untreated control cows (Downer et al., 1993; Eppard et al., 1991; Remond et al., 1991; Rijpkema etal., 1990; Skarda et al., 1992; Thomas et al., 1991; Whitaker et al., 1988). In two trials, a significant increase in CM was reported for bST-treated cows compared with control cows. However, in these trials, cows receiving bST had a significantly greater incidence of CM than did control cows before bST treatment was initiated which might have affected outcomes of the studies (Cole et al., 1992; Pell et al., 1992). In a review of 15 full lactation studies conducted in the US and EuTOpe, 467 lactation records of cows supplemented with bST and 447 lactation records of control cows were summarized (White et al., 1994). A significantly higher rate of CM was observed for the cows that had been treated with bST. However, a significantly higher rate of CM was also apparent during the pretreatment period for cows that were later treated with bST. Thus, the higher rate of CM in treated cows may have been due to relapses of earlier cases because of greater amounts of IMI or increased susceptibility to CM compared with control cows. However, when cows that became mastitic during the pretreatment period were excluded from the analysis, a significantly higher rate of CM was still observed in cows that were treated with bST. This review concluded that bST does not alter the normal relationship (positive correlation) that exists between milk yield 14 and mastitis. While it has been theorized that bST may increase CM incidence, it might be possible for bST to reduce the effects of clinical mastitis on the cow. Workers have observed that cows treated with bST had less lost milk production during a clinical mastitis event and returned to milk production faster following disease resolution compared to untreated controls (V andeputte-Van Messon and Burvenich, 1993). The mechanism responsible for this effect is likely multi-factorial with possible alterations in metabolism accounting for this effect including increased feed intake, enhanced immune cell fimction and increased irnmunoglobulin levels. It has been hypothesized that the decrease in immune function observed with increasing age of mammals may be due to declining somatotropin levels. Support for this observation is that the addition of exogenous growth hormone enhances immune function of older animals to a greater extent than it did in younger animals (Arkins et al., 1993). The primary effect of bST on immune cells is to stimulate both an increase in T-lymphocyte function and numbers. These effects appear to be due to IGF-l , whose production is increased upon administration of bST. Supporting evidence for this finding is that endotoxic challenge was observed to decrease IGF-l levels. Therefore, by stimulating IGF-l production, bST may mitigate some of the negative effects of endotoxemia on immune cell function (Elsasser et al., 1996). In this regard, the Monsanto Company is considering conducting research to determine if bST would be useful as an adjunct therapy for clinical mastitis. 15 Reproductive effects Reproductive performance and milk production levels are two of the most important determinants of dairy farm profitability (Ferguson, 1996; Lean et al., 1989). Ironically, these parameters have been found to be negatively associated with decreasing reproductive function observed as milk production levels increased (Cole et al., 1991; Cole et al., 1992; Harrison et al., 1990). This relationship is in part due to the positive association of increased milk production with the incidence .of diseases which can negatively affect reproductive performance (Deluyker et al. , 1991; Oresnik, 1995). Therefore, technologies that increase milk production must be evaluated for any negative impact they. might have on disease occurrence and reproductive performance. Despite producing increased milk production, if such technologies are associated with reduced reproductive performance, farm profitability could be adversely affected (Lean et al., 1989; Schmidt, 1989). Several disease conditions affecting dairy cows have been found to adversely - affect reproductive function. Diseases which may reduce reproductive performance include clinical mastitis, ketosis, retained placenta, metritis, abortion, cystic ovarian disease and lameness (Cullor, 1991; Etherington et al., 1996; Garverick, 1997; Oresnik, 1995). Because of the diversity of physiological mechanisms involved in the potential interaction between bST, diseases and reproduction, it is important to monitor the combined effects of bST usage on disease incidence and reproductive performance simultaneously. l6 A possible mechanism by which bST affects reproductive performance is by extending the time during which the cow is in negative energy balance. A negative correlation between energy balance and reproductive performance has been proposed (Butler and Smith, 1989; Cole et al., 1991; Harrison et al., 1990). When used according to labeled directions, the commercially-available bST product causes a period of negative energy balance for approximately seven weeks following the initiation of treatment (Bauman et al., 1989). However, it is unclear whether energy balance has any role in mediating the effect of bST on reproduction because reduced reproductive performance has been reported in treated cows that were in positive energy balance (Cole et al., 1991; Eppard etal., 1991). Additionally, energy balance could not have been a factor when reduced estrus expression was. observed in ovariectomized, non- lactating heifers that were treated with bST compared with untreated herdmates (Lefebvre and Block, 1992). Therefore, it appears that bST may affect reproductive performance by acting either directly (or through IGF-l) on the ovary or may influence other aspects of reproductive function indirectly that are not mediated by energy availability (De La Sota et al., 1993; Wells et al., 1995). It has been suggested that the possible adverse affects of this product on reproductive performance may be minimized by delaying bST treatment until after maximum negative energy balance has already occurred during lactation (McClary et al. , 1990). It has been observed that bST acts directly on the corpus luteum and the size of this ovarian structure has been found to increase with bST treatment. This likely accounts for the higher levels of progesterone observed in bST-treated cows (Lucy et 17 al., 1994; Schemm et al., 1990; Gallo and Block, 1991). It has been hypothesized that these elevated progesterone levels may suppress the expression of estrus because reduced estrus behavior was observed in bST—treated, ovariectomized non-lactating heifers (Lefebvre and Block, 1992). However, marginally reduced signs of estrus were also observed even when significantly lower progesterone levels occurred in bST- treated cows (Waterman et al., 1993). Therefore, bST treatment may be associated with reduced estrus expression in cows. While bST acts on the corpus luteum, it primarily affects follicular structures indirectly by increasing IGF-l concentrations (Lucy et al. , 1994). Cows treated with bST have been found to have more Class 1 (3 to 5 mm diameter) and Class 2 (6 to 9 mm diameter) follicles present compared to untreated controls. This may result from a physiological response of the ovary to increased IGF-l concentrations associated with bST treatment or increased responsiveness of the follicles to FSH, LH or both hormones (Lucy et al., 1994, Gong et al., 1991). Also, increased LH pulsatility has been observed ianT-treated cows (Schemm et al. , 1990). This may explain the increased growth rate and size of preovulatory follicles in treated cows (Lucy et al., 1994; De La Sota et al., 1993). Regardless of the underlying mechanism, bST has been associated with increased numbers of the largest follicles present on the ovary. Additionally, bST treatment resulted in the largest (preovulatory) follicles appearing to display less dominance over smaller (subordinate) follicles than in control cows (Lucy et al. , 1994). Increased incidence of twin calvings have been observed with bST treatment in 18 some studies (Cole et al., 1991; Estaban et al., 1994). However, the mechanism responsible for bST-associated twinning is unclear. Even though increased numbers of small follicles have been found to be present in bST-treated cows, an increase in the ovulation rate has not been observed in these cows (De La Sota et al., 1993; Lucy et al., 1994). If multiple ovulations could be demonstrated, this would help explain the increased twinning that may be observed with bST use. It is possible that this indeed does occur but the timing of observations to date have simply not taken place at the correct times (Matt Lucy, personal communication, 1997). Although most studies have found no significant association between fetal loss and bST use, this was observed in at least, one trial (Cole et al. , 1992). It is unknown as to what mechanism may be responsible for this association or if random chance alone was solely responsible. Some preapproval studies observed that bST significantly increased days open of cows (Cole et al., 1992; Burton et al., 1990). However, other workers only observed a negligible increase in days open for bST-treated cows (Leonard et al., 1990; Hansen et al., 1993; Rijpkema et al., 1990; Zhao et al., 1992). Because bST alters the shape of the lactation curve, this distorts the normal relationship between reproductive performance and profitability because bST-treated cows produce milk at higher levels over prolonged periods compared to untreated cows. Based on this reduced cost of additional days open when bST is used, some Workers have proposed that increased profitability will be realized when extended calving intervals (e. g., 16.5 months) are used for bST-treated cows (Galton et al., 1997). 19 Few studies have evaluated the effect of the commercially prepared bST product (Posilac‘, Monsanto and Company, St. Louis, MO).on reproductive performance when used for a complete lactation. It is necessary to observe reproductive performance for an entire lactation because the effect of bST on reproduction likely varies with the dose used and periods of time for which it is used (Esteban et al. , 1994). Because many preapproval studies were of short duration, a study for a complete lactation using the commercially available bST product is necessary. Longevity The assessment of effects on cow longevity are an important step in evaluating new technologies. Any reduction in herd life likely has a severe negative impact on any financial gains incurred through adoption of such technologies (Beaudeau et al., 1995). It is possible for bST to exert both positive and negative influence on cow survival. By increasing milk production, especially in late lactation, bST could delay culling decisions for treated cows. However, if any negative cow events (disease) occurred that were associated with bST use, this effect may shorten the functional longevity of cows. The use of Cox’s proportional hazards model (Cox and Oakes, 1984) gives the best estimate of the relationship between diseases and culling in terms of risk assessment as well as any effect on herd life (Beaudeau et al., 1995). This model allows analysis of the effects that various cow-specific covariates (e. g., parity and herd) and disease variables may have on culling. Through the use of such a model, the impact that bST may have on culling can be evaluated in terms of risk and impact on herd life. 20 Many studies have attempted to determine the reason for culling in dairy cows. The three most common reasons given for culling cows are reproductive failure, low milk production and mastitis (Beaudeau et al., 1995). It is possible for bST to affect all of these reasons for culling cows. Of these reasons for culling, failure to be come pregnant (reproductive failure) is consistently the most frequent reason given for culling cows (Beaudeau et al., 1994; Esslemont and Kossaibati, 1997). If bST use is associated with reproductive failure, either directly itself, or indirectly by increasing the incidence of disease which can adversely affect reproductive performance, this would result in bST having a negative impact on cow longevity. However, in regard to culling for reason of low production, bST likely has a positive effect and may reduce culling for this reason by increasing milk production. DHIA lactation projections The influence that bST may have on DHIA milk production projections has been completely ignored. DHIA currently feels that no significant bias is introduced into lactational milk projections by the use of bST. However, the shape of the lactation curve is significantly altered by the use of bST‘because lactational persistency is increased (Bauman et al., 1989). In addition, bST causes a second peak in production that occurs after a cow’s “natural” peak. Because DHIA projections are often made based on a single days production each month, depending on which day within the l4-day PosilacR injection cycle that this projection is made from also may significantly impact this estimation. Based on these potential sources of variation, it has been suggested that new 21 lactation curves and equations need to be constructed for the accurate prediction of milk production for bST-treated cows (Leitch et al., 1990). Farms currently use DHIA records to support culling decisions. The DHIA Records Manual states that cows with projected 305-day Mature Equivalent milk production of 2,000 lbs. less than herdmates are good candidates for culling (DHIA Records Manual, 1989). As a result of increasing milk production, bST influences cows of lower genetic merit for milk production to more closely resemble cows with high potential for milk production (Bauman et al., 1989) and therefore may decreases the risk of such cows being culled. Economic impact A dairy farm is first and foremost a business and the fundamental principle of any business is to make a profit for its investors. Bovine somatotropin is a management change which is designed to increase profitability. However, most estimates of increased profitability resulting from bST use to date have largely ignored any negative financial effects which may be present or have used estimates of increased costs which may occur (Elbehri and Yonkers, 1995). Most estimates only take into account the cost of bST, administration of the product and additional feed which will be consumed by treated cows. Also, estimates have been used for anticipated profits from increased milk production. However, these estimates are based on data generated from pre-approval trials (Shoeffling et al., 1991), were performed in research-type dairy herds and often did not use the commercially-available bST (PosilacR) product according to label directions. 22 Therefore, these estimates may not be accurate in commercial dairy situations. It has been stated that actual increases in production in commercial dairy herds may be as much as 25% less than were observed during preapproval studies (Fallert et al., 1987). Increased milk production in response to bST has been found to be variable between cows with some cows increasing in production up to 10 kgs / day and others showing no response at all (Hard et al., 1992; McGuire et al., 1992). Therefore, to accurately calculate increased profits from bST use daily, individual-cow milk weights must be used to compare similar groups of bST-treated and untreated herdmates. Any increased cost attributable to bST use must be subtracted from increased profits. Of primary concern is clinical mastitis, which has been observed to increase concurrently with increased milk production (Oltenacu and Ekesbo, 1994) and increased incidence of diseases which may decrease reproductive performance (Deluyker et al., 1991; Oresnik, 1995). Such financial losses would need to be accounted for to arrive at an accurate estimate of increased farm profitability resulting from bST use. Workers have observed decreased milk production in subsequent lactations afier bST use but the magnitude of this effect is not currently known and few studies have addressed this issue (Gibson et al., 1992). If this is a significant effect of bST use, it would need to be counted as a cost of using this product as well. Partial budgeting has been used to examine the financial impact of management changes (Hady and Lloyd et al., 1994). In this analysis, only revenue and expense changes which occur as a result of the management change are considered. In regard to bST use, potential variables to consider are changes in milk production, disease rates, 23 feed costs, labor costs and replacement animals costs (affected by culling rates). When constructing such analyses, accuracy is increased by utilizing farm-specific data (Lloyd et al., 1987). However, most work to date has made use of industry-wide estimates for many of these variables when examining the financial impact of bST use (Shoefflling et al., 1991; Elbehri and Yonkers, 1995) After an estimate is made of a change in farm revenue as a result of a particular management change, it is important to perform a sensitivity analysis for variables which are subject to change and thereby affect the influence of the proposed change (Hady and Lloyd et al., 1994). This is, important because of the inherent volatility in prices which are either paid or received by the farm. Through this analysis, factors which have the largest impact on profitability of the management change can be determined. This allows managers to make better financial choices when deciding whether or not to adopt or continue the use of a particular management program. 24 REFERENCES Arkins, S., R. Dantzer, and K. W. Kelley. 1993. Somatolactogens, somatomedins, and Oimmumity. J. Dairy Sci. 76:2437-2450. Bartlett, P. C., J. Van Wijk. 1991. Temporal patterns of lost milk production following clinical mastitis in a large Michigan Holstein herd. J. Dairy Sci. 74:1561-1572. Bauman, D. E., D. L. Hard, B. A. Crooker, M. S. Partridge, K. Garrick, L. D. Sandles, H. N. Erb, S. E. Franson, G. F. Hartnell, and R. L. Hintz. 1989. Long-term evaluation of a prolonged-release formulation of N-methionyl bovine somatotropin in lactating dairy cows. J. Dairy Sci. 72:642-651. Beaudeau, F ., V. Ducrocq, C. Fourichon, and H. Seegers.1995. Effect of disease on length of productive life of French Holstein dairy cows assessed by survival analysis. J. Dairy Sci. 78:103-117. Beerepoot, G. M. M., A. A. Dykhuizen, M. Nielen, and Y. H. Schukken. 1991. The economics of naturally occurring twinning in dairy cattle. J. Dairy Sci. 75:1044- 1051. Burton, J. L., B. W. McBride, J. H. Burton, and R. G. Effert. 1990. Health and reproductive performance of dairy cows treated for up to two consecutive lactations with bovine somatotropin. J Dairy Sci. 73:3258-3265. Butler, W. R., and R. D. Smith. 1989. Interrelationships between energy balance and postpartum reproductive function in dairy cattle. J Dairy Sci. 72:767-783. Cole, J. W., P. J. Eppard, B. G. Boysen, K. S. Madsen, R. H. Sorbet, M. A. Miller, R. L. Hintz, W. E. Ribelin, B. G. Hammond, R. J. Collier, and G. M. Lanza. 1992. Response of dairy cows to high doses of a sustained-release bovine somatotropin administered during two lactations. 2. Health and reproduction. J. Dairy Sci. 75:11-123. Cole, W. J., K. S. Madsen, R. L. Hintz, and R. J. Collier. 1991. Effect of recombinantly-derived bovine somatotropin on reproductive performance of dairy cattle. Theriogenology. 36:573-595. Cullor, J. S. 1991 . Mastitis in dairy cows: does it hinder reproductive performance? Vet Med. 86: 830-832. 25 Cox, D. R., and D. Oakes. 1984. Analysis of survival data. Chapman and Hall, New York De La Sota, R. L., M. C. Lucy,C. R. Staples, and W. W. Thatcher. 1993. Effects of recombinant bovine somatotropin (sometribove) on ovarian function in lactating and nonlactating dairy cows. J Dairy Sci. 76:1002-1013. Deluyker, H. A., J. M. Gay, L. D. Weaver, A. S. Axari. 1991. Change in milk yield with clinical diseases for a high producing dairy herd. J Dairy Sci. 74:463-445. Downer, J. V., D. L. Patterson, D. W. Rock, W. V. Chalupa, R. M. Cleale, J. L. Firkins, G. L. Lynch, J. H. Clark, B. O. Brodie, B. F. Jenny, and R. De Gregorio. 1993. Dose titration of sustained-release recombinant bovine somatotropin in lactating dairy cows. J. Dairy Sci. 76:1125-1136. Elsasser, T. H., M. Richards, R. Collier, and G. F. Hartnell. 1996. Physiological responses to repeated endotoxin challenge are selectively affected by recombinant bovine somatotropin administration to calves. Dom. An. Endo. 13(1):91-103. Elbehri, A., and R. D. Yonkers. 1995. Economic analysis of the impacts of bovine somatotropin on the profitability of representative dairy farms in the northeast. Ag. and Res. Econ. Rev. 4:88-100. Eppard, R. 1., 5. Hudson, w. J. Cole, R. L. Hintz, G. F. Hartnell, T. w. Hunter, L. E. Metzger, A. R. Torkelson, B. G. Hammond, R. J. Collier, and G. M. Lanza. 1991. Response of dairy cows to high doses of a sustained-release bovine somatotropin administered during two lactations. 1. Production response. J. Dairy Sci. 74:3807-3821. Esslemont, R. J., and M. A. Kossaibati. 1997. Culling in 50 dairy herds in England. Vet Rec. 140:36-39. Esteban, E., P. H. Kass, L. D. Weaver, J. D. Rowe, C. A. Holrnberg, C. E. Franti, and H. F. Trout. 1994. Reproductive performance in high producing dairy cows treated with recombinant bovine somatotropin. J Dairy Sci. 77:3371-3381. Etherington, W. G., M. L. Kinsel, and W. E. Marsh. 1996. Relationship of production to reproductive performance in Ontario dairy cows: herd level and individual animal descriptive statistics. Theriogenology. 46:935-959. Ferguson, J. D. 1996. Diet, production, and reproduction in dairy cows. An Feed Sci Tech. 59:173-184. 26 Firat, M. Z. 1993. An investigation into the effects of clinical mastitis on milk yield in dairy cows. Livest. Prod. Sci. 36:311-321. Gallo, G. F., and E. Block. 1991. Effects of recombinant-bovine somatotropin on hypophyseal and ovarian functions of lactating dairy cows. Can. J. Anirn. Sci. 71:343-353. Galton, D. M., R. W. Everett, M. E. Van Amburgh, D. L. Bauman, and W. A. Knoblauch. 1997. Extended calving intervals with the use of bST. Presented at the Tri-State Dairy Nutrition Conference, Fort Wayne, IN. Garverick, H. A. 1997. Ovarian follicular cysts in dairy cattle. J Dairy Sci. 80:995- 1004. Gibson, J. P., B. W. McBride, J. H. Burton, 1. Politis, and X. Zhao. 1992. Effect on production traits of bovine somatotropin for up to three consecutive lactations. J. Dairy Sci. 75:837-846. Gong, J. G., T. Bramley, and R. Webb. 1991. The effect of recombinant bovine somatotropin on ovarian function in heifers: follicular populations and peripheral hormones. Biol Reprod. 45:941-949. Hady, P. J. and J. W. Lloyd, J. B. Kaneene, and A. L. Skidmore. 1994. Partial budget model for reproductive programs of dairy farm businesses. J. Dairy Sci. 77:482-491. Hansen, W. P., D. E. Otterby, and J. G. Linn. 1993. Multi-farm use of bovine somatotropin for two consecutive lactations and its effects on lactational performance, health, and reproduction. J Dairy Sci. 77:94-110. Hard, D. L., F. Adriaens, C. G. Prosser, J. A. Newbold, R.'B. Heap, R. H. Phipps, and M. J. Hannah. 1992. Effects of recombinant bovine somatotropin (bST) and nutrition on galactopoiesis, hepatic bST receptors and plasma hormone concentrations in the dairy cow. J. Physiol. 446:171P. (Abstr.). Harrison, R. O., S. P. Ford, J. W. Young, A. J. Conley, and A. B. Freeman. 1990. Increased milk production versus reproductive and energy status of high producing dairy cows. J Dairy Sci. 73:2749-2758. 27 Hoblet, K‘. H., G. D. Schnitkey, D. Arbaugh, J. S. Hogan, K. L. Smith, P. S. Schoenberger, D. A. Todhunter, W. D. Hueston, D. E. Pritchard, G. L. Bowman, L. E. Heider, B. L. Brockett, and H. R. Conrad. 1991. Costs associated with selected preventive practices and with episodes of clinical mastitis in nine herds with low somatic cell counts. JAVMA. 199(2): 190-196. Kronfeld, D. S. 1994. Health management of dairy herds treated with bovine somatotropin. JAVMA 204(1):116-130. Lean, I. J ., J. C. Galland, and J. L. Scott. 1989. Relationships between fertility, peak milk yields and lactational persisitency in dairy cows. Theriogenology. 31:1093- 1103. ‘ Lefebvre, D. M. , and E. Block. 1992. Effect of recombinant bovine somatotropin on estradiol-induced estrus behavior in ovariectomized heifers. J Dairy Sci. 75: 1461-1464. Leitch, H. W., E. B. Burnside, and B. W. McBride. 1990. Treatment of dairy cows with recombinant bovine somatotropin: genetic and phenotypic aspects. J. Dairy Sci. 73:181-190. Leonard, M., M. Gallo, G. Gallo, and E. Block. 1990. Effects of a 28-day sustained- release formulation of recombinant bovine somatotropin (rbST) administration to cows over two consecutive lactations. Can J Anim Sci. 70:795-809. Lloyd, J. W., J. B. Kaneene, and S. B. Harsh. 1987. Toward responsible farm-level economic analysis. JAVMA. 191(2):]95-199. Lucy, M. C., T. L. Curran, R. J. Collier, and W. J. Cole. 1994. Extended function of the corpus luteum and earlier development of the second follicular wave in heifers treated with bovine somatotropin. Theriogenology. 41:561-572. McClary, D. S., R. K. McGuffey, and-H. B. Green. 1990. Bovine somatotropin: Part 2. Agri-pract. 11:5-11. McGuire, M. A., D. E. Bauman, M. A. Miller, and G. F. Hartnell. 1992. Response of somatomedins (IGF-I and IGF—II) in lactating cows to variations in dietary energy and protein and treatment with recombinant n-methionyl bovine somatotropin. J. Nutr. 122:128. Michigan Dairy Herd Improvement Association. 1989. Records Manual. p. 2-18. 28 Oltenacu, P. A., and I. Ekesbo. 1994. Epidemiological study of clinical mastitis in dairy cattle. Vet. Res. 25:208-212. Oresnik, A. 1995. Effect of health and reproductive disorders on milk yield and fertility in dairy cows. Bovine Pract. 29:43-45. Pell, A.'N., D. S. Tsang, B. A. Howlett, M. T. Huyler, V. K. Meserole, W. A. Samuels, G. F. Hartnell, and R. L. Hintz. 1992. Effects of a prolonged-release formulation of sometribove (n-methionyl bovine somatotropin) on Jersey cows. J. Dairy Sci. 75:3416-3431. PosilacR product label. 1993. Monsanto Co., St. Louis, MO. Remond, B., M. Cisse, A. Ollier, and Y. Chilliard. 1991. Slow release somatotropin in dairy heifers and cows fed two levels of energy concentrate. 1. Performance and ' body condition. J. Dairy Sci. 74:1370-1381. Rijpkema, Y. S., L. van Reeuwijk, and D. L. Hard. 1990. Responses of dairy cows to treatment with sometribove (r-BST) during three consecutive years. Livest. Prod. Sci.-26:193-216. ' Schemm, S. R., D. R. Deaver, D. R. Griel, and L. D. Muller. 1990. Effects of recombinant bovine somatotropin on luteinizing hormone and ovarian function in lactating dairy cows. Biol Reprod. 42:815-821. Schmidt, G. H. 1989. Economics of using bovine somatotropin in dairy cows and potential impact on the US dairy industry. J. Dairy Sci. 72:737-745. Shoeffling, J. R., R. C. Angus, D. V. Armstrong, and J. T. Huber. 1991. Economic implications of bovine somatotropin use for the Arizona dairy industry. J. Dairy Sci. 74:2347-2352. Skarda, J ., E. Markalous, J. Slaba, P. Krejci, O. Skardova, and J. Zednik. 1992. Effect of methionyl bovine somatotropin in a prolonged-release vehicle on milk production, hormone profiles and health in dairy cattle. J. Dairy Res. 59:499- 506. Thomas, J. W., R. A. Erdman, D. M. Galton, R. C. Lamb, M. J. Arambel, J. D. Olson, K. S. Madsen, W. A. Samuels, C. J. Peel, and G. A. Green. 1991. Responses by lactating cows in commercial dairy herds to recombinant bovine somatotropin. J. Dairy Sci. 74:945-964. 29 Vandeputte-Van Messon, G., and C. Burvenich. 1993. Effect of somatotropin on changes in milk production and composition during coliform mastitis in periparturient cows. J. Dairy Sci. 76:3727-3741. Waterman, D. F., W. J. Silvia, R. W. Hemken, G. Heersche Jr., T. S. Swenson, and R. G. Eggert. 1993. Effect of bovine somatotropin on reproductive function in lactating dairy cattle. Theriogenology. 40: 1015-1028. Wells, S. J ., A. M. Trent, R. S. Collier, and W. J. Cole. 1995. Effect of long-term administration of a prolonged release formulation of bovine somatotropin (sometribove) on clinical lameness in dairy cows. Am. J Vet. Res. 56(8):992- 996. Whitaker, D. A., E. J. Smith, J. M. Kelly, and L. S. Hodgson-Jones. 1988. Health, . welfare and fertility implications of the use of bovine somatotropin in dairy cattle. Vet. Rec. 122:503-505. White, T. C., et. al. 1994. Clinical mastitis in cows treated with sometribove (recombinant bovine somatotropin) and its relationship to milk yield. J. Dairy Sci. 77:2249-2260. Zhao, X., J. H. Burton, and B. W. McBride. 1992. Lactation, health, and reproduction of dairy cows receiving daily injectable or sustained-released somatotropin. J Dairy Sci. 75:3122-3130. 30 CHAPTER THREE RECOMBINANT BOVINE SOMATOTROPIN AND CLINICAL MASTITIS: INCIDENCE AND DAYS OF MILK DISCARDED DUE TO THERAPY ABSTRACT Holstein cows (n = 555) from four Michigan dairy farms were randomly assigned to receive bovine somatotropin (bST) or to serve as untreated controls. Bovine somatotropin (500 mg) was administered subcutaneously every 14 days beginning at 63 to 69 days of lactation and continuing until approximately 21 days prior to end of lactation or until the animal was removed from the herd. Study objectives were to determine the effect of bST on the incidence of clinical mastitis, number of days that milk was withheld because of therapy for clinical mastitis, and culling for mastitis. A total of 127 (22.9%) cases of clinical mastitis occurred during lactation. In the pretrial period (before 63 to 69 days of lactation), 42 (33.1%) cases occurred, and 85 (66.9%) cases occurred during the trial period. Of the 42 pretrial cases, 57.1 % were in the control cows, and 42.9% occurred in the treated cows. Of the 85 trial cases, 47.1% were in the control cows, and 52.9% occurred in the treated cows. Using logistic regression, the odds ratio for the incidence of clinical mastitis in bST-treated cows was 1.06 (95% CI: 0.62 to 1.81). The number of days that milk was discarded following therapy for clinical mastitis and the culling rate for mastitis did not differ between the study treatment groups. 31 INTRODUCTION Mastitis is an economically important disease to the dairy industry with costs associated with mastitis estimated at approximately $100.00 per clinical case (Hoblet et al., 1991, Miller et al., 1993). The risk of CM has been shown to increase as milk production increases (Oltenacu and Ekesbo, 1994) with milk production decreasing following disease onset (Bartlett and Van Wijk, 1991; Firat, 1993; Oltenacu and Ekesbo, 1994). Therefore, technologies that enhance milk production may be positively correlated with CM and thereby increase financial losses on the farm caused by this disease. The approval of bST (PosilacR; Monsanto Company, St. Louis, M0) for lactating dairy cows has caused considerable scientific and public controversy over potential effects this product might have on the incidence of CM and subsequent antibiotic use resulting from therapy (Kronfield, 1994). Of nine trials in which 500- to 700-mg doses of bST were administered at 14-day intervals, seven studies concluded that the incidence of CM did not differ between cows that were administered bST and unsupplemented control cows (Downer et al. , 1993; Eppard et al. , 1991; Remond et al., 1991; Rijkema et al., 1990; Skarda et al., 1992; Thomas et al., 1991; Whitaker et al., 1988). In two trials, a significant increase in CM was reported for treated cows compared with control cows. However, in these trials, cows receiving bST had a significantly greater incidence of CM than did control cows before bST treatment was initiated, which might have affected the outcomes of the studies (Cole et al. , 1992; Pell 32 et al., 1992). In a review of 15 full lactation studies conducted in the US and Europe, 467 lactation records of cows supplemented with bST and 447 lactation records of control cows were summarized (White et al., 1994). A significantly higher rate of CM was observed for the cows that had been treated with bST. However, a significantly higher rate of CM was also apparent during the pretreatment period for cows that were later treated with bST. Thus, the higher rate of CM in treated cows may have been due to relapses of earlier cases because of greater amounts of IMI or increased susceptibility to CM compared with control cows. However, when cows that became mastitic during the pretreatment period were excluded from the analysis, a significantly higher rate of CM was still observed in cows that were treated with bST. This review concluded that bST does not alter the normal relationship (positive correlation) that exists between milk yield and mastitis. The purpose of the present study was to determine the effect of bST on the incidence of CM and the number of days that milk was discarded because of therapy for CM. SPECIFIC METHODS Quarter milk samples in duplicate were collected from the study cows between 30 and 60 days of lactation. Cultures were determined to be positive only if the same organism was observed in both sample replicates. Samples were determined to be contaminated when three or more organisms were isolated. These pretrial milk cultures were not used to assign cows to study groups but were used to compare the prevalence 33 of pretrial intra-mammary infection between the treated and control groups. To compare the incidence of CM before and after trial entry, all cases 0f CM that occurred during the entire lactation were recorded. A case of clinical mastitis was defined as a quarter with abnormal milk observed throughout a particular milking. A new case of CM in a previously affected quarter occurred when a period of three weeks had transpired between the observation of CM. Farm personnel were asked to aseptically collect pretreatment milk samples from affected quarters at the first milking when CM was observed. Subsequently, samples were frozen and retrieved by the investigators during monthly farm visits. All milk samples were cultured for isolation of bacteria by plating 0.01 ml of milk on 5 % sheep blood and MacConkey agar plates. Further identification of isolates-was accomplished .by Gram staining, catalase, coagulase, and CAMP testing (National Mastitis Council, 1987). The SAS System for WindowsR (SAS Institute, 1994) was used for logistic regression analysis of CM incidence, StatistixR (Analytical Software, 1994) was utilized for Student's t-testing of days of milk withheld because of antibiotic therapy, and Epi Info (Centers for Disease Control, 1996) was used to calculate chi-square values for cows culled because of mastitis and the Z-test statistic for the binomial distribution test of CM incidence. Cases of CM for each farm and study group were analyzed by the use of a binomial distribution test; the equation was Z = a - N,M,/T (MlNlNo/Tz)” 34 where Z = calculated Z-test statistic; a = number of cases observed in the treated group; N1 = amount of cow-time experienced by the treated group; N 0 = amount of cow-time experienced by the control group; M1 = total number of cases observed, and T = total amount of cow-time experienced by both study groups. The use of this equation was necessary in order to adjust for multiple cases of CM occurring in the same cow and the different time frame in each treatment group during which these cases occurred. Cows that were culled because of mastitis were compared by Mantel-Haenszel chi-square. The number of days that milk was withheld because of therapy for CM and pretreatment milk production were analyzed by Student's t-test. A logistic regression model was constructed for cases of CM that occurred during the trial period. The dependent variable was the occurrence of CM and the independent variables of farm, lactation, and season of calving were investigated as potential independent variables. The variable for treatment group was forced into the model to assess the effect of bST. The model analyzed was ”(3'me where Y = probability of clinical mastitis occurrence, e = base of natural log, and f (MAST) = intercept + PM + LACT + SEAS + GROUP + error where PM = farm of origin (1, 2, 3, or 4); LACT = lactation (parity) (1 through 8); SEAS = season of calving (1 = summer; 2 = fall; 3 = winter), and GROUP = treatment [0 = control 35 group; 1 = treated (bST) group]. The variable for herd was investigated as both a fixed and random effect. All possible two-way interactions of variables that were significant in the model were assessed as well as the quadratic effect of the continuous variable for lactation. The final model was constructed by initial use of a backward variable selection method and later by a forward model-building strategy. RESULTS Mastitis For the 555 cows in the analysis, 127 cases of CM (22.9% overall incidence) occurred during the test lactation with 64 (50.4%) of the cases in the control group and 63 (49.6%) in the bST-treated group (P = 0.76). Of these cases, 42 (33.1%) occurred before the trial period (< 63 days of lactation); 24 (57 . l %) involved the control group and 18 (42.9%) involved the bST-treated group (P = 0.29). During the trial, 85 (66.9%) cases occurred; 40 (47.1%) involved the control group and 45 (52.9%) involved the bST-treated group (P = 0.37). Rates for CM (cases per 100 cow-months) (Table 3.1) were determined for each farm and study period. 36 TABLE 3.1: Incidence of clinical mastitis (cases per 100 cow-months) in dairy cows administered 500 mg of bST (PosilacR; Monsanto Company, St. Louis, M0) at 14-day intervals and in untreated controls. l - 63 days of lactation > 63 days of lactation Total lactation Farm no. *bST Control Total bST Control Total bST Control Total 1 68 4.40 3.33 3.97 3.66b 0.85' 2.48 3.6 1.31 2.68 2 22 2.43 4.00 3.22 2.15 2.03 2.09 2.1 2.76 2.44 3 13 4.24 2.08 3.24 1.12 1.67 1.39 1.7 1.73 1.72 4 12 0.73b 5.59‘ 3.36 0.88 0.53 0.68 0.7 1.58 1.22 M1 55 2.81 3.90 3.35 1.85 1.52 1.68 1.9 2.08 2.02 "" Rates in the same row with different superscript letters differ (P < 0.05). * Cows that would later be treated with bST. Of the 85 cases of CM occurring in the trial period, 19 (22.4%) of these cases were repeat cases occurring in the same cows. Of these 19 cases, eight cases (42.1%) were detected in control cows and 11 cases (57.9%) occurred in bST-treated cows (P = 0.63). These 19 cases of CM were detected in 14 cows of which seven were control cows and seven had been assigned to bST treatment. Nine cows were observed to have a case of CM in both the pretrial and trial periods with four of these being control cows and five being bST-treated cows (P =0.74). Pretrial milk cultures were obtained from approximately 82 % of the study cows. Of these cultures, 51.8% were from control cows and 48.2% were from cows in the treatment group. The prevalence of IMI between 30 and 60 days of lactation was 35.8% for all trial cows with 35.6% prevalence observed in the control cows and 35.9% found in the treaunent cows (P = 0.94). Coagulase-negative staphylococci and no organism isolated were the predominant culture results (Table 3.2). 37 TABLE 3.2: Milk culture results (duplicate quarter samples) at 30 to 60 days of lactation (pretrial period) in dairy cows later administered 500 mg of bST (PosilacR; Monsanto Company, St. Louis, M0) at 14-day intervals and in untreated controls. Culture result bST Control Total Strep. ag. 1 0 1 Strep. sp. 3 2 5 Staph. aureus 6 9 15 Staph. sp. 54 58 112 Gram-negative 4 4 8 Mixed 6 4 10 Others 5 9 14 No growth 141 150 291 Not cultured 65 34 99 Total 284 1271 555 Milk cultures from CM cases were performed on 45.8% of pretrial cases and 49.4% of the cases occurring during the trial period. Overall, milk cultures were obtained from 47.3% of all CM cases during the study. The number of cultures that were obtained did not differ between the study groups. Gram-negative bacteria, environmental streptococci, and no organism isolated were the predominant culture results (Table 3.3). The type of organism isolated from cases of CM did not differ between treatment groups. 38 TABLE 3.3: Milk culture results (single quarter samples) from clinical mastitis cases in dairy cows administered 500 mg of bST (PosilacR; Monsanto Company, St. Louis, M0) at 14-day intervals and in untreated controls between approximately 63 days of lactation and dry off. Culture result bST Control Total Strep. ag. 0 0 0 Strep. sp. 7 2 9 Staph. aureus 2 3 5 Staph. sp. 3 0 3 Gram-negative 3 7 10 Mixed 0 1 1 Others 0 0 0 N 0 growth 8 8 16 Not cultured 22 19 41 Totals 45 40 85 Using daily individual cow milk weights, milk production in the pretreatment period was compared between treated and control cows; no significant difference was found in milk production between treatment groups (P = 0.70). A logistic regression model was used to adjust for possible confounding variables and odds ratios were calculated for variables in the fitted model (Table 3.4). An odds ratio measures the likelihood of disease occurrence for a given risk exposure when compared to those individuals without the exposure. The odds ratio for the bST treatment effect was found to be 1.06 (95 % CI: 0.62 to 1.81) when the independent variable for farm was considered as a fixed effect. This was interpreted as cows treated with bST were 6% more likely to have an Occurrence of CM than those who were not treated. However, since the 95 % CI contained one (the point of no difference in odds 39 of disease occurrence between treatment groups), the true point estimate of the odds ratio is likely not significantly different than one. When the independent variable for farm was investigated as a random effect, the model output for the effect of bST changed insignificantly. Therefore, the independent variable for farm effect was considered to be a fixed effect in the final model. Model fit was accessed by the Hosmer-Lemeshow test; the goodness-of-fit statistic calculated was 7.91 with 9 df. The corresponding probability (P = 0.54) indicated that the model fit the data quite well. TABLE 3.4: Logistic regression analysis results of clinical mastitis in dairy cows administered 500 mg of bST (PosilacR; Monsanto Company, St. Louis, M0) at 14-day intervals and in unsupplemented controls. Source of variation df Parameter Significance Odds ratio (95 % CI) estimate Farm 1 1 1.6575 0.0037 5.25 (1.71 tol6.06) Farm 2 1 1.7559 0.0004 5.79 (2.19 t015.29) Farm 3 1 1.0251 0.0685 2.79 (0.93 to 8.40) Lactation 1 1.1290 0.0063 3.09 (1.38 to 6.95) Lactation“ 1 -0.1179 0.0599 0.89 (0.79 to 1.01) Study group 1 0.0585 0.8302 1.06 (0.62 to 1.81) ' The quadratic effect of the variable for lactation. Discarded milk The mean (j-_SEM) number of days that milk was discarded following all cases of CM during the trial period was 7.5 i 1.7 days for control cows and 4.2 _1- 1.1 days for bST-treated cows (P = 0.11). For CM cases during the trial period where 40 antibiotics were administered (52.9% of the cases), the mean number of days that milk was discarded was 11.5 i 2.3 days for control cows and 10.0 _-t-_ 1.9 days for bST- treated cows (P = 0.63). DISCUSSION Field trials are susceptible to an underreporting bias because of potential deficiencies in the recording of animal health events. Previous researchers have identified this source of error and the variation in compliance that exists among farms (Bartlett et al. , 1992). In the present study, all herds had disease-recording systems in place at the start of the trial that assisted in the accurate recording of disease events, subsequent treatments, and periods of discarded milk. Twenty-two cows that were originally selected for the trial were in poor health at the time when treatment was to begin and therefore, were excluded from the analysis. This exclusion was in accordance with the product label that stated that cows must be determined to be healthy if bST treatment is to begin during the 9th week of lactation (Monsanto Company, 1993). Because less than 5% of the cows selected for the trial were excluded, and that treated and control cows were excluded based on the same criteria, it is unlikely that a sizable bias was introduced from this omission. The overall incidence of CM in found in this study (2.02 cases / 100 cow- months) is lower than that observed in Ohio [3.06 cases / 100 cow-months; (Bartlett et al., 1992), California [3.10 cases / 100 cow-months; (Weigher et al., 1990)], and Pennsylvania [4.20 cases / 100 cow-months; (Erskine etal., 1988)]. The relatively low 41 rates of CM observed in this study might have resulted from several managerial and environmental factors- including the use of teat dip prior to and after milking, utilizing dry cow therapy on all cows, milking all cows three times per day, providing housing for lactating cows that emphasized clean and dry free stalls, and having low herd prevalences of contagious mastitis organisms. In this study, the rate of CM experienced during the first 63 days of lactation was approximately double the rate for the remainder of lactation (3.35 compared to 1.68 cases / 100 cow-months). This result is in agreement with previous studies that also found increased risk of CM in early lactation (Erskine et al. , 1988; Houben et a1. , 1993; Oltenacu and Ekesbo, 1994). Thus, potential effects of bST on CM are reduced when treatment is initiated after the first 63 days of lactation because the administration is during a period of lower risk. The incidence of CM incidence during the pretreatment period, although not statistically different between the study treatment groups, varied among farms and between treatment groups within farms. For example, the rate of CM in cows assigned to the control group for farm four was significantly higher (P -- 0.02) than cows assigned to the bST-treated group. Primarily because of the influence of this farm, the cows assigned to the study control group had a slightly higher, but not significantly different (P = 0.29), rate of CM in the pretreatment period. This result is in contrast to previous studies where a bias toward a higher incidence of CM in the treated group of cows was observed before bST treatment began (White et al. , 1994). However, farm one had higher rate of CM in the bST-treated cows during the treatment period (P = 0.04) which was not observed in the 42 pretreatment period on this farm. Nonetheless, despite herd variations, the treatment period and overall lactation incidence of CM was not different between the study treatment groups The number of cows with multiple cases of CM in the treatment period was found to be similar between groups. The treatment groups did not differ in either the number of cows experiencing repeat cases or in the number of repeat cases occurring in each group. Also, no difference between the study groups was observed in the number of cows that experienced CM during both the pretreatment and treatment periods. The type of organism that was isolated from cows with CM did not differ between groups which is in agreement with other trials that found no effect of bST on the type of pathogen isolated from cases of CM (Lissemore et al., 1991). Although samples were collected from only 47.3% of recorded cases of CM in this study, I believe that the results of cultures that were obtained are representative of all cases. Furthermore, since the number of cultures performed were similar between the treatment groups, no significant bias occurred. However, because a large number of cases were not cultured, small differences in the type of organism isolated might not have been detected. The predominant isolates (Streptococcus sp. , Staphylococcus sp. , and Gram-negative rods) were similar to those found in previous studies (Bartlett et al. , 1992; Bartlett and Van Wijk, 1991; Houben et al., 1993). In the logistic regression model, no. variable was included for milk production. Milk production has been viewed as an intervening variable in the relationship between bST and CM. 1 hypothesized that the mechanism responsible for this association is that 43 bST increases CM indirectly through enhanced milk production. Milk production was not included in this statistical model because it is not appropriate to use intervening variables as independent variables in multivariate models (Kleinbaum etal., 1982). The odds ratio of this model determined that bST-treated cows were 6% more likely to experience CM than were control cows. Such small differences in estimated relative risk of CM are likely of little biological significance. Furthermore, because the 95 % confidence interval for the odds ratio includes one, there is likely no significant difference in the estimated risk of CM between the treatment groups. However, since the variable for treatment effect was not statistically significant in the model, some loss of precision occurred in regard to this variable when it was included in the final model. The number of days of milk production that were discarded per case of CM was slightly less for the treated .cows (P = 0.11). However, the primary reason for this difference was the influence of farm one, which followed a non-antibiotic course of therapy for cases of CM. Consequently, few days of milk production were discarded. On this farm, 12 bST-treated cows experienced CM compared to only two control cows. Therefore, 27.3% of all cases of CM in the study bST-treated cows compared to only 5.0% of all cases of CM in the study control cows occurred on farm one. Thus, this one farm had a disproportional effect on the number of days that milk was discarded between the study groups. When only cases of CM for which antibiotic therapy was administered were analyzed, no difference was observed between the treatment groups (P = 0.63). CONCLUSIONS It has been hypothesized that bST effects CM by way of an indirect mechanism that acts through increased milk production. In this trial, the effect of bST which was evaluated included both indirect and direct effects that bST may have on CM. By not utilizing a independent variable for milk production in our statistical model, the effect ' of bST on CM that we evaluated included both possible mechanisms combined. While herd variation in the data was apparent, overall, no differences were found for any measured feature of CM between the study groups in our trial. It is important to note that this study was conducted on well-managed farms that had controlled contagious mastitis and also experienced relatively low rates of CM caused by environmental organisms. Therefore, it would not be appropriate to extrapolate our data to farms which have higher prevalences of contagious mastitis organisms and incidences of CM above those experienced by our study herds. I concluded that regardless of the mechanism considered, bST was not associated with an increase in CM incidence, milk discarded because of therapy for CM, or culling because of mastitis in the herds studied. 45 REFERENCES Bartlett, P. C., G. Y. Miller, S. E. Lance, and L. E. Heider. 1992. Clinical mastitis and intramammary infections on Ohio dairy farms. Prev. Vet. Med. 12:59-71. Bartlett, P. C ., J. Van Wijk. 1991. Temporal patterns of lost milk production following clinical mastitis in a large Michigan Holstein herd. J. Dairy Sci. 74:1561-1572. Cole, J. W., P. J. Eppard, B. G. Boysen, K. S. Madsen, R. H. Sorbet, M. A. Miller, R. L. Hintz, W. E. Ribelin, B. G. Hammond, R. J. Collier, and G. M. Lanza. 1992. Response of dairy cows to high doses of a sustained-release bovine somatotropin administered during two lactations. 2. Health and reproduction. J. Dairy Sci. 75:11-123. Downer, J. V., D. L. Patterson, D. W. Rock, W. V. Chalupa, R. M. Cleale, J. L. Firkins, G. L. Lynch, J. H. Clark, B. O. Brodie, B. F. Jenny, and R. De Gregorio. 1993. Dose titration of sustained-release recombinant bovine somatotropin in lactating dairy cows. J. Dairy Sci. 76:1125-1136. Epi Info“, Version 6.03, 1996. Centers for Disease Control, Atlanta, GA. Eppard, R. J., 8. Hudson, W. J. Cole, R. L. Hintz, G. F. Hartnell, T. W. Hunter, L. E. Metzger, A. R. Torkelson, B. G. Hammond, R. J. Collier, and G. M. Lanza. 1991. Response of dairy cows to high doses of a sustained-release bovine somatotropin administered during two lactations. 1. Production response. J. Dairy Sci. 74:3807-3821. Erskine, R. J ., R. J. Eberhart, L. J. Hutchinson, S. B. Spencer, and M. A. Campbell. 1988. Incidence and types of clinical mastitis in dairy herds with high and low somatic cell counts. JAVMA l90(11):1411-1416. Excel“, Version 5.0. 1993. Microsoft Corporation, Redmond, WA. Firat, M. Z. 1993. An investigation into the effects of clinical mastitis on milk yield in dairy cows. Livest. Prod. Sci. 36:311-321. Hoblet, K. H., G. D. Schnitkey, D. Arbaugh, J. S. Hogan, K. L. Smith, P. S. Schoenberger, D. A. Todhunter, W. D. Hueston, D. E. Pritchard, G. L. Bowman, L. E. Heider, B. L. Brockett, and H. R. Conrad. 1991. Costs associated with selected preventive practices and with episodes of clinical mastitis in nine herds with low somatic cell counts. JAVMA 199(2): 190-196. 46 Houben, E. H. P., A. A. Dijkhuizen, J. A. M. Van Arendonk, and R. B. M. Huirne. 1993. Short and long-terrn production losses and repeatability of clinical mastitis in dairy cattle. J. Dairy Sci. 76:2561-2578. Kleibaum, D. G., Kupper, L. L., and H. Morgenstern. 1982. Epidemiologic Research. Lifetime Learning Publications, Belmont, CA. Kronfeld, D. S. 1994. Health management of dairy herds treated with bovine somatotropin. JAVMA 204(1):116-130. Lissemore, K. D., K. E. Leslie, B. W. McBride, J. H. Burton, A. R. Willan, and K. G. Bateman. 1991. Observations on intramammary infection and somatic cell counts in cows treated with recombinant bovine somatotropin. Can. J. Vet. Res. 55:196-198. Miller, G. Y., P. C. Bartlett, S. E. Lance, J. D. Anderson, and L. E. Heider. Costs of clinical mastitis and mastitis prevention in dairy herds. 1993. JAVMA 202(8): 1230-1236. National Mastitis Council. 1987. Field Handbook on Bovine Mastitis. W. D. Hoards and Sons Co., Fort Atkinson, WI. Oltenacu, P. A., I. Ekesbo. 1994. Epidemiological study of clinical mastitis in dairy cattle. Vet. Res.925:208-212. Pell, A. N., D. S. Tsang, B. A. Howlett, M. T. Huyler, V. K. Meserole, W. A. Samuels, G. F. Hartnell, and R. L. Hintz. 1992. Effects of a prolonged-release formulation of sometribove (n-methionyl bovine somatotropin) on Jersey cows. J. Dairy Sci. 75:3416-3431. PosilacR product label. 1993. Monsanto Co., St. Louis, MO. Remond, B., M. Cisse, A. Ollier, and Y. Chilliard. 1991. Slow release somatotropin in dairy heifers and cows fed two levels of energy concentrate. 1. Performance and body condition. J. Dairy Sci. 74:1370-1381. Rijpkema, Y. S., L. van Reeuwijk, and D. L. Hard. 1990. Responses of dairy cows to treatment with sometribove (r-BST) during three consecutive years. Livest. Prod. Sci. 26:193-216. SAS System for Windows“, Version 6.10, 1994. SAS Institute, Cary, NC. 47 Skarda, J ., E. Markalous, J. Slaba, P. Krejci, O. Skardova, and J. Zednik. 1992. Effect of methionyl bovine somatotropin in a prolonged-release vehicle on milk production, hormone profiles and health in dairy cattle. J. Dairy Res. 59:499- 506. Statistix“, Version 4.1, 1994. Analytical Software, St. Paul, MN. Thomas, J. W., R. A. Erdman, D. M. Galton, R. C. Lamb, M. J. Arambel, J. D. Olson, K. S. Madsen, W. A. Samuels, C. J. Peel, and G. A. Green. 1991. Responses by lactating cows in commercial dairy herds to recombinant bovine somatotropin. J. Dairy Sci. 74:945-964. Weigher, B. J., D. W. Hird, W. M. Sischo, J. C. Hohnes, C. Danaye-Elrni, C. W. Palmer, and W. W. Utterback. 1990. Veterinary and non-veterinary costs of disease in 29 California dairies participating in the National Animal Health Monitoring System from 1988 to 1989. JAVMA 196(12):]945-1949. Whitaker, D. A., E. J. Smith, J. M. Kelly, and L. S. Hodgson-Jones. 1988. Health, welfare and fertility implications of the use of bovine somatotropin in dairy cattle. Vet. Rec. 122:503-505. White, T. .C., et. a1. 1994. Clinical mastitis in cows treated with sometribove (recombinant bovine somatotropin) and its relationship to milk yield. J. Dairy Sci. 77 :2249-2260. 48 CHAPTER FOUR RECOMBINAN T BOVINE SOMATOTROPIN : ASSOCIATION WITH REPRODUCTIVE PERFORMANCE IN DAIRY COWS ABSTRACT A clinical trial was performed to determine the effect of bST on reproductive performance of dairy cows in four Michigan dairy farms when bST was used according to labeled directions. Holstein cows (11 = 555) at four Michigan dairy farms were randomly assigned to receive bST or to serve as untreated controls. Bovine somatotropin (500 mg) was administered subcutaneously every 14 days beginning at 63 to 69 days of lactation and continuing until approximately 21 days prior to the end of lactation or until the producer removed the animal from the herd. A total of 229 , (42.57%) animals consisting of 112 (48.91%) controls and 117 (51.09%) bST-treated cows were not diagnosed pregnant and were considered as having reproductive failure by 150 days of lactation and 92 (17.10%) of the animals consisting of 45 (48.91%) controls and 47 (51.09%) bST-treated cows had reproductive failure at the end of their lactation. No significant difference in the incidence of reproductive failure existed between'the study treatment groups during either period of time. Cows with cystic ovarian disease had more instances of reproductive failure at 150 days of lactation than cows without this disease. Cows with dystocia, twin births, metritis, displaced abomasum and cystic ovarian disease had more instances of reproductive failure by the end of lactation than cows without these problems. The bST-treated cows developed 49 more nonspecific off feed conditions during the study lactation and had more twin births at the calving following bST administration compared to control cows. Measured and calculated reproductive parameters did not differ between the study treatment groups. Logistic regression analysis found no significant effect of bST on reproductive failure; the odds ratios for the effect of bST treatment on reproductive failure at 150 days of lactation and reproductive failure at the end of lactation were 0.95 (95% CI: 0.67 to 1.36) and 0.95 (95% CI: 0.59 to 1.54), respectively. Overall, bST did have a significant impact on reproductive performance in the herds studied. 50 INTRODUCTION Milk production and reproductive performance are two important determinants of dairy farm profitability (Ferguson, 1996; Lean et al., 1989). As milk production rises, costs due to reproductive problems usually increase (Cole et al., 1991; Cole et al., 1992; Gong et a1. , 1991). Because bovine somatotropin has been found to significantly increase milk production (Hartnell et al. , 1991, Downer et a1. , 1993, Thomas et al., 1991, Eppard et al., 1991; Pell et al., 1992; Rijpkema et al., 1990), it has been suggested that some of the financial gains resulting from bST-assisted milk production may be offset by a concurrent reduction in reproductive performance (Lean et al., 1989). Bovine somatotropin might affect reproductive performance by extending the period of negative energy balance in cows. A negative correlation between energy balance and indicators of reproductive performance (e.g. , days to first ovulation, days to first estrus and days open) has been reported (Butler and Smith, 1995; Cole et al. , 1991, Harrison et al. , 1990). When used according to label directions, commercially- available bST caused negative energy balance for approximately 7 weeks following the initiation of treatment (Bauman et al. , 1989, Peel and Bauman, 1987). This time period coincides with the desired time of conception on most farms. While delaying bST treatment until 63 days in milk increases the percentage of cows in positive energy balance when bST treatment is initiated, the increased number of days that cows spend in negative energy balance resulting from the use of bST could nevertheless decrease 51 future reproductive performance. The relationship between milk production and reproductive performance may be due, in part, to the positive association between increased milk production and diseases such as clinical mastitis, ketosis, retained placenta, metritis, abortion, cystic ovarian disease and lameness which can reduce reproductive performance (Cullor, 1991; Deluyker et al., 1991; Etherington et al., 1996; Garverick, 1997; Oresnik, 1995). By increasing milk production, bST may be associated with increased incidence of these diseases as well as a consequent decreased in reproductive performance. Bovine somatotropin has been widely studied and many pre-approval trials have been conducted. However, most of these trials had an insufficient sample size to detect differences in reproductiveperformance (Cole et al., 1992). Nevertheless, some of these studies concluded that bST significantly increased the number of days open (Burton etal., 1990; Cole et al. , 1991). Alternatively, other studies found a negligible increase in days open associated with use of bST (Hansen et al., 1993; Leonard et al. , 1990; Rijpkema et al., 1990; Zhao et al., 1992). Because bST alters the shape of the lactation curve, the usual relationship between days open and profitability (Ferguson, 1996) is distorted. Therefore, the cost of additional days open needs to be re-evaluated in cows being treated with bST. The effect of bST on the incidence of twin births remains controversial. Some studies have observed no association between twinning and bST treatment, while others have found that twinning was associated with use of bST (Burton et al. , 1990; Esteban et a1. , 1994). In addition to the decreased calf value incurred because of freemartinism, 52 cows that deliver twins are at greater risk to develop other diseases such as retained placenta and metritis (N ielen et al. , 1989). As result of twin births, affected cows produce less milk than herdmates that deliver single calves, which leads to economic loss (Beerepoot and Dykhuizen, 1992; Nielen et al., 1989). The objective of this study was to determine the effect of bST on reproductive performance of dairy cows when bST was used according to label directions (Monsanto Company, 1993). SPECIFIC METHODS A voluntary 45 day postpartum waiting period prior to breeding was followed at all farms. Cows that were inseminated prior to the end of the voluntary waiting period were excluded from the analysis. This resulted in 17 cows (10 control and 7 bST- treated cows) being eliminated from the analysis. Therefore, the final analysis utilized a total of 538 cows (261 control and 277 bST-treated cows). DairyComp 305R (Valley Agricultural'Software, 1990) was used for management of cow-specific data by all study herds. Recorded data included calving dates, artificial insemination dates, results of veterinary reproductive examinations, disease conditions as well as disease and reproductive treatments. Reproductive management regarding estrus detection, inseminations and reproductive examinations was directed by output from this computer program based on the same parameters for both the bST-treated and control cows. For example, selection of cows for insemination was made without regard to study treatment group. 53 The HeatwatchR (DDx Incorporated, 1992) program was used to record estrus events for two of the herds, while visual observation for estrus detection was used at the remaining two herds. All farms artificially inseminated cows approximately 12 hours after observed standing estrus. Prostaglandin F2“ was used to synchronize estrus on all four farms. . The SAS System for WindowsR (SAS Institute, 1994) and Epi Info (Centers for Disease Control, 1996) were used for statistical analysis of reproductive data. As an initial screening process, parameters of reproductive efficiency were calculated based on the following equations: Average days to first service was defined as the total days from calving to first insemination divided by the total number of cows inseminated; average days open was defined as the total days from calving to conception divided by the total number of cows conceiving (becoming pregnant); first service conception rate was defined as the total number of cows pregnant to first insemination divided by the total number of cows inseminated; overall conception rate was defined as the total number of cows pregnant for all inseminations divided by the total number of inseminations in pregnant cows; pregnancy rate was defined as the total number of cows that became pregnant divided by the total number of cows in the breeding group; heat detection rate was defined as the total number of inseminations in pregnant cows past the voluntary waiting period divided by the total number of days from the voluntary waiting period to conception (pregnancy); services per conception was defined as the total number of inseminations in pregnant cows divided by the total number of cows conceiving (becoming pregnant); calving interval was defined as the 54 total days from calving to begin the study lactation until the next calving. Statistical analysis of the effect of bST on reproduction was conducted in the following manner: average days open, days to first service and calving interval were compared between the study groups by Student's t-test. Conception rate (first service), pregnancy rate and reproductive failure were evaluated by chi-square analysis. The estrus detection rate and services per pregnancy (overall conception rate) were compared by binomial distribution (incidence density) testing. Chi-square was also used to compare the incidence of twin birth between the study treatment groups. The variable for reproductive failure at 150 days of lactation was defined as follows: 1) cows were not confirmed pregnant by 150 days of lactation despite being a member of the breeding herd or 2) cows were not inseminated while in the breeding herd for at least 105 days (until at least 150 days of lactation). Cows that were culled or died before 150 days of lactation were considered reproductive failures if they met either or both of the above criteria; otherwise, they were considered reproductive successes. A total of 41 cows either died or were culled before 150 days of lactation (24 control and 17 bST-treated cows); of these, 20 cows (11 control and 9 bST-treated cows) were considered to be cases of reproductive failure and 21 cows (13 control and 8 bST-treated cows) were classified as reproductive successes. The variable for reproductive failure at the end of lactation was defined similarly except that cows not diagnosed pregnant were not classified as failures until the end of their lactation. Cows which aborted were not considered to have reproductive failure and contributed data only up until the time of abortion. Diseases which may haVe affected either variable for 55 reproductive failure were analyzed in a bivariate analysis to determine whether they were possible confounding or intervening variables. Bivariate analysis was used to assess the degree of collinearity between diseases and study treatment group, both before and after the initiation of bST treatment. Based on the results of initial screening, and because reproductive failure encompasses and summarizes all reproductive performance measures, it was chosen for further analysis. Multivariable logistic regression models were constructed to evaluate reproductive performance using reproductive failure as the dependent variable. The independent variables of farm, parity, season of calving, pretreatment milk production, occurrence of twin birth and instances of diseases were investigated as potential confounding variables. Disease presence was determined in the same time period as reproductive failure used in the model. The variable for study treatment group (treatment vs. control) was forced into the model to assess the effect of bST. The model analyzed was Y = 1 1 +e-f(REPRO) where Y = probability of RF, e = base of natural log, and f (REPRO) = intercept + FM + PARITY + GROUP + PREMILK + DA+ CYSTIC + TWIN + DIARRHEA + OFFEED + ABORT + CM + MF + KET + DYSTOCIA + RP + MET + error, where the independent variables are defined as FM = farm of origin (1, 2, 3, or 4); PARIT Y = parity (lactation) (1 through 8); GROUP = study group (1 = bST treatment group, 0 = no treatment (control group); PREMILK = (milk production 56 before the bST treatment period ( < 63 DIM)); DA = displaced abomasum; CYSTIC = cystic ovarian disease; TWIN = twin birth; DIARRHEA = nonspecific diarrhea; OFFEED = nonspecific off feed condition; ABORT = abortion; CM = clinical mastitis; MF = milk fever; KET = ketosis; DYSTOCIA = dystocia; RP = retained placenta; MET = metritis. The farm effect was investigated as both a fixed and a random effect; this variable was included in the models regardless of significance. Other independent variables were eliminated based on insignificance at P > 0.10. All two-way, biologically-relevant interactions, defined according to the additive definition of interaction (Rothman, 1986), were assessed as well as the quadratic effect of the continuous variables for parity and pretreatment milk production. Models for reproductive failure at 150 days and at the end of lactation were constructed by a manual step-down method and were confirmed by a manual step-up method. RESULTS For the 538 cows included in the final analysis, no significant effect of bST was found for any of the measured reproductive parameters (Table 4.1). However, a higher pregnancy rate was observed in primiparous cows compared to multiparous cows (P = 0.0001) that was independent of bST treatment effect. In bivariate analysis, cows with cystic ovarian disease had a significantly greater risk of reproductive failure at 150 days of lactation (P = 0.001), and the combined effect of all diseases studied resulted in significantly greater risk of reproductive failure 57 at 150 days of lactation (P = 0.03). Cows that experienced dystocia, twin births, metritis, displaced abomasum or cystic ovarian disease also had a significantly greater risk of reproductive failure at the end of lactation (P < 0.05) and the combined effect of all diseases resulted in significantly greater risk of reproductive failure at the end of lactation (P = 0.04). TABLE 4.1: Reproductive parameters for cows administered 500 mg of bST at 14-day intervals from 63 days of lactation until approximately 21 days before the end of lactation and in untreated controls. ' Parameter bST Control Total P -value Average days to first (261) 95.3 (240) 98.3 (501) 96.8 0.52 service ‘ (n). days Average days open (222) 131.6 (203) 134.3 (425) 133.0 0.72 (n). days Calving interval, (192) 418.0 (175) 416.2 (367) 417.1 0.84 (n). days ‘ Conception rate first (261) 45.82 (240) 49.99 - (501) 45.43 0.43 service (n), % Conception rate, all (261) 54.35 (240) 52.08 (501) 53.19 0.64 services (n), % Estrus detection rate (261) 44.69 (240) 44.46 (501) 44.58 0.96 (n). % Pregnancy rate (277) 80.14 (261) 77.78 (538) 79.00 0.50 (n). % Number of services per (261) 1.84 (240) 1.92 (501) 1.88 0.64 pregnancy (n), n Reproductive failure: (277) 51.09 (261) 48.91 (538) 42.57 0.88 150 days lactation (n). % Reproductive failure: (277) 51.09 (261) 48.91 (538) 17.10 0.93 end of lactation (n), % 58 In the bivariate comparison of diseases and study group, bST-treated cows experienced significantly more nonspecific off feed conditions than did control cows. However, the incidence of all diseases combined was not different between study groups (Table 4.2). In the bivariate analysis of diseases occurring only after the initiation of bST treatment (> 63 days of lactation), incidence of abortion (P = 0.11) and nonspecific off feed conditions (P = 0.11)) were marginally greater in the bST-treated cows than in the controls. Overall, the bST-treated cows had marginally more disease during the treatment period than the controls (P = 0.10). Bivariate analysis also demonstrated an association between the incidence of twin births and use of bST at the calving following the lactation under study; cows treated with bST had more twin births than the control cows (P = 0.05).. However, this difference was not observed at the calving occurring immediately before the initiation of this study (Table 4.2). 59 TABLE 4.2: Cows with a disease occurring at anytime during lactation when administered 500 mg of bST at 14-day intervals from 63 days of lactation until approximately 21 day before the end of lactation and in untreated controls. Lactation Incidence bST Control Total Incidence Disease condition (n) % (n) % (n) % P-value Displaced abomasum (16) 0.058 (16) 0.061 (32) 0.060 0.86 Cystic ovarian disease (28) 0.101 (20) 0.077 (48) 0.089 0.32 Abortion (26) 0.094 (16) 0.061 (42) 0.078 0.16 Nonspecific diarrhea (3) 0.011 (5) 0.019 (8) 0.015 0.43 Nonspecific off feed - (20) 0.072 (9) 0.035 (29) 0.054 0.05 Twin birth2 (19) 0.069 (13) 0.050 (32) 0.060 0.36 Milk fever (4) 0.014 (7) 0.027 (11) 0.020 0.31 Clinical mastitis (46) 0.166 (51) 0.195 (97) 0.180 0.38 Ketosis (21) 0.076 (12) 0.046 (33) 0.061 0.15 Dystocia (3) 0.011 (2) 0.008 (5) 0.009 0.70 Retained placenta _ (17) 0.061 (10) 0.038 (27) 0.050 0.22 Metritis (54) 0.195 (47) 0.180 (101) 0.188 0.66 Any disease occurrenceb (153) 0.539 (133) 0.491 (286) 0.532 0.26 Total (277) 0.515 (261) 0.485 (538) 1.000 - 2Incidence of twin birth prior to the study lactation bDiseases do not total because of multiple diseases occurring in the same cow Using multiple logistic regression to model reproductive failure at 150 days of lactation, the odds ratio and 95% confidence interval for the bST treatment effect was 0.95 (95% CI: 0.67 to 1.36) (Table 4.3). Model fit was assessed by the Hosmer- 60 Lemeshow test; the goodness-of-fit statistic was 2.33 with a degrees of freedom of six and a corresponding probability of P = 0.89 indicating good model fit. TABLE 4.3: Logistic regression analysis of reproductive failure at 150 days of lactation for cows administered 500 mg of bST at 14—day intervals from 63 days of lactation until approximately 21 days before the end of lactation and in untreated controls. Source of variation df Parameter estimate P Odds ratio (95 % CI) Farm 1 1 -0.0241 0.9422 0.98 (0.51 to 1.87) Farm 2 1 0.5438 0.0252 1.72 (1.07 to 2.77) Farm 3 1 0.6450 0.0166 1.91 (1.12 to 3.23) Twin birth 1 0.6895 0.0694 1.99 (0.95 to 4.19) Cystic ovarian disease 1 1.3654 0.0006 3.92 (1.79 to 8.58) bST treatment . 1 -0.0476 0.7911 0.95 (0.67 to 1.36) TABLE 4.4: Logistic regression analysis of reproductive failure by the end of lactation for cows administered 500 mg of bST at 14-day intervals from 63 days of lactation until approximately 21 days before the end of lactation and in untreated controls. Source of variation df Parameter estimate P Odds ratio (95 % CI) Farm 1 1 -03077 0.4395 0.74 (0.34 to 1.60) Farm 2 1 -1.0122 0.0021 0.36 (0.19 to 0.69) Farm 3 1 -01545 0.6394 0.86 (0.45 to 1.64) Parity 1 0.2901 0.0013 1.34 (1.12 to 1.59) Twin birth 1 0.8802 0.0368 2.41 (1.06 to 5.51) Displaced abomasum 1 0.9850 0.0183 2.68 (1.18 to 6.07) Cystic ovarian disease 1 0.9575 0.0079 2.61 (1.29 to 5.28) Dystocia 1 2.1427 0.0314 8.52 (1.28 to 60.00) bST treatment 1 0.0474 0.8457 0.95 (0.59 to 1.54) 61 Similarly, multiple logistic regression found no significant effect of bST on reproductive failure at the end of lactation (Table 4.4). The odds ratio and 95 % confidence interval for bST treatment was 0.95 (95% CI: 0.59 to 1.54). Model fit was assessed by the Hosmer-Lemeshow test; the goodness-of-fit statistic was 3.86 with a degrees of freedom of seven and a corresponding probability of P = 0.80 indicating that the model fit the data well. Thus, in these models, cows treated with bST were 5% less likely to experience reproductive failure than the controls. However, these estimates were not statistically significant. When the independent variable for farm was considered as a random effect variable, the coefficients and significance of bST did not change significantly. Therefore, the variable for farm effect was considered to be a fixed effect in the final models. DISCUSSION The objective was to determine theeffect of bST on reproductive performance in dairy cows on four Michigan farms when bST was used according to label directions (Monsanto Company, 1993). Of the 580 cows originally selected for this study, 22 were in poor health at the time when bST treatment was to begin (63 to 69 days in milk) and therefore were excluded from the study (Monsanto Company, 1993). Also, 3 cows were excluded for clerical errors unrelated to their health status. An additional 17 cows were excluded from the analysis because they were inseminated prior to the end of the established voluntary waiting period. Because less than 8% of the cows were 62 excluded from the analysis, and because both bST-treated and control cows were excluded based on the same criteria, it is unlikely that these exclusions created a significant bias in the data. No blinding of study farm personnel was used in this trial. I believe that it would have been futile to attempt blinding because the study herds had daily milk weights available for all cows. Therefore, herd managers could easily determine which cows were bST-treated based on their initial production response. I discussed with herdsmen the importance of equal handling of cows within each study group. While it is possible that some bias may have occurred as a result of herd managers not being blinded, I believe the data supports our conclusion that a significant bias did not occur from a lack of blinding. It has been documented that several diseases of dairy cows can reduce reproductive performance (Cullor, 1991; Harman et al. , 1996; Mohammed et al., 1991; Etherington et al., 1996). Because failure of a cow to conceive is one of the most common reasons given for culling cows (Esslemont and Kossabati, 1997; Etherington et al. , 1996), the occurrence of diseases which adversely affect reproduction can thereby decrease the productive lifetime (longevity) of cows (Galligan, 1997). Diseases that have been associated with decreased reproductive performance are clinical mastitis, ketosis, retained placenta, metritis, abortion, cystic ovarian disease and lameness (Cullor, 1991; Etherington et al., 1996; Garverick, 1997). In the current study, the lactational incidence of most diseases were higher than in herds observed in Ontario (Etherington et al. , 1996) and were more similar to those observed in Finland (Harman 63 et al., 1996) and California (Deluyker et al., 1991). The differences observed in disease rates may have been due to either under-reporting or over-reporting of disease, differences in disease diagnosis by herd managers in these study herds, or a result of real differences in disease rates among herds in different geographical locations. While this study was not specifically designed to detect differences in disease rates between the study groups, the disease incidences observed in this study were comparable to those found in other studies. In the current study, dystocia, twin birth, metritis, displaced abomasum and cystic ovarian disease were associated with increased reproductive failure at the end of lactation. However, when the effect of confounding variables was removed by use of multivariate logistic regression, metritis did not have a significant deleterious effect. lameness was .not addressed in this study, but others have associated it with reduced reproductive performance (Grohn et al., 1994; Harman et al., 1996; Wells et al.,, _ 1995). However, in another study, bST was not associated with lameness (Wells et al., 1995). The underlying causes of the observed nonspecific off feed and diarrheal conditions were not known in this study. The occurrence of more frequent off feed conditions in bST-treated cows was, however, observed in one preapproval trial (Cole et al., 1992). It was speculated that higher dry matter intakes in bST-treated cows compared to control cows may have increased the risk of digestive disturbance (Cole et al. , 1992). Diarrheal conditions probably had multiple causes such as Bovine Virus Diarrhea and J ohnnes disease, however, the occurrence of diarrheal conditions did not 64 differ between the treated and control groups. The effect of bST on the incidence of twin births has varied among farms and studies. In the present study, a significant positive effect of bST on twin births was found. Similar increases in twinning rates associated with use of bST have been found in other studies (Cole et al., 1991; Esteban et al., 1994). However, the mechanism responsible for bST-associated twinning is not known. Even though increased numbers of small (preovulatory) follicles have been found in bST-treated cows, an increased ovulation rate has not been observed (De La Sota et al., 1993; Lucy et al., 1994). Multiple ovulations may explain the increased twinning that some workers have observed. However, studies to date have failed to document their occurrence. Nevertheless, the FDA recently granted the Monsanto Company, based on data from a post-approval monitoring project, permission to remove a cautionary statement from their bST product label regarding the risk of twin births (Monsanto Company, 1993). The overall occurrence of reproductive failure at the end of lactation was similar to that observed in other studies (Cole et al.,, 1991; Esteban et al. , 1994). As expected, when a shorter period of time for conception was allowed (reproductive failure at 150 days of lactation), a greater number of reproductive failures was observed. However, bST treatment was not associated with either time period that was defined for reproductive failure. This finding is similar to that of other studies in which bST was not observed to influence reproductive performance. Moreover, similar to that of other studies, no effect of bST was found on any other measure of reproductive efficiency (Ceelen, 1995; Hansen et al., 1993; Leonard et al., 1990; Pellet al., 1992; Remond et 65 al., 1991; Rijpkema et al. , 1990). However, other studies found that bST had a negative effect on reproductive parameters (Burton et al., 1990; Burton: et al., 1990; Butler and Smith, 1989; Cole et al., 1992; Esteban et al., 1994; Hemken et al., 1991; Oldebroek and Garsen, 1993; Whitaker et al., 1988; Zhao et al., 1992). Most recently, the Monsanto Company completed a post-approval monitoring project required by the FDA when it approved bST for commercial use. For reproductive analysis, this study involved 1,126 cows (564 control and 562 treated cows) in 28 herds across the US, and found a significant increase in the number of days open for bST-treated primiparous cows and a decreased pregnancy rate for bST-treated multiparous cows compared with that of the control cows (Monsanto Company, 1997). However, the present study did not observe any reduction in measures of reproductive performance for cows of any parity. The observed lack of effect of bST on reproduction may have been due to the smaller number of cows in our study. Milk production has long been associated with reproductive performance in dairy cattle with reproductive performance tending to decrease as production increases (Etherington et al., 1996; Ferguson, 1996; Grohn et al., 1994; Harrison et al., 1990; Lean et al. , 1989). In the present study, milk yield prior to 63 days of lactation was not a significant predictor of reproductive failure at either 150 days of milk or at the end of lactation. Milk yield during the bST treatment period was intentionally not used as an independent variable in our logistic regression models. This is because an underlying assumption of our statistical models was that increased milk production in the bST-treated group was the direct result of bST treatment. Therefore, milk yield 66 was considered to be an intervening variable between bST and reproductive failure, and, as such, it is not appropriate to include intervening variables as independent variables in regression models. This is because the inclusion of intervening variables in multivariate models can mask associations between independent variables (risk factors) and the disease being studied (dependent variable) (Kleinbaum etal., 1982). However, others have observed that milk production was not nearly as important in determining reproductive performance as was dry matter intake. By being the most important influence on energy balance, dry matter intake determines the reproductive performance of cows more directly than does milk production (V ila-Godoy et al., 1988). CONCLUSIONS Bovine somatotropin, either by directly altering energy metabolism and nutrient apportionment or through association with diseases which affect reproduction, has the potential to adversely impact reproductive performance in dairy cows. Therefore, it is important to evaluate bST usage for any negative impact it may have on reproduction. In this study, bST-treated cows had a marginal increase in abortions and a significant increase in the incidence of twin births at the subsequent calving. However, despite being associated with nonspecific off feed conditions and twin births, bST did not significantly affect the ability of cows to become pregnant in‘ the herds studied. Dairy producers are encouraged to monitor their cows for any change in disease rates or reproductive performance that may occur with the use of bST. 67 REFERENCES Bartlett, P. C., G. Y. Miller, S. E. Lance, and L. E. Heider. 1992. Clinical mastitis and intramammary infections on Ohio dairy farms. Prev. Vet. Med. 12:59-71. Bauman, D. E., D. L. Hard, B. A. Crooker, M. S. Partridge, K. Garrick, L. D. Sandles, H. ' N. Erb, S. E. Franson, G. F. Hartnell, and R. L. Hintz. 1989. Long-terrn evaluation of a prolonged-release formulation of n-methionyl bovine somatotropin in lactating dairy cows. J. Dairy Sci. 72:642-651. Beerepoot, G. M. M., and A. A. Dykhuizen. 1992. The economics of naturally occuring twinning in dairy cattle. J Dairy Sci. 75:104-4-1051. Burton, J. L., B. W. McBride, J. H. Burton, and R. G. Effert. 1990. Health and reproductive performance of dairy cows treated for up to two consecutive lactations with bovine somatotropin. J. Dairy Sci. 73:3258-3265. Burton, J. H., G. K. MacLeod, B. W. McBride, J. L. Burton, K. Bateman, I. McMillan, and R. G. Eggert. 1990. Overall efficacy of chronically administered recombinant bovine somatotropin to lactating dairy cows. J. Dairy Sci. 73:2157-2167. Butler, W. R., and R. D. Smith. 1989. Interrelationships between energy balance and postpartum reproductive function in dairy cattle. J. Dairy Sci. 72:767-783. Ceelen, H. J. 1995. Bovine somatotoropin and cow health - what are the facts? Can. Vet. J. 36:25-27. Cole, W. J ., K. S. Madsen, R. L. Hintz, and R. J. Collier. 1991. Effect of recombinantly- derived bovine somatotropin on reproductive performance of dairy cattle. Therio. 36:573-595. Cole, J. W., P. J. Eppard, B. G. Boysen, K. S. Madsen, R. H. Sorbet, M. A. Miller, R. L. Hintz, W. E. Ribelin, B. G. Hammond, R. J. Collier, and G. M. Lanza. 1992. Response of dairy cows to high doses of a sustained-release bovine somatotropin administered during two lactations. 2. Health and reproduction. J. Dairy Sci. 75:11-123. Cullor, J. S. 1991. Mastitis in dairy cows: does it hinder reproductive performance? Vet Med. 86: 830—832. Dairy Comp 305“, Version 3.1. Valley Agricultural Software. Tulare CA. 1990. 68 De La Sota, R. L., M. C. Lucy, C. R. Staples, and W. W. Thatcher. 1993. Effects of recombinant bovine somatotropin (Sometribove) on ovarian function in lactating and nonlactating dairy cows. J. Dairy Sci. 76:1002-1013. Deluyker, H. A., J. M. Gay, L. D. Weaver, and A. S. Axari. 1991. Change in milk yield with clinical diseases for a high producing dairy herd. J. Dairy Sci. 74. 463-445. Downer, J. V., D. L. Patterson, D. W. Rock, W. V. Chalupa, R. M. Cleale, J. L. Firkins, G. L. Lynch, J. H. Clark, B. O. Brodie, B. F. Jenny, and R. De Gregorio. 1993. Dose titration of sustained-release recombinant bovine somatotropin in lactating dairy cows. J. Dairy Sci. 76:1125-1136. Epi Info, Version 6.03. Centers for Disease Control, Atlanta GA. 1996. Eppard, R. J ., S. Hudson, W. J. Cole, R. L. Hintz, G. F. Hartnell, T. W. Hunter, L. E. Metzger, A. R. Torkelson, B. G. Hammond, R. J. Collier, and G. M. Lanza. 1991. Response of dairy cows to high doses of a sustained-release bovine somatotropin administered during two lactations. 1. Production response. J. Dairy Sci. 74:3807-3821. Esslemont, R. J ., and M. A. Kossaibati. 1997. Culling in 50 dairy herds in England. Vet. Rec. 140:36-39. Estban, E., P. H. Kass, L. D. Weaver, J. D. Rowe, C. A. Holrnberg, C. E. Franti, and H. F. Trout. 1994. Interval from calving to conception in high producing dairy cows treated with recombinant bovine somatotropin. J. Dairy Sci. 7 7 2549-2561. - Esteban, E., P. H. Kass, L. D. Weaver, J. D. Rowe, C. A. Holrnberg, C. E. Franti, and H. F. Trout. 1994. Pregnancy incidence in high producing dairy cows treated with recombinant bovine somatotropin. J. Dairy Sci. 77:468-481. Esteban, E., P. H. Kass, L. D. Weaver, J. D. Rowe, C. A. Holrnberg, C. E. Franti, and H. F. Trout. 1994. Reproductive performance in high producing dairy cows treated with recombinant bovine somatotropin. J. Dairy Sci. 77:3371-3381. Etherington, W. G., M. L. Kinsel, and W. E. Marsh. 1996. Relationship of production to reproductive performance in ontario dairy cows: herd level and individual animal descriptive statistics. Therio. 46:935-959. Excel“, Version 5.0. Microsoft Corporation. Redmond WA. 1993. 69 Ferguson, J. D. 1996. Diet, production, and reproduction in dairy cows. An. Feed Sci. Tech. 59:173-184. Forar, A. L., J. M. Gay and D. D. Hancock. 1995. The frequency of endemic fetal loss in dairy cattle: A Review. Therio. 43:989-1000. Galligan, D. M. 1997. Extended calving intervals with the use of bST. Protiva Reproductive Management Symposium (in conjunction with the Tri-State Dairy Nutrition Conference) p. 11-17. Garverick, H. A. 1997. Ovarian follicular cysts in dairy cattle. J Dairy Sci. 80:995- . 1004. Gong, J. G., T. Bramley, and R. Webb. 1991. The effect of recombinant bovine somatotropin on ovarian function in heifers: follicular populations and peripheral hormones. Biol Reprod. 45:941-949. Grohn, Y. T., J. A. Hertl, and J. L. Harman. 1994. Effect of early lactation milk yield on reproductive disorders in dairy cows. Am. J. Vet. Res. 55(11):1521- 1528. Hansen, W. P., D. E. Otterby, and J. G. Linn. 1993. Multi-farm use of bovine somatotropin for two consecutive lactations and its effects on lactational performance, health, andreproduction. J. Dairy. Sci. 77:94-110. Harman, J. L., Y. T. Grohn, H. N. Erb, and G. Casella. 1996. Event-tirne analysis of the effect of season on parturition, parity, and concurrent disease on parturition-to-conception interval in dairy cows. Am. J. Vet. Res. 57(5):640- 645. Harrison, R. O., S. P. Ford, J. W. Young, A. J. Conley, and A. E. Freeman. 1990. Increased milk production versus reproductive and energy status of high producing dairy cows. J. Dairy Sci. 73:2749-2758. Hartnell, G. F., S. E. Franson, D. E. Bauman, H. H. Head, J. T. Huber, R. C. Lamb, K. S. Madsen, W. J. Cole, and R. L. Hintz. 1991. Evaluation of sometribove in a prolonged-release system in lactating dairy cows - production responses. J Dairy Sci. 74:2645-2663. HeatWatch“, Version 1.0. DDx Inc.. Denver CO. 1992. 70 Hemken, R. W., R. J. Harmon, W. J. Silvia, W. B. Tucker, G. Heersche, and R. G. Eggert. 1991. Effect of dietary energy and previous bovine somatotropin on milk yield, mastitis, and reproduction in dairy cows. J. Dairy Sci. 74:4265- 4272. Judge, L. J., R. J. Erskine, and P. C. Bartlett. 1997. Recombinant bovine somatotropin and clinical mastitis: incidence, discarded milk following therapy, and culling. J Dairy Sci. 80:3212-3218. Kleibaum, D. G., Kupper, L. L., and H. Morgenstern. 1982. Epidemiologic research. Lifetime Learning Publications, Belmont, CA. Lean, I. J ., J. C. Galland, and J. L. Scott. 1989. Relationships between fertility, peak milk yields and lactational persisitency in dairy cows. Therio. 31(5): 1093-1103. Lefebvre, D. M., and E. Block. 1992. Effect of recombinant bovine somatotropin on estradiol-induced estrus behavior in ovariectomized heifers. J. Dairy Sci. 75:1461-1464. Leonard, M., M. Gallo, G. Gallo, and E. Block. 1990. Effects of a 28-day sustained- release formulation. of recombinant bovine somatotropin (rbST) administration to cows over two consecutive lactations. Can. J. Animal Sci. 70:795-809. Lucy, M. C., J. D. Savio, L. Badinga, R. L. De La Sota, and W. W. Thatcher. 1992. Factors that affect ovarian follicular dynamics in cattle. J Anim Sci. 70:3615- 3626. Lucy, M. C., T. L. Curran, R. J. Collier, and W. J. Cole. 1994. Extended function of the corpus luteum and earlier development of the second follicular wave in heifers treated with bovine somatotropin. Therio. 41:561-572. Mohammed, H. O., M. E. White, and N. LaFaunce. 1991. Multivariate analysis of factors associated with calving rate in dairy cows. Therio. 35(2):443-449. McClary, M. S., R. K. McGuffey, and H. B. Green. 1990. Bovine somatotropin: part 2. Agri-pract. 11(6):5-11. Nielen, M., Y. H. Schukken, and D. T. Scholl. 1989. Twinning in dairy cattle: a study of risk factors and effects. Therio. 32:845-862. Oldenbroek, J. K., and G. J. Garsen. 1993. Effects of treatment of dairy cows with recombinant bovine somatotropin over three or four lactations. J. Dairy Sci. 76:453-467. 71 01:5 Peel. II P Hit P811 P051- Rem Oresnik, A. 1995. Effect of health and reproductive disorders on milk yield and fertility in dairy cows. Bovine Pract. 29:43-45. Peel, C. J ., and D. E. Bauman. 1987. Somatotropin and lactation. J Dairy Sci. 70:474- 486. Pell, A. N., D. S. Tsang, B. A. Howlett, and M. T. Huyler. 1992. Effects ofa prolonged-release formulation of sometribove (n-methionyl bovine somatotropin) on jersey cows. J. Dairy Sci. 75:3416-3431. Posilac“ product label. Monsanto C0.. St. Louis MO. 1993. Post-approval Monitoring Program (PAMP). Protiva, a division of Monsanto C0.. St. Louis MO. 1997. Remond, B., M. Cisse, A. Ollier, and Y. Chilliard. 1991. Slow release somatotropin in dairy heifers and cows fed. two levels of energy concentrate. 1. performance and body condition. J .‘ Dairy Sci. 74:1370-1381. Rijpkema, Y. S., L. van Reeuwijk, and D. L. Hard. 1990. Responses of dairy cows to treatment with sometribove (r-BST) during three consecutive years. Livest. Prod. Sci. 26:193-216. Rothman, K. J. 1986. Interactions between causes. Modern Epidemiology: Boston: .Little, Brown and Company. SAS System for Windows“, Version 6.10. Cary NC, SAS Institute, 1994. Schemm, S. R., D. R. Deaver, D. R. Griel, and L. D. Muller. 1990. Effects of recombinant bovine somatotropin on luteinizing hormone and ovarian function in lactating dairy cows. Biol Reprod. 42:815-821. Schmidt, G. H. 1989. Effect of length of calving intervals on income over feed and variable costs. J. Dairy Sci. 72:1605-1611. Thomas, J. W., R. A. Erdman, D. M. Galton, R. C. Lamb, M. J. Arambel, J. D. Olson, K. S. Madsen, W. A. Samuels, C. J. Peel, and G. A. Green. 1991. Responses by lactating cows in commercial dairy herds to recombinant bovine somatotropin. J Dairy Sci. 74:945-964. Vila-Godoy, A., T. L. Hughes, R. S. Emery, L. T. Chapin, and R. L. Fogwell. 1988. Association between energy balance and luteal function in lactating dairy cows. J Dairy Sci. 71:1063-1072. 72 Waterman, D. F., W. J. Silvia, R. W. Hemken, G. Heersche, Jr., T. S. Swenson, and R. G. Eggert. 1993. Effect of bovine somatotropin on reproductive function in lactating dairy cattle. Therio. 40:1015-1028. Wells, S. J ., A. M. Trent, R. J. Collier, and W. J. Cole. 1995. Effect of long—term administration of a prolonged release fomulation of bovine somatotropin (Sometribove) on clinical lameness in dairy cows. Am. J. Vet. Res. 56(8):992- 996. Whitaker, D. A., E. J. Smith, J. M. Kelly, and L. S. Hodgson-Jones. 1988. Health, welfare and fertility implications of the use of bovine somatotropin in dairy cattle. Vet. Rec. 122:503-505. Zhao, X., J. H. Burton, and B. W. McBride. 1992. Lactation, health, and reproduction of dairy cows receiving daily injectable or sustained-released somatotropin. J. Dairy Sci. 75:3122-3130. 73 ' CHAPTER FIVE SURVIVAL “AN ALYSIS OF THE EFFECT OF BOVINE SOMATOTROPIN ON DEATH AND CULLING OF DAIRY COWS ABSTRACT Holstein cows (11 = 555) from four Michigan dairy farms were randomly assigned as untreated controls or to receive bovine somatotropin (500 mg) administered subcutaneously every 14 days beginning at 63 to 69 days of lactation and continuing until approximately 21 days prior to the end of lactation or until the animal was removed from the herd. Survival analysis was used to determine the effect of bST on removal from the herd (due to death or culling) for dairy cows. By bivariate analysis, none of the comparisons made of longevity between the study treatment groups were significantly different. Survival curves were analyzed by log-rank testing and did not differ between the treatment groups for any farm or between primiparous, multiparous or all cows. Significantly more control cows were culled for reason of low milk production compared to bST-treated cows during the study lactation. More control cows were culled for reason of injury compared to bST-treated cows during the first 63 days of the subsequent lactation. Cox proportional hazards models constructed found that abortions and injuries during the study lactation were associated with increased risk of removal from the herd and greater risk of uterine torsion and higher pretreatment milk production were associated with reduced risk of removal from the herd. In this model, the hazard ratio for bST treatment was 0.85 (95% CI: 0.62 to 1.16). During 74 the first 63 days of the subsequent lactation, a similar model found that cows experiencing pariparturient hypocalcernia, displaced abomasum, injury, peritonitis, uterine prolapse or respiratory disease had increased risk of removal. In this model, the hazard ratio for bST treatment was 0.92 (95% CI: 0.53 to 1.61). 75 INTRODUCTION For dairy cows, the part of lifetime when milk production occurs is often referred to as the period of longevity (Essl, 1998). Longevity of dairy cows has a profound effect on dairy farm profitability and affects many facets of the farm, such as herd replacement and feed inventory requirements. Therefore, any management change that affects longevity has the potential to impact many aspects of the dairy farm economy. Increased longevity is economically desirable because as the functional lifetime of cows increases, replacement costs are reduced and more cows reach the most profitable lactations (Essl, 1998; Jagannatha et al., 1998; Van Arendonk, 1991). In fact, the most profitable lactation of a dairy cow is the last lactation attained, in part because older cows produce more milk than do younger cows, and because cows spend most of their first lactation recovering costs associated with replacement rearing (Jagannatha et al., 1998) . Many diseases, suchas mastitis, reproductive failure and lameness, may influence rates of culling and therefore impact cow longevity and economics. In fact, studies have found that the major portion of disease economic impact (approximately 75% of the total cost) is mediated through the cost of deaths and culling (W eigler et al., 1990). Because bovine somatotropin may affect disease rates, it might impact culling rates and thereby decreased profitability. Survival analysis is a technique that allows for the objective comparison of 76 culling (removal) rates. Furthermore, the results of SA are given in time-specific probabilities, which are readily interpretable in both statistically and economically meaningful terms (Lee et al., 1989). Another major advantage of SA is it’s ability to use data from study subjects which never experienced the outcome of interest, referred to as “censored” observations. However, these subjects still contribute useful information to the analysis, and thereby minimize the loss of information (Eicker et al. , 1996). The Cox proportional hazards model has become the model of choice for SA (Cox and Cakes, 1984). This model is able to account for the effect of several confounding factors that may affect removal rates from the herd. The key assumption of this model is that the hazard (risk) of death or culling is balanced (proportional) between the study treatment groups. Other attractive features of this model is that it is not necessary to specify the distribution of .the hazard function and that the estimated coefficients produced by the model are not biased by the inclusion of censored data (Selvin et al., 1996). For our study, the use of Cox models allowed for the simultaneous evaluation of several risk factors suspected of being associated with herd removal rates. The objective was to determine the effect of bST (used according to label directions) on death and culling as determined by survival analysis techniques (removal from the herd). 77 SPECIFIC METHODS The outcome of interest in this study was removal from the herd. A cow could be removed from the herd by either being culled (sold for slaughter purposes) or dying. No cows were sold for dairy purposes during this study- Cows that were not removed from the herd during the study lactation were considered censored at the time when they calved again to started a new lactation. Cows which aborted during the study lactation were considered to have calved again only if they aborted after 250 days of gestation. Otherwise, aborting cows were considered to still be in production during the study lactation. A spreadsheet program (Excel“) (Microsoft Corporation, 1993) was used to manage the data and The SAS System for Windows“ (SAS Institute, 1994) was used for all statistical analysis. The study treatment groups were. compared by chi-square for bivariate analysis of culling and death. Life tables were constructed by dividing the study lactation into three-month intervals. The probability of survival was then calculated for each interval by dividing the number of cows that either were removed from the herd during the interval by the total number of cows at risk during the interval. Cumulative survival was calculated by dividing the number of cows remaining in the herd at the end of the interval by the total number of cows at the beginning of the study. Kaplan-Meier (product limit) survival function estimates were analyzed by use of the log-rank test and used to compare survival curves between the study treatment and parity groups. This test is based on a calculation of the odds of 78 survival for cows in each study treatment group and then comparing the expected to the observed result. The calculated test statistic is distributed approximately as a chi- square and was used to obtain the corresponding P-value. This analysis was also used to find the median survival time for each study and parity group and for the entire study population. Cox proportional hazards models were constructed for cows that either were culled or died during the study lactation or during the first 63 days of their subsequent lactation. The total time contributed to the analysis by each study treatment group was initially compared by Student’s t-test. The Cox model utilized was: h(t) = he“) eXP(l3X) where h(t) = the hazard (risk) of removal experienced by a particular cow at a given time; ho(t) = the average (baseline) hazard of removal; and exp(Bx) = the vector of the coefficient for independent variables utilized in the model. Independent variables were included in the model if they were significant at P < 0.10. Dummy variables were created for the four study farms and were forced into all models to adjust for any confounding attributable to unmeasured factors present in these herds. Also, the variable for study treatment group, being the principle variable of interest, was also forced into the models. Lactation (parity) and diseases specific to the time frame under analysis (either the study lactation or the first 63 days of the subsequent lactation) were included in the models if they were significant at the inclusion criterion. Independent variables were selected for inclusion, first by a step-up variable selection method and later confirmed by a step-down procedure. The independent variables used were 79 considered to be proportional between the study treatment groups for all cows and herds. All biologically significant, two-way interactions were accessed and considered for inclusion in the model if they met the inclusion criterion. A hazard ratio was calculated for each independent dichotomus variable that was included in the model. Both martingale and deviance residuals were plotted to identify possible outlying observations. Because cows were randomly assigned to study treatment groups, milk production in the pretreatment period (the first 63 days of the study lactation) should theoretically have been approximately balanced between the study groups. However, production was different between the study groups on one farm (farm three) with control cows producing 192.2 kg more milk in the pretreatment period compared to cows that would later be treated with bST (P = 0.03). No significant difference in pretreatment milk production was found between the study groups in the remaining three herds. For the four herds analyzed together, pretreatment milk production was similar between the study groups with a 47.1 kg difference observed between the treatment groups (P = 0.34). In a previous study, reproductive parameters were found to be approximately balanced between the study treatment groups (Judge et a1. , 1999). Because reproductive performance has a large influence on culling, if bST had been associated with any reproductive parameter, it could have been viewed as an intervening variable between bST use and culling. However, because these data demonstrated that the study treatment groups were approximately balanced with regard to reproductive parameters, 80 culling due to poor reproductive function was determined to be independent of any bST treatment effect. The lactational incidence of most diseases were similar between the study treatment groups. The only exception was that bST-treated cows experienced more non-specific off feed conditions compared to control cows (P = 0.05) (Judge et al. , 1999). Because this and other diseases were not significantly associated with the rate of removal of cows, any change in risk of removal as a result of diseases in this study was determined to be independent of any affect bST may have had on disease rates. RESULTS Bivariate analysis of removal from the herd Cows that were removed from the herd (culled or died) during the study lactation were compared between the treatment and parity groups (Table 5.1). 81 TABLE 5.1: Analysis of culling, deaths and removal from the herd (culling plus deaths = total) for primiparous, multiparous and all study cows from calving until the end of lactation for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co. , St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls (number of cows removed (percent of treatment group removed)). Treatment bST Control Total group Parity (n) % (n) % (n) % P-value Odds ratio (95% CI) Primiparous Culled (22) 0.21 (12) 0.14 (34) 0.18 0.23 1.60 (0.70, 3.79) Died (4) 0.04 (2) 0.02 (6) 0.03 0.57 1.63 (0.29, 9.12) Total (26) 0.24 (14) 0.16 (40) 0.21 0.17 1.65 (0.80, 3.40) Multiparous Culled (53) 0.30 (67) 0.36 (120) 0.33 0.21 0.75 (0.47, 1.20) Died (7) 0.04 (7) 0.04 (14) 0.04 0.93 1.05 (0.36, 3.05) Total (60) 0.34 (74) 0.40 (134) 0.37 0.23 0.77 (0.49, 1.21) All cows Culled (75) 0.26 (79) 0.29 (154) 0.28 0.95 0.99 (0.64, 1.53) Died (11) 0.04 (9) 0.03 (20) 0.04 0.73 1.17 (0.48, 2.88) Total (86) 0.30 (88) 0.33 (174) 0.31 0.85 1.04 (0.68, 1.61) None of the comparisons made were statistically significant but a trend toward higher rates of removal was apparent for bST-treated primiparous cows compared to primiparous controls. During the first 63 days of the subsequent lactation, the opposite result was observed with a trend toward fewer deaths and removals for cows that were primiparous and treated with bST in the previous lactation (Table 5.2). A similar trend was found for all study cows during the first 63 days of the subsequent lactation. 82 TABLE 5.2: Analysis of culling, deaths and removal from the herd (culling plus deaths = total) for primiparous, multiparous and all study cows during the first 63 days of the subsequent lactation for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls (number of cows removed (percent of group removed)). Treatment , bST Contr01 Total group Parity (n) % (n) % (n) % P-value Odds ratio (95 % CI) Primiparous Culled (4) 0.05 (6) 0.08 (10) 0.07 0.40 0.57 (0.11, 2.54) Died (2) 0.03 (5) 0.07 ~ (7) 0.05 0.19 0.34 (0.06, 1.81) Total (6) 0.07 (11) 0.15 (17) 0.11 0.12 0.44 (0.16, 1.27) Multiparous Culled (9) 0.08 (11) 0.10 (20) 0.09 0.57 0.77 (0.27, 2.13) Died (10) 0.09 (13) 0.12 (23) 0.10 0.44 0.71 (0.30, 1.70) Total (19) 0.16 (24) 0.21 (43) 0.19 0.32 0.71 (0.37, 1.39) All cows Culled (13) 0.07 (17) 0.09 (30) 0.08 0.33 0.69 (0.30, 1.56) Died (12) 0.06 (18) 0.10 (30) 0.08 0.18 0.60 (0.28, 1.27) Total (25) 0.13 (35) 0.19 (60) 0.16 0.09 0.62 (0.35, 1.08) Survival measured by life table analysis Cows removed from the herds were analyzed by life table analysis for primiparous, multiparous and for all study cows (Tables 5.3, 5.4 and 5.5). Statistical analysis of the three survivorship curves by log-rank testing found median survival time was not significantly different between the any study treatment groups. Median 83 survival times of 588 days (95 % CI: 520 to 658 days) and 595 days (95% CI: 526 to 801 days) were calculated for bST-treated and control primiparous cows, respectively (P = 0.28). TABLE 5.3: Life table analysis of removal from the herd for primiparous cows from calving until the end of lactation for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St- Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls. No. of Events Surviving proportion Censored Removals No. at risk Interval (%) Overall (%) Interval (mos) .- bST Control bST. Control bST Control bST Control bST Control 0-3 0 0 1 2 107 86 99.1 97.7 - - 3-6 0 0 0 0 106 84 100 100 99.1 97 .7 6-9 0 0 5 1 106 84 95.3 98.8 94.3 96.5 9-12 27 27 2 4 101 83 98.0 95.2 92.5 91.9 12-15 33 27 4 2 72 52 94.5 96.2 88.8 89.5 15-18 17 14 9 2 35 23 74.3 91.3 80.4 87.2 18-21 1 3 4 2 9 7 55.6 71.4 76.6 84.9 21-24 2 1 l 0 4 2 75.0 100 75.7 84.9 24-27 1 0 0 l 1 l 100 0 75.7 83.7 Total 81 72 26 14 107 86 - - 75.7 83.7 For multiparous cows, median survival times of 581 days (95% CI 470 to 819 days) and 508 (95% C1450 to 685 days) were found for bST-treated and control cows, respectively (P = 0.34) 84 TABLE 5.4: Life table analysis of removal from the herd for multiparous cows from calving until the end of lactation for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls. No. of Events Surviving proportion Censored Removed No. at risk Interval ( %) Overall (%) Interval (mos) bST Control bST Control bST Control bST Control bST ' Control 0-3 0 0 3 5 177 185 98.3 97.3 - - 3-6 0 0 10 11 174 180 94.3 93.9 92.7 91.4 6-9 0 0 13 16 164 169 92.1 90.5 85.3 82.7 9-12 42 25 14 20 151 153 90.7 86.9 77.4 71.9 12-15 44 58 12 12 95 108 87.4“ 88.9 70.6 65.4 15-18 21 18 4 6 39 38 89.7 84.2 68.4 62.2 18-21 5 6 3 2 14 14 78.6 85.7 66.7 61.1 21-24 4 3 0 1 6 6 100 83.3 66.7 60.5 24-27 1 l 1 1 2 2 50.0 50.0 66.1 60.0 Total 117 111 60 74 177 185 - - 66.1 60.0 Median survival times were 581 days (95% CI: 501 to 615 days) and 552 days (95 % CI: 508 to 685 days) for all study treatment and control cows analyzed together, respectively (P = 0.51). 85 TABLE 5.5: Life table analysis of removal from the herd for all study cows from calving until the end of lactation for all study cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls. No. of Events Surviving proportion Censored Removed No. at risk Interval (%) Overall (%) Interval (mos) bST Control bST Control bST Control bST Control bST Control 0-3 0 0 ’ 4 7 284 271 98.6 97.4 - - 3-6 0 0 10 11 280 264 96.4 95.8 95.1 93.4 6—9 0 0 18 17 270 253 93.3 “93.3 88.7 87.1 9-12 69 52 16 24 252 236 93.7 89.8 83.1 78.2 12-15 77 85 16 14 167 160 90.4 91.3 77.5 73.1 15-18 38 32 13 8 74 61 82.4 86.9 72.9 70.1 18-21 6 9 7 4 23 21 69.6 81.0 70.4 68.6 21-24 6 4 1 l 10 8 90.0 87.5 70.1 68.3 24-27 2 1 1 2 3 3 66.7 33.3 69.7 67.5 Total 198 183 86 88 284 271 - - 69.7 67.5 A trend toward reduced longevity was observed for bST-treated primiparous cows (Figure 5.1) while the opposite effect was found for bST-treated multiparous cows (Figure 5.2). For all study cows, both study treatment groups had similar survivorship curves (Figure 5.3). However, none of the differences in survival times were significant for any comparisons between study treatment groups. For the subsequent lactation, life tables were not constructed because of the short follow-up time (only the first 63 days of lactation). 86 FIGURE 5.1: Survivorship curve for primiparous cows from calving until the end of lactation for cows supplemented with 500 mg of bST (—) (Posilac“, Monsanto Co. , St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the . end of lactation and untreated controls H). Vertical lines on x-axis denote median survival times. 1.0- 0.9- 0-8‘ 0.7- 0.5- 0.5‘ 0.4‘ 0.3“ l 0.2“ 0.1‘ Survival Function Estimate 0.0' ITIIITIIIJIII'IIIIIIIHIIIll lllllllllllllllll 0 100 200 300 400 500 00 700 800 900 Survival Time (days) 87 FIGURE 5.2: Survivorship curve for multiparous cows from calving until the end of lactation for cows supplemented with 500 mg of bST (—) (Posilac“, Monsanto Co. , St. Louis, M0) at 14—day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls (——). Vertical lines on x-axis denote median survival times. 1.01 Survival. Function Estimate 0.1‘ FIIIIHIIIIWTIIITIIIIIIIllIll[Tlllllll111f11l 0 100 200 300 400 500 00 700 800 900 Survival time (days) 88 FIGURE 5.3: Survivorship curve for all study cows from calving until the end of lactation for cows supplemented with 500 mg-ofl bST( ) (PosilacR, Monsanto Co. , St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls ( ). Vertical lines on x-axis denote median survival times. ' m 1.0... 4.. . g 0.93 f, 0.8“ m r t: 0.7- 0 . a: 0.1;1 o . g or: U. _ or (g 1 'EE 0.3‘ 3 0.2- (I) . o.i- IIITIJIIIIWIIIIWIITIII1117 IIITIIIIJIIJIII1 0 100 200 300 400 500 600 700 800 900 Survival Time (days) 89 Reasons for removal from the herd Reasons given by herd managers for removal of cows from the herd were compared between the treatment and parity groups. The only significant result during the study lactation was that fewer bST-treated cows were reported culled because of low milk production (P = 0.003) (Table 5.6). TABLE 5.6: Reasons for removal from the herd for all study cows from calving until the end of lactation for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co. , St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls (no. removed (percent of total removed)). Treatment group bST Controls Total Reason for removal (11) % (n) % (n) % P-value Reproduction (27) 31.4 . (23) 26.1 (50) 28.7 0.68 Lameness (11) 12.8 (18) 20.5 (29) 16.7 0.14 Mastitis (12) 14.0 (10) 11.4 (22) 12.6 0.75 Low milk production (4) 4.7 (17) 19.3 (21) 12.1 0.003 Physical injury (8) 9.3 (4) 4.6 (12) 6.9 0.28 Abortion (6) 7.0 (5) 5.7 (11) 6.6 0.82 Unspecified reason (7) 8.1 (4) 4.6 (11) 6.3 0.40 Displaced abomasum (2) 2.3 (3) 3.4 (5) 2.9 0.62 Old age (2) 2.3 . (0) 0.0 (2) 1.2 0.17 Udder condition (1) 1.2 (1) 1.1 (2) 1.2 0.98 Respiratory disease (1) 1.2 (1) 1.1 (2) 1.2 0.97 Cancer (0) 0.0 (2) 2.3 (2) 1 .2 0. 15 Hardware disease (1) 1.2 (0) 0.0 (1) 0.7 0.33 Infectious disease (1) 1.2 (0) 0.0 (1) 0.7 0.33 Organ failure (1) 1.2 (0) 0.0 (1) 0.7 0.33 Peritonitis (1) 1.2 (0) 0.0 (1) 0.7 0.33 Metabolic disease ( 1) 1.2 (0) 0.0 (l) 0.7 0.33 Total _ (86) 49.4 88 50.6 (174) 100.0 0.59 90 This effect was observed for all parity and study group comparisons. During the first 63 days of the subsequent lactation, significantly fewer bST-treated cows were removed from the herd for reason of physical injury compared to control cows during this time frame (P = 0.003) (Table 5.7). TABLE 5.7: Reasons for removal from the herd for all study cows during the first 63 days of the subsequent lactation for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, M0) at l4-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls (number of cows removed (percent of total cows removed)). Treatment group bST Controls Total Reason for culling (n) % (n) % (n) % P-value Unspecified reason (6) 24.0 (10) 28.6 (16) 26.7 0.24 Displaced abomasum (4) 16.0 (5) 14.3 (9) 15.0 0.65 lameness (7) 28.0 (2) 5.7' (9) 15.0 0.12 Physical injury _ (0) 0.0 (8) 22.9 (8) 13.3 0.003 Mastitis (2) 8.0 ' (3) 8.6 (5) 8.3 0.60 Metabolic disease (3) 12.0 (2) 5.7. (5) 8.3 0.71 Dystocia (0) 0.0 (3) 8.6 (3) 5.0 0.07 Respiratory disease (0) 0.0 (2) 5.7 (2) 3.3 0.14 Lowmilkproduction (1) 4.0 . (0) 0.0 (1) 1.7 0.34 Peritonitis (1) 4.0 (0) 0.0 (1) 1.7 0.34 m (1) 4-0 (0) 0n (1) 1 7 0 34... Total (2_5) 41.7 35 (58.3) (60) 15.8 0.08 In our study, bST-treated primiparous cows displayed a trend toward a greater culling rate for reproductive failure than did similarly aged controls (P = 0.11). This occurred because these bST-treated cows tended to have a lower pregnancy rate than did comparable controls (P = 0.13). As a result, a trend toward a higher removal rate was found for bST-treated primiparous cows as compared to primiparous controls 91 during the study lactation (P = 0.17). All other reasons for removal were similar between the treatment groups for parity comparisons and all study cows analyzed together as well. Survival analyzed by proportional hazards multivariate modeling The 555 cows included in this analysis contributed a total of 214,725 cow-day's with bST-treated cows contributing 110,991 cow—days and control cows contributing 103,734 cow—days. Average lactation (calving until the subsequent calving or removal) length was similar between the study treatment groups with bST-treated primiparous cows averaging 413.5 days and controls averaging 402.3 days (P = 0.45). Multiparous cows that were treated with bST “had an average of 377.2 days of lactation length with controls averaging 373.7 days (P = 0.80) and all study cows together averaging 390.8 days of lactation length for bST-treated cows and 382.8 days for control cows (P = 0.43). However, primiparous cows lactation length averaged 33.1 days longer than did multiparous cows (P = 0.001). Over all parities, study cows averaged 386.9 days of lactation length. Separate models were constructed for primiparous, multiparous and for all study cows that were removed from the herds during the treatment lactation and for the first 63 days of the subsequent lactation. During the study lactation, for primiparous cows the conditions of abortion (P = 0.0003) and injury (P < 0.0001) were significantly associated with increased risk of removal while becoming pregnant decreased risk of removal (P < 0.0001). For this model, a hazard ratio of 0.89 (95% CI: 0.43, 1.86) 92 for bST treatment effect was calculated but was not significant in the model (P = 0.76). When removal from the herd was modeled for multiparous cows, cows aborting were at significantly greater risk (P = 0.005) of being culled while those becoming pregnant had reduced risk for removal (P < 0.0001). For this model, a hazard‘ratio of 0.88 (95 % CI: 0.62, 1.25) for bST treatment effect was calculated for bST treatment effect but was not significant in the model (P = 0.47). When all study cows were included, lactation (parity) was observed to have a significant effect (P < 0.001) on herd removal with survival decreasing as the number of lactations increased (Table 7.8). TABLE 5.8: Cox proportional hazards modeling of removal from the herd for all study cows from calving until the end of lactation for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, M0) at l4-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls. Independent Parameter Standard Hazard variable estimate error P-value ratio 95 % CI Farm 1 I 0.3221 0.2876 0.2627 1.38 0.79, 2.43 Farm 2 -0.2211 0.2373 0.3516 0.80 0.50, 1.28 Farm 3 0.4809 0.2293 0.0360 1.62 1.03, 2.54 Lactation 1.1822 0.2737 - <0.0001 3.26 1.91, 5.58 Lactation2 -0.1422' 0.0401 0.0004 0.87 0.80, 0.94 Pregnancy status -2.4611 0.1819 <0.0001 0.09 0.06, 0.12 Abortion 1.1411 0.2471 <0.0001 3.13 1.93, 5.08 Physical injury 1.5182 0.3745 <0.0001 4.56 2.19, 9.51 Uterine torsion -2.2952 0.7766 0.0031 0.10 0.02, 0.46 Pretreatment -0.0002 <0.0001 0.0299 1.00 1.00, 1.00 milk production Treatment group -0.1627 0.1594 0.3073 0.85 0.62, 1.16 Lactation2 = the quadratic term of the independent variable for lactation. 93 However, a curvilinear affect of this age-dependent variable was evident because the quadratic term of this variable was significant in the model as well (P = 0.0004). Therefore, the youngest and oldest cows were at the greatest risk of removal compared to cows in the middle of the age distribution. In this model,- abortions and physical injuries increased risk of removal while uterine torsions and greater pretreatment milk production decreased risk of removal from the herd. The interactions of parity and abortion (P = 0.02) as well as physical injury and pretreatment milk production (P = 0.03) were significant in the model for all study cows. However, the addition of these terms to the model was not determined to change the parameter estimate for bST treatment effect significantly. Therefore, these variables were not included in the final model. For the study lactation, a hazard ratio of 0.85 (95% CI: 0.62, 1.16) for bST treatment effect, but this variable was not significant in the model (P = 0.31). During the first 63 days of subsequent lactation, previously primiparous cows that calved again tended to be removed at a lower rate if they had been treated with bST during the study (previous) lactation. In this model, diseases associated with increased risk of removal were displaced abomasum (P < 0.0001), metritis (P = 0.04), injuries (P = 0.0001) and dystocia (P = 0.009). The hazard ratio of bST treatment effect was 0.54 (95% CI: 0.16, 1.88), but this estimate was not statistically significant (P = 0.34). Cows that had been multiparous during the previous (study) lactation and experienced periparturient hypocalcemia (P = 0.004), uterine prolapse (P < 0.0001), displaced abomasum (P = 0.0003), injuries (P = 0.0005) or peritonitis (P = 0.02) had 94 increased risk of removal. Cows that had been previously treated with bST had a small increase in risk of removal (hazard ratio = 1.01 (95% CI: 0.51, 2.01)) but this estimate was not statistically significant (P = 0.97) . In the model of all study cows, the interaction of lactation and injury was significant (P = 0.03). However, this interaction term did not change the parameter estimate for bST treatment effect significantly and therefore was not included in the final model. When all study cows were included, all diseases significant in the model increased the risk of removal, while cows previously treated with bST had a slightly reduced risk of removal (hazard ratio = 0.92 (95% CI: 0.53, 1.61) but this estimate was not significant in the model (P = 0.78) (Table 5.9) .- TABLE 5.9: Cox proportional hazard modeling of removal from the herd for all study cows during the first 63 days of the subsequent lactation for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation and untreated controls. Independent variable Parameter Standard Hazard estimate error P-value ratio 95 % CI Farm 1 0.3883 0.5171 0.4527 1.48 0.54, 4.06 Farm 2 0.3200 0.3932 0.4158 1.38 0.64, 2.98 Farm 3 0.1558 0.4668 0.7386 1.17 0.47, 2.92 lactation 0.2256 0.1074 0.0356 1.25 1.02, 1.55 Milk fever 1.3030 0.4225 0.0020 3 .68 1 .61 , 8.42 Injury 2.4751 0.4874 < 0.0001 11.88 4.57, 30.89 Peritonitis 1.9455 0.7701 0.0115 7.00 1.55, 31.65 Uterine prolapse 5.9240 1.1282 < 0.0001 375 .44 40.97, 3412.51 Respiratory disease 1.3432 0.7799 0.0850 3.83 0.83, 17.67 Displaced abomasum 1.7654 0.3566 < 0.0001 5.84 2.91 , 11.76 Treatment group 007% 0.2847 0.7797 0.92 0.53, 1.61 95 For all models, residual analysis found no outlying observations that required removal from the data. However, a pattern was observed in plotted residuals for both models of all study cows presented above. Because no objective means of assessing fit for these models currently exists, visual judgement of the residual plots is relied upon for this determination. However, it is possible for systematic patterns to be present even when the best fitting model has been constructed (Collet, 1994). DISCUSSION Twenty-two cows that were originally selected for the trial were in poor health at the time when treatment was to begin and therefore, were excluded from the analysis. This exclusion was in accordance with the product label that states cows must be determined healthy if bST treatment is to begin during the ninth week of lactation (Monsanto Company, 1993). Because less than 5% of the cows selected for the study were excluded, and because bST-treated and control cows were excluded based on the same criteria, it is unlikely that a sizable bias was introduced from this omission. Milk production in the pretreatment period was well balanced between the treatment groups except for one farm where control cows produced significantly more milk than those treated with bST. However, for the entire data set, milk production in the treatment groups during the pretreatment period was similar during the first 63 days of lactation. In this study, pre-existing intra-mammary infections (IMI) were balanced between bST-treated and control cows (Judge et al. , 1997) and therefore should not have affected any milk production differences observed. This issue is important 96 because cows with IMI likely produce less milk than cows without infection. During the study lactation, the number of cows removed from the study was similar to that observed in other studies (Grohn et al., 1998; Karuppanan et al. , 1997; Ruegg et al., 1998). However, because cows in the present study could not have been removed before the ninth week of lactation, when labeled bST treatment commenced, the rates of death and culling we calculated underestimate the true lactation rates. In SA, the median is used as the measure of central tendency rather than the mean because the distribution of time to removal from the herd is usually skewed. Therefore, mean survival time is biased and tends to overestimate actual survival time of cows. Median survival time was not significantly different for any of the comparisons made. Analysis of survivorship curves revealed opposite culling trends for cows of differing ages. While primiparous cows treated with bST tended to be culled somewhat more frequently than control cows, multiparous bST-treated cows were retained in the herd slightly longer than similarly aged controls. This result is contrary to that which was reported by Canadian reviewers who theorized that older cows may experience greater rates of diseases leading to. increased culling of these cows (Canadian Veterinary Medical Association Expert Panel, 1998). While multiparous cows in our study experienced greater rates of most diseases compared to younger cows, none of these diseases were associated with increased rates of culling for bST- treated cows. Reasons given for culling because of reproductive problems and lameness were Sinlilar between our study and data collected by the NAHMS (Centers for 97 Epidemiology and Animal Health, 1996). However, I observed less culling for reasons of mastitis and low milk production than was reported by NAHMS. Control cows were culled for reason of low milk production at a similar rate to that reported by NAHMS, but bST-treated cows had less culling for this reason. More control cows of all ages were culled during the study lactation because of low milk production than were bST- treated cows. This resulted in a significant protective effect of bST treatment on culling for reason of low milk production and was observed for primiparous, multiparous and all study cows analyzed together. Because bST-treated cows produce more milk than untreated cows, it was not unexpected that managers kept treated cows in the herd longer. Other studies have observed that cows with higher milk production have reduced risk of culling relative to their herdmates (V ollema and Groen, 1998). In a bST dose titrating study, culling was reduced at all levels of bST treatment compared to control cows (McDaniel et al. , 1990) likely because of greater production in bST- treated cows. Across all study cows, there were no differences in culling between the study treaunent groups for any reason except for reason of low milk production during the study lactation. During the first 63 days of the subsequent lactation following bST treatment, a trend was observed for formerly bST-treated primiparous cows to be culled at a lower rate than were comparable controls. This was the opposite trend observed for this cohort of cows during the study lactation. This may have occurred because culling was less frequent for certain diseases in these bST-treated cows. Analysis of removal from the herd for primiparous, multiparous and all study 98 , cows by Cox proportional hazards modeling of the study lactation revealed two significant predictors common to all three models: Cows which became pregnant were 91% less likely to be culled compared to those not conceiving while cows that aborted were at least three times as likely to be removed from the herd compared to those not aborting. These models demonstrate the large impact that reproductive performance has on the longevity of cows. This is in agreement with previous studies examining factors that affect culling rates for cows (Grohn et al. , 1998). In this study, bST did not affect any reproductive parameters significantly. If bST had influenced reproductive function in this study cows, it would have been difficult to distinguish between the direct effect bST may have had on culling and any indirect effects bST might have had on culling mediated through reduced reproductive performance (Esteban et al., 1994). I chose not to include milk production during the study lactation as an independent variable in any of our models. This is because bST increased milk production in the study cows and milk yield has the potential to be an indirect measure of bST treatment effect (intervening variable). Because of this potential for mixing of effects, it is not appropriate to include such variables as independent variables in multivariate models (Grohn et al. , 1997; Kleinbaum et al. , 1982). Because diseases have a strong influence on culling (Grohn et al., 1998), these variables were chosen for inclusion in the models. It is desirable to include such variables to account for potential confounding by herd effects. Also, because these variables are likely confounders of other risk factors of interest (e.g. , twin calvings), they warrant inclusion in these 99 models (Ross et al., 1998). Furthermore, because diseases often result in reduced milk production, these variables actually account for both the direct effect that the disease may have had on culling and any indirect effect that changes in milk production because of disease conditions might have had on culling. In another study, the level of milk production during the previous 305-day lactation was not associated with any disease except mastitis (Grohn et al., 1998). Therefore, bST would not be expected to influence most disease rates and subsequent culling indirectly via its effect on milk production in the previous lactation. All variables for disease occurrence were considered to be time-independent in the study models. It has been stated that this assumption is valid when diseases occur early in lactation (Grohn et al., 1997). In the present study models, diseases included as potential confounders primarily occurred early in lactation and therefore were treated as time-independent covariates. The Monsanto Company was required to conduct a post-approval study of bST as a condition of approval for commercial sale. In this study, multiparous cows were found to have greater risk of culling (relative risk = 1.38) but no significant effect was found for primiparous cows (Monsanto Company, 1997). This may have resulted from reduced reproductive performance in these cows. Canadian researchers conducted a meta-analysis of bST trial data during their review process for considering granting approval for its use in Canada. In their analysis, bST was found to increase culling rates by as much as 25 % (Canadian Veterinary Medical Association Expert Panel, 1998). However, this effect was found to be marginally statistically significant (P = 0.06). Of the six studies included in this Canadian analysis, only one had a 95% 100 confidence interval that did not include one (the point of no effect). In this analysis, the definition of culling was not consistent across all studies with some studies including cows culled because of poor reproductive performance while others did not. In the present study, all cows that were removed from the herd (either were culled or died) for any reason were included and the final disposition (either culling, death or subsequent calving) was known for all study cows. CONCLUSIONS Functional longevity of dairy cows has a major impact on the profitability of the dairy farm. Therefore, any factor that alters the length of the productive lifetime of cows could influence animal welfare and economics of the farm. Much controversy has existed regarding the increased risk of removal from the herd that may be associated with bST use. While some studies have supported this conclusion, others found no effect of bST on survival. In this study, bST had no significant effect on either culling or death rates. However, a significant association was observed between bST use and reduced culling for the cited reason of low production. I concluded that overall bST did not significantly affect the rate of removal of cows from the study herds. 101 REFERENCES Canadian Veterinary Medical Association Expert Panel. 1998. Report of the Canadian Veterinary Medical Association Expert Panel on rBST.: Ottawa, CA. Centers for Epidemiology and Animal Health. 1996. National Animal Health Monitoring System, Dairy ‘96. Fort Collins, CO. Collet, D., 1994. Modelling Survival Data in Medical Research. Chapman and Hall, New York. Cox, D. R., and D. Oakes. 1984. Analysis of survival data. Chapman and Hall, London. Dairy Comp 305“, Version 3.1. Valley Agricultural Software. Tulare CA. 1990. Eicker, S. W., Y. T. Grohn, and J. A. Hertl. 1996. The association between cumulative milk yield, days open, and days to first breeding in New York Holstein cows. J. Dairy Sci. 79, 235-241. Essl, A., 1998. Longevity in dairy cattle breeding: a review. Livest. Prod. Sci. 57, 79- 89. Esteban, E., P. H. Kass, L. D. Weaver, J. D. Rowe, C. A. Holrnberg, C. E. Franti, and H. F. Trout. 1994. Interval from calving to conception in high producing dairy cows treated with recombinant bovine somatotropin. J. Dairy Sci. 77, 2549-2561. Excel“, Version 5.0. 1993. Microsoft Corporation, Redmond, WA. Grohn, Y. T., V. Ducrocq,and J. A. Hertl. 1997. Modeling the effect of a disease on culling: an illustration of the use of time-dependent covariates for survival analysis. J. Dairy Sci. 80, 1755-1766. Grohn, Y. T., S. W. Eicker,V. Ducrocq, and J. A. Hertl. 1998. Effect of diseases on the culling of Holstein dairy cows in New York state. J. Dairy Sci. 81 , 966-978. Jagannatha, S., J. F. Keown, and L. D. VanVleck,L. 1998. Estimation of relative economic value for herd life of dairy cattle from profile equations. J. Dairy Sci. 81, 1702-1708. 102 Judge, L. J ., P. C. Bartlet, J. W. Lloyd, and R. J. Erskine. The impact of recombinant somatotropin on net farm income: analysis of four Michigan dairy farms. Submitted for publication in J. Dairy Sci. Judge, L. J ., P. C. Bartlett, J. W. Lloyd, and R. J. Erskine. 1999. Recombinant bovine somatotropin: association with reproductive performance in dairy cows. Theriogenology. 52, 481-496. Judge, L. J ., R. J. Erskine, and P. C. Bartlett. 1997. Recombinant bovine somatotropin and clinical mastitis: incidence, discarded milk following therapy, and culling. J. Dairy Sci. 80, 3212-3218. Karuppanan, P., M. C. Thurmond, and I. A. Gardner. 1997. Survivorship approaches to measuring and comparing cull rates for dairies. Prev. Vet. Med. 30, 171-179. Kleinbaum, D. G., L. L. Kupper, and H. Morgenstern. 1982. Epidemiologic research: principles and quantitative methods. Lifetime Learning Publications, London. Lee, L. A., J. D. Ferguson, and D. T. Galligan. 1989. Effect of disease on days open assessed by survival analysis. J. Dairy Sci. 72, 1020-1026. McDaniel, B. T., W. E. Bell, J. Fetrow, B. D. Harrington, and J. D. Rehman. 1990. Survival rates and reasons for removal of cows injected with rBST. J. Dairy Sci. 73(suppl 1):159. Parmar, K. B., and D. Machin. 1995. Survival analysis: a practical approach. John Wiley and sons, New York. Posilac“ product label, 1993. Monsanto Co., St. Louis, MO. Post-approval Monitoring Program (PAMP). Protiva, a division of Monsanto C0.. St. Louis MO. 1997. Ross, W. A., J. B. Kaneene, and J. C. Gardiner. 1998. Survival analysis of risk factors associated with the occurrence of lameness in a Michigan horse population. Am J Vet Res. 59(1), 23-29. Ruegg, P. L., A. Fabellar, and R. L. Hintz. 1998. Effect of use of bovine somatotropin on culling practices in thirty-two dairy herds in Indiana, Michigan, and Ohio. J. Dairy Sci. 81:1262-1266. SAS System for Windows“, Version 6.10, 1994. SAS Institute, Cary, North Carolina. 103 Selvin, S., 1996. Statistical analysis of epidemiologic data. Oxford University Press, New York. Van Arendonk, J. A. M., 1991. Use of profit equations to determine relative economic value of dairy cattle herd life and production from field data. J Dairy Sci 74, 1 101-1 107. Vollema, A. R., and A. F. Groen. 1998. A comparison of breeding value predictors and longevity using a linear model and survival analysis. J Dairy Sci 81, 3315- 3320. Weigler, B. J., D. W. Hird,W. M. Sischo, J. C. Holmes,C. Danaye-Ehni, C. W. Palmer, and W. W. Utterback. 1990. Veterinary and nonveterinary costs of disease in 29 California dairies participating in the national animal health monitoring system from 1988 to 1989. JAVMA 196(12), 1945-1949. 104 CHAPTER SIX THE EFFECT OF RECOMBINANT BOVINE SOMATOTROPIN ON MILK PRODUCTION AND LACTATION MILK ESTIMATES FOR COWS MILKED THREE TIMES PER DAY ABSTRACT Holstein cows (11 = 555) from four Michigan dairy farms were randomly assigned as untreated controls or to receive bovine somatotropin (500 mg) administered subcutaneously every 14 days beginning at 63 to 69 days of lactation and continuing until approximately 21 days prior to the end of lactation or until the animal was removed from the herd. Determination of the patterns of bST-assisted milk production and the effect of bST on lactational milk estimates were made. Correlation was high (r > 0.97) between actual milk produced (measured by daily, individual-cow milk weights) and Dairy Herd Improvement Association production estimates at 305 days of lactation for all study cows. Standard deviation (variance) in daily and 305-day production did not differ significantly between the treatment groups. While positive production response to bST was apparent at the first administration, maximum response was not apparent until approximately the middle of the 238 day treatment period. Both amount of, and days to, peak milk production were significantly greater for bST-treated cows compared to controls and bST-treated cows were significantly more persistent in lactation than controls. Dairy Herd Improvement Association measurements underestimated actual 305-day production more for bST treated-cows than it did for 105 control cows at the first and second monthly tests after bST initiation. However, this difference was not significant by the third monthly test date. By 305 days of lactation (final lactational production estimate), the estimates were equally accurate for both bST-treated and control cows. 106 INTRODUCTION Although bovine somatotropin (Posilac“; Monsanto Company, St. Louis, MO) has been used commercially in the US since 1994, the temporal pattern of observed milk production response has not been characterized in commercial dairy herds when bST is used according to label directions for a complete lactation (Pelissier et al. , 1978). Any affect that bST may have on daily variability of milk production, the amount of, and time to, peak production, as well as lactational persistency has not been widely reported since the approval of bST for commercial use. Also, little is known regarding any affect bST may have on DHIA lactational production estimates that are based on sampling milk production approximately every 30 days. By altering the pattern of milk production (Bauman, 1992) bST may adversely affect the accuracy of DHIA milk production estimates. Currently in the US, DHIA lactational yield estimates made from test day data are most frequently performed using the test interval method (Norman et al., 1999; US Department of Agriculture, Science and Education Administration, 1980). This procedure involves dividing the interval between successive test days into two parts: milk produced during the first half of the interval is based on the first test day data and production during the second half of the interval “is based on the next monthly test day yield (Norman et al. , 1999). The equations and factors used adjust for season of calving, region of the US that the herd is located in, the production level of the herd and production on the last sampling day of the record (US Department of Agriculture, 107 Science and Education Administration, 1980). It is important that these production estimates be accurate because herd management (e. g. , culling) decisions are frequently based on this information (Jones, 1997; Norman et al., 1999). Also, the prediction of production for lactations currently in progress is used in evaluating new sires and allows the records of as many daughters as possible to be included in this process (Keown and Van Vleck, 1972). However, it has been stated that monophasic models (such as the test interval method) may not accurately predict lactational milk yield in bST-treated cows without some modifications (W eigel et al. , 1992). Such modifications may be needed because bST can cause a second peak in production and has been associated with increased lactational persistency (slower rate of production decline) when administration begins at 63 days of lactation (labeled use) (Bauman, 1992). In doing so, bST changes the shape of the lactation curve (Weigel et al., 1992) and therefore necessitates the examination of methods used by DHIA to project milk production for accuracy in bST-treated cows (Jones, 1997). Milk production in bST-treated cows has been observed to follow a cyclical pattern during each 14-day injection period, with several injections necessary before peak production response is attained (Bauman et al. , 1989). Whether the day that milk production is measured during this cycle has a significant influence on the accuracy of lactational milk projections has not been widely studied. Because DHIA production measurements usually occur at approximately 30 day intervals for most farms, it is likely that lactational production estimates are based on milk yields occurring on approximately same day of the 14-day bST injection cycle each month. Therefore, if 108 daily production is significantly more variable in bST-treated cows, this inconsistence may negatively affect the accuracy of DHIA lactational production estimates. The most precise measurement of the production response attributable to bST requires the use of daily, individual-cow milk weights by which all milked produced by each cow is recorded, and are currently considered to be the “gold standard” of milk yield measurements. However, this technology is currently quite expensive and not in widespread use. Therefore, few farms have the ability to make daily milk measurements and the management of such data is difficult and time-consuming. As a result, most farms using bST lack the ability to directly measure the effect bST has on milk production in their cows. Consequently, most producers must rely on DHIA milk production projections for making management decisions in bST-treated cows. This study was a clinical trial involving four Michigan dairy herds that had electronically recorded, daily, individual-cow milk weights. The objective was to determine the effect of bST (used according to label directions) on patterns of milk production and the effect of bST on the accuracy of DHIA lactational milk estimates. SPECIFIC METHODS Milk production before the bST treatment period (the first 63 days of the study lactation) was compared between study groups for each farm (pretreatment period). For most results from the treatment (bST administration) period, cows which did not complete their 305-day lactation because of death or culling were included in this analysis. However, for the comparison of DHIA production estimates with daily, 109 individual-cow milk measurements, only cows which completed at least 305 days of lactation were used. Actual milk produced and DHIA production estimates were compared at four times during the study lactation: At the first, second and third monthly DHIA tests after bST treatment was initiated, and at 305 days of lactation (final DHIA production estimate). The variation in daily milk production within each bST injection cycle was found by first limiting the data to only complete injection cycles (milk weights present for all 14 days of the cows bST injection cycle). Standard deviation in milk production was found for each injection cycle and for all 17 injection cycles together as a measure of production variability. lactational persistency (rate of decline in daily production) was calculated by finding the average milk production within each complete 14-day bST injection period and then dividing average production for each of these 17 cycles by the average of the first (baseline) injection cycle (Amos et al., 1985). The persistency for each injection cycle and for all 17 injection cycles averaged together was then calculated for primiparous, multiparous and all study cows and then compared between bST-treated and control cows. The SAS System for Windows“ was used for all statistical analysis (SAS Institute, 1994). All milk production measurement variables, variability in production, lactational persistency and stage (day) of lactation calculations were compared between bST-treated and control cows by Student’s t-test. A repeated measures AN OVA was used to model the difference in production between actual milk produced (measured by 110 daily, individual-cow milk weights) and DHIA estimated lactation production. Difference in production between these two measures was used as the repeated (dependent) variable and dummy variables for herd effect were included in the model as well. Pearson’s correlation coefficient was used to analyze the correlation of DHIA lactational milk estimates with actual milk produced. These correlation coefficients were then compared between the treatment groups by use of a test statistic which utilized a Fisher 2 transformation (Neter et al. , 1996). RESULTS Pretrial milk production Because cows were randomly assigned to study treatment groups, milk production in the pretreatment period should theoretically have been approximately balanced between the study groups (Table 6.1). However, production differed between the treatment groups on one farm (Farm 3) with control cows producing 192.2 kg more milk in the pretreatment period compared to bST-treated cows (P = 0.03). In the remaining three herds, no significant difference in pretreatment milk production was found between the study groups. For all study cows, pretreatment milk production was similar between the study groups with a 47.1 kg difference observed between the treatment groups (P = 0.34). 111 TABLE 6.1: Pretreatment (from 4 to 63 days of lactation) milk production estimates for actual milk produced for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co. , St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows (kg / first 63 days of lactation). bST, (n) Control, (11) Total n P- value Farml 2471.2 (41) 2572.7 (27) . 68 0.36 Farm2 1766.5 (111) 1745.0 (113) 224 0.77 Farm3 2170.0 (69) 2362.2 (66) 128 0.03 Farm4 2073.1 (63) 2177.1 (65) 135 0.23 Lact = 1 1721.8 (107) 1675.7 (86) 193 0.45 lact > 1 2223.1 (177) 2270.0 (185) 362 0.43 Total 2034.3 (284) 2081.4 (271) 555 0.34 Patterns of milk production Variability (measured by standard deviation) in milk production did not differ significantly among the 17, 14-day bST injection cycles and between bST-treated and control cows (Figure 6.1). Standard deviation (variance) in production tended to differ more during the middle of the 17 injection trial period than during either early or later in lactation. Standard deviation in production tended to differ more during the interval between the sixth and eighth injections but was similar between the treatment groups at all other injection cycles and the mean standard deviation of all 17 injection cycles considered together was marginally greater in the bST-treated compared to the control group ( P = 0.08). However, when a Bonferroni adjustment was employed because of multiple 112 comparisons made (N eter et al. , 1996), none of these comparisons were significant at the required value of P = 0.003 (0.05 / 18). FIGURE 6.1: Variability (standard deviation) in daily milk production within each of 17 injection cycles for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co. , St. Louis, MO) (A) at l4—day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows ( 0). Stars indicate standard deviation of milk production differed between the study groups (P < 0.05). as! so. as. C .2 fl .2 > O can. '6 h N 13 I: g .. 755‘ 710. 1....L 1* “IT‘UVT*IUIIIIIUIII 01234567891011121314151617 IMGIQde 113 Production diverged rapidly between the study treatment groups during the first three bST injection cycles (Figure 6.2). The upward shift in production as a result of bST treatment was most apparent by observing the divergence in milk production on the first day of each of the 17 injection cycles (Figure 6.3). Production response to bST was apparent at the first administration of the product in both primiparous (Figure 6.4), multiparous (Figure 6.5) and for all study cows (Figure 6.6). However, the maximal response was not apparent until approximately the middle of the 17 injection treatment period. FIGURE 6.2: Daily milk production during the first three of 17 injection cycles for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co. , St. Louis, MO) (A) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows ( O) . Vertical lines indicate day of bST administration. 41- 381 Milk Preduction (kg I day) 37- 36 1 I I T I I I I I II 1 14 27 40 Days after bST initiation 114 FIGURE 6.3: Milk production on the first day of each of 17 injection cycles for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co. , St. Louis, MO) (A) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows ( O). 40 384 > N '5 ‘ O 35. 5 C .9 a 3 .5 33. O h n. i E 30. 284 a U I U I U I U I I I' U f I U 1234567891011121314151617 lriectlonO/cle 115 FIGURE 6.4: Actual (five-day interval average) milk production for primiparous cows supplemented with 500 mg of bST (Posilac“, Monsanto Co. , St. Louis, MO) (A) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows. Vertical line indicates initiation of bST treatment. 53. .l 50. . 3:47. (B 2 . ‘E2414. 8 ‘ 1 “341. If 3 13 e 4 138.. 5 . E .I 35. 32. . aITIIIIIIIIIIIIIjIIfiIlrWIIIIIIII 8 63 118 173 228 283 Daysinlwlk 116 FIGURE 6.5: Actual (five-day interval average) milk production for multiparous cows supplemented with 500 mg of bST (Posilac“, Monsanto Co. , St. Louis, MO) (A) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows. Vertical line indicates initiation of bST treatment. 47 .. MiIIiRIlhddofilmt MIMI h 2 3 4i 5 6 7 8 91011121314151617 29 WIITIIIIIIIIII 8 63 118 173 228 283 Days in Milk 117 FIGURE 6.6: Actual (five-day interval average) milk production for all cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, MO) (A) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows ( O). Vertical line indicates initiation of bST treatment. 53. . 50.. 3:47- I’" g . . a «/ .544. C .2 . ’6411 3 U . o .l h 0-38. as . 2 as! 32. 29fITIIIIIIIIIIIIIIIIIIIIIITIIII 8 63 118 173 228 283 Days in Milk 118 Average peak milk production was 50.8 kg / day and occurred at an average of 113.9 days of lactation for bST-treated cows while average peak production was 48.9 kg per day occurring at an average of 86.4 days of lactation for control cows (Table 6.2). Both average peak production and average time to peak production significantly greater for bST-treated cows compared to controls (P = 0.01 and P < 0.0001, respectively). While average peak production was not greatly different between bST-treated and control multiparous cows (P = 0.09), bST-treated primiparous cows had higher average peak milk production when compared to similar aged controls (P < 0.0001). TABLE 6.2: Peak milk production and time to peak milk production for cows supplemented with 500 mg of bST (P0silac“, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows. Milk production bST Control Difference Total P -value Peak milk production (kg / day) lact = 1 46.4 42.3 4.1 44.6 0.0001 lact > 1 53.4 51.9 1.5 52.6 0.0925 Total 50.8 48.9 1.9 49.8 0.0119 Time to peak production (dayS) lact = 1 138.0 108.8 29.2 125.0 0.0010 lact > 1 99.4 76.0 23.4 87.4 0.0001 Total 113.9 86.4 27.5 100.5 0.0001 Study cows treated with bST were significantly more persistent in lactation (7.0% greater lactational persistency) compared to control cows (P < 0.0001) with bST-treated 119 primiparous cows 6.3 % more persistent (P = 0.0003) and bST-treated multiparous cows 6.1% more persistent (P < 0.0001) and than control cows. Correlation and accuracy of milk production measurements Actual milk produced and all DHIA production estimates analyzed were found to be significantly correlated (P < 0.001) when compared by Pearson correlation analysis (Table 6.3). However, correlation coefficients for all DHIA estimates and actual milk produced were not significantly different between the study treatment groups for any of the four comparisons made (first, second, third monthly tests after bST treatment initiation and final (305-day) DHIA production estimates). TABLE 6.3: Correlation (Pearson’s) of the first three (afier bST treatment initiation) and final (305-day) DHIA milk production estimates with actual milk produced (daily milk weight measurements) for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows. Milk production bST (rbST) Control (rm) P -value measure First DHIA estimate 0.7242 0.7334 0.41 Second DHIA estimate 0.8329 0.8323 1.00 Third DHIA estimate 0.8501 0.8382 0.35 Final DHIA estimate 0.9753 0.9824 1.00 While correlations were not different, the accuracy of DHIA production estimates was significantly affected by the amount time elapsed since bST initiation (Table 6.4). 120 Milk production at the first DHIA test date after bST initiation underestimated actual production by 4.7% for bST-treated cows and by 0.8% for controls yielding a 3.9% difference in accuracy (P = 0.001). Similarly, DHIA underestimated production by 2.7 % for bST-treated cows and overestimated production by 0.1% for control cows (2.8% absolute difference) at the second test date after bST initiation (P = 0.002). However, the difference in accuracy (0.3 % absolute difference ) was not significant by the third DHIA test date (P = 0.19). Therefore, by 305 days of lactation (final estimate), DHIA overestimated actual production by 2.4% for both the bST-treated and control groups, and accuracy had improved to the point that no significant difference existed between the treatrnent groups (P = 0.68). A repeated measures ANOVA model was also constructed to analyzed the effect of bST on DHIA lactation milk estimates. In this model, the repeated measure dependent variable) was the difference between the actual milk produced during lactation (as measured by daily, individual cow milk weights) and the DHIA estimates made at approximately 30, 60 and 90 days after bST initiation and the final (305-day) estimate. The independent variables of farm and study treatment group were included in this model. The difference in production between actual and estimated production was analyzed by orthogonally contrasting this difference at the first three time frames (30, 60 and 90 days after bST initiation) with the difference at the final (305-day) estimate. The results of this analysis confirmed the bivariate comparison because the variable for bST treatment effect was significant (P < 0.001) at the first two time periods but was not significant by the third time frame (P <0.091). 121 TABLE 6.4: Differences in milk production between the first three (after bST treatment initiation) and final (305-day) DHIA milk production estimates and actual milk produced (measured by daily milk weight measurements) for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, M0) at 14—day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows. bST Control Difference P - Milk production kg / lactation kg / lactation (bST - control)3 value measure Actual milk production 10195.9 9645.0 550.9 0.004 First DHIA estimate 9721.8 9572.3 149.5 0.417 Difference (actual - 509.6 73.1 436.5 0.001 first DHIA estimate) Second DHIA estimate 9916.4 9657.3 259.1 0.148 Difference (actual - 315.3 (12.3) 327.6 0.002 second DHIA estimate) Third DHIA estimate 10128.6 9681.4 447.2 0.019 Difference (actual - 102.9 (36.3) 139.2 0.188 third DHIA estimate) ' Final DHIA estimate 10436.4 9869.6 566.8 0.003 Difference (actual - (240.4) (224.5) 15.9 0.679 final DHIA estimate) aAbsolute difference in milk production between bST-treated and control cows DISCUSSION Twenty-two cows that were originally selected for the trial were in poor health at the time when bST treatment was to begin and therefore, were excluded from the analysis. 122 This exclusion was in accordance with the product label that states cows must be determined healthy if bST treatment is to begin during the ninth week of lactation (Monsanto Company, 1993). Because less than 5 % of the cows selected for the study were excluded, and that cows scheduled to be treated with bST and control cows were excluded based on the same criteria, it is unlikely that a sizable bias was introduced from this omission. Milk production in the pretreatment period was well balanced between the treatment groups except for one farm (farm three) where control cows produced significantly more milk than those treated with bST. However, for all study cows production during the pretreatment period was similar during the first 63 days of lactation. In this study, pre-existing IMI were balanced between treatment and control cows (Judge et al. , 1997) and therefore should not have affected any milk production differences observed. This was important because cows with IMI may produce less milk than cows without infection. While intuitively it would seem milk production might be more variable in bST- treated cows, I observed similar overall variation in daily milk production between bST- treated and control cows (Figure 6.1). All cows tend to vary in production from day to day, and in this study bST-treated cows were observed to vary in daily production no more than control cows. However, production did vary during each 14-day injection period. A similar pattern of production was found in the present study compared to that reported by other workers (Bauman et al. , 1989), with production increasing from day one to seven, maximum response occurring on approximately day nine and then gradually decreasing until day 12 (Figure 6.2). Response was generally in decline on the 13Uh and 14‘h days, but 123 bST-treated cows still produced more milk than control cows during all 17 injection cycles on these days as well. The milk production response observed in this study was less than observed by other workers (Bauman et al., 1989; Remond et al., 1991) but similar to that found in a large study conducted in the Northeastern US (Bauman et al., 1999). The results of the present study are in contrast to the greater change in milk yield observed as a result of bST treatment by other workers (Bauman et al., 1989; Remond et al., 1991; Skarda et al., 1992). However, cows milked 3X may respond less to bST treatment than do those milked 2X (Speicher et al. , 1994). If cows in the present study had been milk 2X and responded similarly to those in another study, a production response of 4.9 kg / day (3.6 kg / day x 0.37 greater response) and 4.3 kg / day (2.8 kg / day x 0.54 greater response) response during the trial period between the study treatment groups for primiparous and multiparous cows might have been observed, respectively (Speicher et al. , 1994). However, Speicher et al. did not use commercially-available bST (14 mg bST was administered daily) (Speicher et al., 1994) and this difference may be responsible for the disparity in production response between studies. Nevertheless, the results of this study should be representative of those herds who milk cows 3X, they may not be as accurate for herds that milk cows 2X. Parity can influence the magnitude of production response to bST (Hartnell, 1994), and in the present study a 27 % greater production response from bST use was observed in primiparous compared to multiparous cows. Most studies have observed either multiparous cows responding either more favorably (Crooker and Otterby, 1991; Thomas 124 et al. , 1991) or the same as primiparous cows (Bauman et al. , 1989; Remond et a1. , 1991), I with one study observing a greater response in primiparous cows (lamb et al., 1988). However, independent of bST treatment, primiparous cows have been found to respond more favorably to 3X milking than multiparous cows (Allen et al., 1986; Amos et al. , 1985). The greater response to increased milking frequency by younger cows may be because their udder capacity is the major production limiting factor. Therefore, more frequent emptying of smaller udders may result in a proportionally greater response to both bST treatment and milking frequency compared to more mature cows with larger udders (Amos et al. , 1985). However, the increased lactational persistency characteristic of primiparous cows is greater than can be explained by a reduction in intra-mammary pressure alone (Pelissier et al. , 1978). Additionally, one author suggested that primiparous animals responded greater to bST than multiparous cows when all cows are milked 3X because multiparous cows are already producing near their genetic potential prior to the initiation of bST (Speicher et al. , 1994). However, other workers have concluded that production response to bST is independent of milk frequency (Hartnell, 1994; Thomas et al.», 1991). As lactation progressed, production response to bST declined slightly (Figures 6.4, 6.5 and 6.6). This result was also observed in other studies and was attributed to either a decrease in body condition or a reduction in the number or activity of mammary secretory cells (Remond et al. , 1991). In this study, all primiparous cows peaked at an average of 125 days of lactation and all multiparous cows peaked at average 87.4 days of lactation with all cows peaking on the average at 100.5 days of lactation. These production peaks occurred later than 125 those observed by other workers (Amos et al. , 1985) with bST-treated cows peaking nearly one month later in lactation than did control cows. Therefore, within each production group, bST treatment created a second peak in production occurring approximately one month after the first peak (Figures 6.4, 6.5 and 6.6). These results have implications for producers who manage bST-treated cows because the significant change in the shape of the lactation curve affects the nutritional needs of these cows. Therefore, producers would be advised to monitor milk production as well as body condition and use both of these parameters when making nutritional decisions for bST-treated cows rather than simply basing these decisions on stage of lactation (Crooker and Otterby, 1991). While production response was apparent after the first bST administration, maximal response was not observed until several injections were performed (Figures 6.2 and 6.3). Other workers have found that maximal response was attained approximately 5 weeks after the initiation of treatment (Downer et al. , 1993). However, in this study maximal response was not observed until approximately the middle of the 238-day treatment period. Nevertheless, it was clear that bST was associated with greater lactational persistency and therefore altered the shape of the lactation curve. ' In this study, both primiparous and multiparous cows were nearly equally persistent in lactation with primiparous cows displaying slightly greater persistency. Overall, bST-treated cows displayed approximately 7 % greater lactational persistency than did control cows. This result is in agreement with other studies measuring bST-associated increased lactational persistency (Bauman, 1992). However, other workers observed greater persistency in primiparous compared to multiparous cows (Thomas et al. , 1991) while other studies found that persistency was no 126 greater in bST-treated compared to control cows and that production differed between the treatment groups only as a result of higher peak production (Thomas et al. , 1991). The findings of the present study concur with other results that contend the increase in 305-day milk production results from both increased peak milk production and greater lactational persistency (Bauman, 1992). Because DHIA 305-day milk production models were developed prior to the commercial use of bST, these models were not designed to adjust for either the second peak in production or the increased lactational persistency exhibited by bST-treated cows. Therefore, these models tend to underestimate bST-assisted milk production until they can compensate for the change in shape of the lactation curve caused by bST. At the first DHIA test date following bST initiation, production estimates were significantly less accurate: for bST-treated compared to control cows. This inaccuracy persisted until the third DHIA production estimate was made. Therefore, these models significantly underestimated actual 305—day production for approximately 100 days (approximately the seventh injection . cycle) after bST treatment was initiated. However, by 305 days of lactation, accuracy was improved to the point that no significant difference in the accuracy of DHIA milk production estimates was observed between the treatment groups. Therefore, while significantly greater error existed for early lactational estimates, it was clear that by the end of lactation DHIA models accurately predicted actual milk production for both bST-treated and control cows. The primary reason for this inaccuracy is that the equations used are not sensitive to large upward shifts in production such as normally occurs at the time of bST initiation 127 (Galligan et al. , 1992). Related to this sudden increase in production, these models also have difficulty accounting for more than a single peak in production during a single lactation (Weigel et al., 1992). However, because milk production was found to be no more variable for bST-treated compared to control cows, the day during the 14-day bST injection cycle on which milk measurements are taken for the purpose of making DHIA production estimates should not contribute to the observed inaccuracy. Nevertheless, while these models seem to be adequate for estimating 305-day milk production after the first 100 days of bST treatment, additional (adjustment) factors appear to be needed to improve accuracy during the early bST treatment period. The major implication of the inaccuracy observed for DHIA estimates is that improper culling decisions may be made when cows are culled because of low milk production when such decisions, based on DHIA production estimates, are made during the early bST treatment period (Norman et al. , 1999). Dairy producers and industry consultants need to take this feature of bST-assisted milk production into consideration when making culling decisions based on these production estimates for bST-treated cows. Better decisions may be made .by delaying this type of culling decision for bST-treated cows until at least the third DHIA production estimate is available. At this point, more confidence could be placed the production estimates made by DHIA and therefore in the culling decision process as well. 128 CONCLUSIONS Despite changing the shape of the lactation curve, because of both later and greater peak milk production as well as increased lactational persistency, bST did not significantly decrease the accuracy of 305—day (final) DHIA milk production estimates. Therefore, producers should have confidence in these final lactational production estimates regardless of whether cows are treated with bST or not. However, earlier DHIA estimates contain significantly more error for bST-treated compared to control cows. Therefore, while it appears that separate regression models are not necessary for the accurate prediction of final lactational milk production, adjustrnent factors may be useful for improving the accuracy of production estimates for bST-treated cows during the first approximately 100 days following the initiation of bST treatment. Producers should take this period of inaccuracy into consideration when making culling decisions for bST-treated cows based on DHIA-predicted milk production. Also, as a result of causing a second peak in production occurring after the normal peak, producers need to consider body condition more so than stage of lactation when making production management decisions for bST- treated cows. 129 REFERENCES Allen, D. B., E. J. DePeters, and R. C. Labin. 1986. Three times a day milking: effects on milk production, reproductive efficiency and udder health. J. Dairy Sci. 69: 1441. Amos, H. E., T. Kiser, and M. Loewenstein. 1985. Influence of milk frequency on productive and reproductive efficiencies of dairy cows. J. Dairy Sci. 68:732. Bauman, D. E.. 1992. Bovine Somatotropin: Review of an emerging animal technology. J. Dairy Sci. 75:3432-3451. Bauman, D. E., R. W. Everett, W. H. Weiland, and R. J. Collier. 1999. Production Responses to Bovine Somatotropin in Northeast Dairy Herds. J. Dairy Sci. 82:2564-2573. Bauman, D. E., D. L. Hard, B. A. Crooker, M. S. Partridge, K. Garrick, L. D. Sandles, H. N. Erb, S. E. Franson, G. F. Hartnell, and. R. L. Hintz. 1989. Long-term evaluation of a prolonged-release formulation of N -methionyl bovine somatotropin in lactating dairy cows. J. Dairy Sci. 72:642-651. Brown, C. A., P. T. Chandler, and J. B. Holter.‘ 1977. Developement of Predictive equations for milk yield and dry matter intake in. lactating cows. J. Dairy Sci. 60:1739-1754. Congleton, W. R., Jr., and R. W. Everett. 1980. Application of the incomplete gamma function to predict cumulative milk production. J. Dairy Sci. 63: 109-1 19. Crooker, B. A.and D. E. Otterby. 1991. Management of the dairy herd treated with bovine somatotropin. Vet. Clin. of North Am. Food Anim. Pract. 7(2):417. Dairy Comp 305“, Version 3.1. Valley Agricultural Software. Tulare CA. 1990. Downer, J. V., D. L. Patterson, D. W. Rock, W. V. Chalupa, R. M. Cleale, J. L. Firkins, G. L. Lynch, J. H. Clark, B. O. Brodie, B. F. Jenny, and R. De Gregorio. 1993. Dose titration of sustained-release recombinant bovine somatotropin in lactating dairy cows. J. Dairy Sci. 76:1125-1136. Excel“, Version 5.0. 1993. Microsoft Corporation, Redmond, WA. 130 Erdman, R. A., and M. Varner. 1995. Fixed yield responses to increased milking frequency. J. Dairy Sci. 78:119—1203. Galligan, D. T. , J. D. Ferguson, W. Chalupa, C. F. Ramberg. 1992. Quantifying changes in milk production: don’t let someone throw you a curve. Compend. l4(4):549- 553. Gill, J. L. 1978. Design and Analysis of Experiments. The Iowa State University Press, Ames, IA. Hartnell, G. F. 1994. Bovine somatotropin in the dairy industry: a review. Prof. An. Sci.10:85-101. Judge, L. J ., P. C. Bartlett, J. W. Lloyd and R. J. Erskine. The impact of recombinant somatotropin on net farm income: analysis of four Michigan dairy farms. Submitted for publication in J. Dairy Sci. Judge, L. J ., P. C. Bartlett, J. W. Lloyd and R. J. Erskine. 1999. Recombinant bovine somatotropin: association with reproductive performance in dairy cows. Theriogenology. 52:481-496. Judge, L. J ., R. J. Erskine and P. C. Bartlett. 1997 . Recombinant bovine somatotropin and clinical mastitis: incidence, discarded milk following therapy. and culling. J. Dairy Sci. 80:3212-3218. Jones, T.. 1997. Empirical Bayes prediction of 305-day milk production. J. Dairy Sci. 80:1060-1075. Keown, J. F. , and L. D. Van Vleck. 1972. Extending lactation records in progress to 305- day equivalent. J. Dairy Sci. 56:1070-1079. Lamb, R. C., M. J. Anderson, S. L. Henderson, J. W. Call, R. J. Callan, D. L Hard, and L. Kung, Jr. 1988. Production response of Holstein cows to sometribove USAN (recombinant methionyl bovine somatotropin) in a prolonged release system for one lactation. J. Dairy Sci. 71 (Suppl. l):208(Abstract). Neter, J ., M. H. Kutner, C. J. Nachtsheirn, and W. Wasserrnan. 1996. Applied Linear Statistical Models. Irwin, Chicago and Boston. Norman, H. D., P. M. VanRaden, J. R. Wright, and J. S. C . 1999. Comparison of test interval and best prediction methods for estimation of lactation yield from monthly, a.m.-p.m. , and trirnonthly testing. J. Dairy Sci. 82:438-444. 131 Pelissier, C. L., L. J. Koong, and L. F. Bennett. 1978. Influence of 3X milking on milk and milkfat production. J. Dairy Sci. (Suppl.):132. (Abstr.). Posilac“ product label. 1993. Monsanto Co., St. Louis, MO. Remond, B., M. Cisse, A. Ollier, and Y. Chilliard. 1991. Slow release somatotropin in dairy heifers and cows fed two levels of energy concentrate. 1. Performance and body condition. J. Dairy Sci. 74:1370-1381. SAS System for Windows“, Version 6.10, 1994. SAS Institute, Cary, NC. Skarda, J ., E. Markalous, J. Slaba, P. Krejci, O. Skardova, and J. Zednik. 1992. Effect of methionyl bovine somatotropin in a prolonged-release vehicle on milk production, hormone profiles and health in dairy cattle. J. Dairy Res. 59:499-506. Spartan Dairy Ration Evaluator/Balancer, Version 2.01. 1992. Cooperative Extension Service and the Agricultural Experiment Station. Michigan State University, E. Lansing, MI. Spreicher, J. A., H. A. Tucker, R. W. Ashley, E. P. Stanisewski, J. F. Boucher, and C. J. Sniffen. 1994. Production responses of cows to recombinantly derived bovine somatotropin and to frequency of milking. J. Dairy Sci. 77:2509-2517. Thomas, J. W., R. A. Erdman, D. M. Galton, R. C. Lamb, M. J. Arambel, J. D. Olsen, K. S. Madsen, W. A. Samuels, C. J. Peel, and G. A. Green. 1991. Responses by lactating cows in commercial dairy herds to recombinant bovine somatotropin. J. Dairy Sci. 74:945-964. US Dept. of Agriculture, Science and Education Administration, Dairy Herd Improvement Letter. 1980. Projection factors for milk and fat lactation records. 56(1), March. Weigel, K. A., B. A. Craig, T. R. Bidwell, and D. M. Bates. 1992. Comparison of alternative diphasic lactation curve models under bovine somatotropin administration. J. Dairy Sci. 75:580-589. 132 CHAPTER SEVEN THE IMPACT OF RECOMBINANT BOVINE SOMATOTROPIN ON NET FARM INCOME: ANALYSIS OF FOUR MICHIGAN DAIRY FARMS ABSTRACT Holstein cows (11 = 555) from four Michigan dairy farms were randomly assigned as untreated controls or to receive bovine somatotropin (500 mg) administered subcutaneously every 14 days beginning at 63 to 69 days of lactation and continuing until approximately 21 days prior to the end of lactation or until removal from the herd. The objective was to determine the effect of bST used according to labeled directions on Net Farm Income per cow when direct costs of bST use (cost of bST, labor for bST administration and additional feed) were considered. Net Farm Income from bST initiation (ninth week of lactation) until 301 days of lactation (17 injections) was increased by an average of $83.14 per bST-treated primiparous cow, $32.98 per bST-treated multiparous cow and $43 .01 per cow for all bST-treated cows compared to untreated controls. For the four study farms, bST use changed Net Farm Income by $96.21 , $3.57, $78.71 and $-7. 15 per bST-treated cow, respectively. Break-even analysis found 2.3 kg of additional milk per day per bST-treated cow necessary to cover direct costs of bST use, and at the observed level of production response (2.9 kg), a break-even milk price of $21.86 per 100 kg was calculated. Sensitivity analysis found the change in Net Farm Income from bST use most sensitive to changes in milk price and bST production response while less sensitive to changes-in the price of bST and feed. Twin calvings reduced bST-associated 133 increased Net Farm Income by $0.78 per bST-treated cow which decreased the expected increase in Net Farm Income to $42.23 per bST-treated cow. While not observed in this study, diseases that may be associated with bST could reduce profits further. Results of this study are consistent with previous work supporting the ability of bST to enhance dairy farm profits. 134 INTRODUCTION Bovine somatotropin (Posilac“; Monsanto Company, St. Louis, MO) was approved for lactating dairy cows in November 1993, with the first commercial use occurring in early 1994. From its release for commercial sale until the present, well over 100 million doses of bST have been purchased by US dairy producers, and in 1996 it was estimated that approximately 10% of eligible cows in the US were treated with bST (Centers for Epidemiology and Animal Health, 1996; Dickrell, 1999). While many dairy producers have perceived an increase in milk production and profitability with bST use, most estimates of increased production and financial gains occurred during the preapproval period (Elbehri and Yonkers,'1995; Marion and Willis, 1990; Remond et al., 1991, Shoeffling et al. , 1991). Because preapproval studies used various bST doses and lengths of treatment, these estimates may not be applicable when bST is used according to labeled directions (Fallert et al., 1987, Marion and Willis, 1990). In fact, it has been suggested that the production response from bST use in commercial herds may be as much as 25 % less than was observed under experimental conditions (Fallert, 1987). Despite preapproval studies which predicted that milk prices would likely fall because of increased production nationwide, milk production per cow and total US production have not changed dramatically following the introduction of bST (Dickrell, 1999; US Department of Agriculture, National Agricultural Statistical Service, 1997). While it is clear this prediction did not materialize, several factors (environmental conditions, fluctuations in numbers of cows, etc.) may have masked the effect bST use may 135 have had on US milk production. Nevertheless, the extent to which individual dairy farms may have benefitted from bST use is currently unclear and has not been widely studied since its introduction. To achieve accurate measurements of the production response and the associated change in N F1 from bST use requires the utilization of daily, individual-cow milk weights. However, few farms have the ability to make daily milk measurements and the management of such data is difficult and time-consuming. As a result, most farms using bST lack the ability to accurately measure the effect bST has on milk production and farm profitability. In addition, to accurately measure bST profitability, any increase in disease- related costs associated with bST use must be subtracted from any financial gains. ~ However, most prior estimates of profits generated from bST-assisted milk production have been made without considering any increase in disease resulting from use of bST. Recently, an expert panel in Canada concluded that bST increases the risk of clinical mastitis by 25 % and clinical lameness by 50% (Canadian Veterinarian Medical Expert Panel, 1998). Such significant increases in disease rates might substantially increase the cost of bST use. Partial budgets are useful for estimating expected changes in revenues and expenses resulting from a management change such as use of bST (Lloyd and Hady, 1994). By use of daily, individual-cow milk weights to calculate increased milk production and then subtracting direct costs attributable to bST (bST product purchase, increased feed and labor and any increased disease costs) it is possible to estimate the change in profit per bST- treated cow. 136 This study was a clinical trial involving four Michigan dairy herds that had daily, individual-cow milk weight capability. Our objective was to determine the effect of bST (used according to labeled directions) on Net Farm Income per bST-treated cow per lactation. SPECIFIC METHODS Clinical mastitis results from our data have been previously reported (Judge et al. , 1997). Because mastitis can affect milk production (Bartlett et al. , 1990; Deluyker er al. , 1993) it was necessary to document that mastitis parameters were similar between the study treatment groups before calculating production differences. A cost estimate was calculated for'cows experiencing twin calvings. By using the method of Beerepoot et al. (Beerepoot et al. , 1991), an increase in revenue from extra calves produced was found. Significant costs observed in our data resulting from twin calvings included increased calf death loss, increased rates of retained placenta and metritis, increased days open and increased culling rates. These costs were subtracted from the increased calf value to arrive at an estimated cost per case of twins. . The only costs of retainedplacenta and metritiswere determined to be for therapy used to treat these diseases. Using the method of Hady et al. (Hady et al., 1994), an average value for the cost of 39 extra days spent open ($48.75) was used. This value ($1.25 / day) is consistent with values found by review of the literature (Bauman et al., 1989; Huber et al., 1997; Remond et a1. , 1991). The cost of culling was found by first finding the midpoint (average) of the dairy (maximum) value and the cull (minimum) value. Then cull 137 (slaughter) value was subtracted from this calculation to find the relative value of the cow at the time of culling. This value was then multiplied by the relative increase in culling rate that was associated with twinning. A dairy value of $1300, $1200, $1000, $800 and $700 was assigned to cows of parity one, two, three, four and five or greater, respectively. Cull value was taken from market receipts maintained by the farm. For cows missing receipts, the average cull value of other coWs having receipts and culled after experiencing a twin birth were used to complete the data. Cows that died were assigned a zero cull value. Calves were valued at $4.14 and $2.07 per kg of body weight for female and male calves, respectively. The final estimated cost per case of twins was found by discounting the calculated cost by 10% because this loss was not realized until the subsequent lactation. Milk production before the bST trial period (the first 63 days of the study lactation) was compared between study groups for each farm (pretrial period). Similarly, milk production was compared for the bST trial period (from 63 days of lactation up to a maximum of 301 days of lactation = 238 days maximum length). Cows not completing the pretrial period were excluded from the study. However, cows not completing the trial period (either because of death or culling) were included in the analysis until the time of their removal from the herd. Therefore, together the pretrial and trial periods accounted for the first 301 days of lactation. Milk production was also compared during the post-trial period (the remainder of lactation after the first 301 days of lactation up to a maximum of 425 days of lactation = 124 days maximum length) and for the entire study period (from 63 days of lactation up to a maximum of 425 days of lactation or the end of lactation = 362 days maximum length). Milk that was discarded because of disease therapy was not 138 included in the analysis of any milk production variable. The SAS System for Windows“ was used for all statistical analysis (SAS Institute, 1994). Milk production during the pretrial, trial, post-trial and entire study periods was compared between bST-treated and control cows by Student’s t-test. The comparison of rates of disease and reproductive parameters as well as culling between the Study treatment groups has been described in previously (Chapters 3, 4 and 5). Costs of inputs (bST, labor to administer bST and additional feed consumed) were based on data gathered from the study farms, Monsanto Company and the US Department of Agriculture, National Agricultural Statistics Service (USDA, NASS). All prices used in this analysis reflected actual costs incurred during the time frame of this study (1994 through 1995) (Table 7.1). TABLE 7.1: Feed prices (during 1-994-95) used to calculate costs for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows. Feed ingredient Cost per unit Cost per kg Alfalfa silage $70.00/ton $0.08 Corn silage $35.00/ton $0.04 Corn grain $2.29/bu $0.08 Soybean meal $224.00/ton $0.24 Whole cotton seed $100.00/ton $0.11 Dairy feed, 16% $1 80.00/ton $0.20 For feed price calculations, the rations fed by each farm were entered into the Spartan Dairy Ration Balancer (Cooperative Extension Service, Michigan State University , 139 1992). Feed costs used were taken from USDA, NASS historical data (USDA, NASS, 1997). The cost per Meal of each ration was then used to calculate the additional feed cost attributable to bST use and a conversion factor from the National Research Council (0.691 Mcal of NEL /kg of 3.5 % fat milk produced) was utilized to arrive at the total additional feed cost (National Research Council, 1988). Three different costs were used for each farm: 1) the highest-producing multiparous cows, 2) primiparous cows and, 3) a pooled cost for all study cows on the farm. The cost of bST used was the base price of the commercially-available product ($5.86 per . dose (bST base product price plus $0.06 shipping charge per dose)). The financial measure of most interest in this study was N FI. This parameter was chosen because it captures the expected change in farm profitability resulting from bST use and conforms to the recommendations put forth by the Farm Financial Standards Council (Farm Financial Standards Council, 1997). A partial budget (Harsh et al. , 1981) was used to calculate the change in NFI for each farm and the entire study population using only those costs that changed as a result of bST use. For each farm and the entire study, partial budgets were constructed separately for primiparous and multiparous cows. This was necessary due to varying production responses in cows of differing parities and in different herds. Partial budgets used in this analysis were based on the template constructed by Willett et al. (W illett et al. , 1994). Using the results of these partial budgets, a break-even milk price was calculated for the primiparous and multiparous cows on each farm and for the entire study population. Sensitivity analysis was performed on the following variables to examine their 140 effect on the change “in NFI resulting from bST use: Production response to bST, price received for milk, bST product price and feed price. To determine which of these were most sensitive, the values of these variables were varied to determine their effect on the change in NFI resulting from bST use. This was performed by dividing the percent change in the dependent variable (the change in NFI) by the percent change in the independent variable (either response to bST, or milk, bST or feed price). A milk price of $28.09 / 100 kg was used for all analysis. This was the average all-milk wholesale price reported by the National Milk Producers Federation during the period the study 'data were collected (1994 through 1995) (National Milk Producers Federation, 1997). This price does not include any premiums that may have been received by farms for fat, protein, and bacteria or SCC levels. For all inputs and outputs used in this financial analysis, no adjustment was made for inflation, opportunity costs or returns to management. Discounting was determined to be necessary only for the financial estimate made for costs associated with twin calvings in this study. RESULTS - Pretrial milk production Because cows were randomly assigned to study treatment groups, milk production in the pretrial period was expected to be balanced between the study groups. However, production was different between the study groups on one farm (farm three) with the control cows producing 192.3 kg more milk in the pretrial period compared to the bST— treated cows (P = 0.03). In the remaining three herds, no significant difference in pretrial 141 milk production was found between the study groups. For the entire study, pretrial milk production was similar between the study treatment groups with a 47.1 kg difference observed during the first 63 days of lactation (P = 0.34). Disease and reproductive measures The effect of bST on diseases in this study has been previously discussed (Judge et al., 1999). In short, nonspecific off feed conditions (P = 0.05) and twin conceptions during the study lactation (P = 0.05) were found to be positively associated with bST use. However, all other diseases studied were similar between the study treatment groups. All reproductive indices calculated were similar between the study treatment groups (Judge et al. ,' 1999). Milk production measures A wide variation in response to bST was observed between different farms and cows of differing parities (Table 7.2). When considering all parities together, milk production during the trial period (from 63 to 301 days of lactation) differed significantly between the study treatment groups for all comparisons except for farm four. . During the total study period (238-day trial + 124-day post-trial period = 362-day bST administration period) production differed significantly between the treatment groups for all but one farm (farm two). However, milk production during the post-trial period (from 301 to 425 days of lactation) differed significantly between the study groups for only farm three and all farms together. Milk production was significantly different between the study treatment 142 groups when all farms were analyzed together except when primiparous and multiparous cows were considered Overall, cows treated with bST produced 2.9 kg more milk per day through the 238 day trial period (P = 0.0001) with primiparous cows increasing 3.6 kg per day (P = 0.0002) and multiparous cows increasing 2.8 kg per day (P = 0.0004) compared to controls. In the 124-day post-trial period, bST-treated cows yielded 2.2 kg more milk per cow per day“ (P = 0.006) with primiparous cows increasing 2.2 kg (P = 0.05) and multiparous cows increasing 1.7 kg per day (P = 0.10) compared to controls. During the entire study, bST-treated cows yielded 2.8 kg more milk per cow per day (P = 0.0001) with primiparous cows increasing 3.1 kg (P = 0.0007) and multiparous cows increasing 2.8 kg per day (P = 0.0002) compared to controls. 143 TABLE 7.2: Daily milk production of cows supplemented with 500 mg of bST (Posilac“, Monsanto Co. , St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until 362 days of lactation compared to untreated controls (mean kg / cow / day). Farm 1 Farm 2 Farm 3 Farm 4 Total Study Trial period Group kg (n) kg (n) kg (11) kg (11) kg (11) Trial period Primiparous Control 33.67 (4) 26.93 (38) 34.03‘(22) 28.46 (22) 29.45' (86) bST 36.11 (15) 28.53 (32) 36.97" (39) 30.66 (21) 33.09” (107) Multiparous Control 34.32 (23) 29.72‘ (75) 36.13‘(44) 32.01 (43) 32.35' (185) bST 39.41 (26) 31.88" (79) 41.78" (30) 33.72 (42) 35.10”(177) Total Control 34.22‘(27) 28.78‘ (113) 35.43' (66) 30.81 (65) 31.43‘ (271) bST 38.20” (41) 30.91"(111) 39.06" (69) 32.70 (63) 34.34” (284) Post-trial period Primiparous Control 18.26 (3) ' 20.84 (26) 24.69 (20) 24.74 (10) 22.68 (59) bST 25.83 (6) 23.13 (21) 26.09 (36) 23.88 (12) 24.89 (75) Multiparous Control 22.35 (13) 18.33 (46) 17.15' (29) 17.21 (17) 18.32 (105) bST 24.68 (14) 17.62 (55) 24.89” (19) 17.67 (12) 20.00 (100) Total Control 21.59 (16) 19.24 (72) 20.23‘ (49) 20.00 (27) 19.89‘ (164) bST 25.02 (20) 19.14 (76) 25.68” (55) 20.78 (24) 22.09” (175) All study period Primiparous Control 32.31 (4) - 25.89 (38) 32.62 (22) 27.89 (22) 28.42' (86) bST 32.32 (15) 27.06 (32) 34.55 (39) 29.86 (21) 31.50" (107) Multiparous Control 32.52 (23) 27.91 (75) 32.89' (44) 30.91 (43) 30.37' (185) bST 37.53 (26) 29.51 (79) 39.17" (30) 32.85 (42) 33.12” (177) Total Control 3249‘ (27) 27.23 (113) 32.80'(66) 29.89‘(65) 29.75' (271) bST 36.72"(41) 28.80 (111) 36.56"(69) 31.85"(63) 32.51"(284) ‘b Different subscripts in each column within each parity group differed in milk production between bST-treated cows and untreated controls (P < 0.05). 144 Financial measures Calculation of the change in NFI attributable to bST was made considering costs of the bST product, labor to administer injections and additional feed required to support extra milk produced. The use of bST changed NFI by $96.21, $3.57, $78.71 and $-7.15 per bST-treated cow, respectively, for the trial period. The overall average change in N FI attributable to bST was $43.01 per bST-treated cow for the trial period. A large variation in N FI gain was found among the four farms studied (Table 7.3). An increase in NFI of $96.21 for the trial period was observed on one study farm while another farm experienced a $7.15 loss resulting from bST use during this time. Because bST-treated primiparous cows experienced an approximately 27% greater production response to bST compared with multiparous cows, the younger cows returned a two-fold greater increase in NFI for the trial period compared with older cows (Table 7.4). 145 TABLE 7.3: Partial budget analysis for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, M0) at l4-day intervals beginning at approximately 63 days of lactation until approximately 301 days of lactation ($ per supplemented period per cow). separately during the post-trial period. Farm 1 Farm 2 Farm 3 Farm 4 A. INCREASED INCOME l. Supplemented days per lactation 238 days 238 days 238 days 238 days 2. Average daily increased milk 3.9 kg 2.1 kg “3.6 kg 1.9 kg 3. Total additional production from bST 928.2 kg 499.8 kg 856.8 kg 452.2 kg 4. Gross milk price per 100 kg $28.09 $28.09 $28.09 $28.09 5. Total added revenue from bST $260.73 $140.39 $240.68 $127.02 B. REDUCED COSTS None identified in this analysis $0.0 $0.0 $0.0 $0.0 C. SUBTOTAL (section A + B) $260.73 $140.39 $240.68 $127.02 D. REDUCED INCOME None identified in this analysis $0.0 $0.0 $0.0 $0.0 E. INCREASED COSTS 6. Number of bST doses l7 17 17 17 7. Cost per dose of bST $5.86 $5.86 $5.86 $5.86 8. Total bST product cost $99.62 $99.62 $99.62 $99.62 9. Time required to inject cows per dose 1 min. 1 min. 1 min. 1 min. 10. Labor cost per hour $10.00 $10.00 $10.00 $10.00 11. Total labor cost $2.83 $2.83 $2.83 $2.83 12. Additional Meal required‘ 641 Meal 345 Meal 592 Meal 312 Meal 13. Feed cost per Meal $0.087 $0.082 $0.083 $0.084 14. Total additional feed cost $55.80 $28.32 $49.14 $26.25 15. Other added costs” $11.23 $6.05 $10.37 $5.47 16. Total additional costs $169.49 $136.82 $161.96 $134.17 F. SUBTOTAL (section D + E) $169.49 $136.82 $161.96 $134.17 NET GAIN (line F - line C) $91.25 $3.57 $78.71 $(7.15) ‘ (0.691 Meal / kg of milk) 5 (milk marketing = $1.21 per 100 kg) 146 TABLE 7.4: Partial budget analysis for cows supplemented with 500 mg of bST (Posilae“, Monsanto Co. , St. Louis, M0) at 14—day intervals beginning at approximately 63 days of lactation until approximately 301 days of lactation compared to untreated control cows ($ per supplemented period per cow). lact = 1 lact > 1 Total A. INCREASED INCOME 1. Supplemented days per lactation 238 days 238 days 238 days 2. Average daily increased milk production 3.6 kg I 2.8 kg 2.9 kg 3. Total additional production due to bST (line 1 x 2) 880.6 kg 642.6 690.2 kg 4. Gross milk price per 100 kg $28.09 $28.09 $28.09 5. Total added revenue from bST (line 3 / 100 x line 4) $247.36 $180.51 $193.88 B. REDUCED COSTS None identified in this analysis $0.0 $0.0 $0.0 C. SUBTOTAL (line A + line B) $247.36 $180.51 $193.88 1). REDUCED INCOME None identified in this analysis $0.0 _ $0.0 $0.0 E. INCREASED COSTS 6. Number of bST doses. 17 17 17 7. Cost per dose of bST I $5.86 $5.86 $5.86 8. Total bST product cost $99.62 $99.62 $99.62 9. Time required to inject each cow per dose 1 min. 1 min. 1 min. 10. labor cost per hour $10.00 $10.00 $10.00 11. Total labor cost (line 6x line 9 / 60 x line 10) $2.83 $2.83 $2.83 12. Additional Meal. required‘ 608 Meal 444 Meal 477 Meal 13. Feed cost per Meal $0.084 $0.084 14. Total additional feed cost (line 12 x line 13) $51.11 $37.30 $40.06 15. Other added costs” $10.66 $7.78 $8.53 16. Total additonal costs (line 8 +line ll +lines 14 & 15) $164.22 $147.53 $150.87 F. SUBTOTAL (line C + line D) $164.22 $147.53 $150.87 NET GAIN (line c - line F) 1 $78.12 $37.99 $43.01 ' (0.691 Meal / kg of milk) " (milk marketing = $1.21 per 100 kg) 147 When break-even milk production was calculated, a response of 2.3 kg per bST- treated ’cow per day was necessary to cover additional costs incurred from bST use (bST product, administration and additional feed costs) (Table 7.5). Therefore, any greater production response would be profitable when only these direct costs were attributed to bST use. For the observed level of production response (2.9 kg per bST-treated cow per day), a break-even milk price of $21.86 / 100 kg was calculated. Because of the greater production response in primiparous cows compared to multiparous cows, in these cows bST use was profitable at approximately $2.89 / 100 kg below the all-cow break-even price. TABLE 7.5: Break-even analysis for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co. , St. Louis, MO) at 14-day intervals beginning at approximately 63 days of lactation until 301 days of lactation compared to untreated control cows assuming costs of $0.084 per Meal for additional feed, $5.86 per'dose' of bST and a $28.09 per 100 kg milk price (per bST-treated cow). Additional milk Break—even Break-even needed to cover increase in daily milk price per costs (kg) milk yield (kg) 100 kg (S) Farm 1 603.4 2.5 18.26 Farm 2 487.1 2.1 27.38 Farm 3 576.6 2.4 18.90 Farm 4 477.7 2.0 29.67 lact = 1 584.6 2.5 18.65 Lact > 1 525.2 2.2 22.96 Total 537.1 2.3 21.86 148 Sensitivity analysis found milk price and bST production response both had profound effects on the change in NFI per bST-treated cow (Table 7.6). A small (5%) increase in production response increased NFI by 14.5% but doubling this value (10% increased response) increased NFI by 25.3%, which was less than the expected 29.0% increase. When milk price dropped by 10% , the change in N FI decreased dramatically (by 45.1%) with a proportional decrease when milk price fell by 20%. Compared to production response and milk price, profitability was only slightly sensitive to the price of bST and feed. A decrease in bST price of 5 % increased the change in NFI by only 10.3 % while a 5 % difference in feed price of altered the change in NFI by just slightly over 4%. The relative financial impact of a change in bST response or price, feed cost or milk price was accessed by dividing the percent change in the dependent variable (change in NFI) by the percent change in the independent variable (change in bST response, price or feed cost). Change in milk price was found to have a proportional effect on NFI (N Fl fell by a similar amount when milk price was decreased by 10% to 20%) whereas bST production response and product price did not pr0portionally affect the change in NFI (the relative affect on the change in N F1 was not as great when response or price was changed from 5 % to 10%). 149 TABLE 7.6: Results of sensitivity analysis for several variables for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, M0) at 14-day intervals beginning at approximately 63 days of lactation until 301 days of lactation compared to untreated control cows assuming $0.084 per Meal for additional feed and $5.86 per dose of bST. Change in assumptions Difference in NFI in Percent change in the $ per cow dependent variable / (percentage change percent change in the in NFI from base independent variable assumption) Base assumption: 2.9 kg $43.01 (0.0%) 0.0 increased production per bST-treated cow per day 5% increase in bST $50.28 (14.5 %) 2.9 production response 10% increase in bST $57.56 (25.3%) 3.4 production response 10% decrease in milk $23.62 (- 45.1%) (- 4.5) price received 20% decrease in milk $4.22 (- 90.2 %) (- 4.5) price received 5% decrease in bST $47.94 (10.3%) 2.1 product price 10% decrease in bST $52.87 (18.7%) 1.9 product price 5% decrease in feed $45.01 . (4.4%) 0.9 price 5% increase in feed $41.01 (- 4.7%) (- 0.9) price 150 In these data, cows bearing twins incurred a loss of $92.83 per case compared to non-twinning herdmates (Table 7.7). This value was then discounted by 10% to arrive at a final cost estimate of $84.39 per case of twins. Retained placentas that were associated with twin births (P = 0.001) as well as cases of metritis (P = 0.02) were considered a cost of bST use. However, because these conditions were not significantly associated with other costs (i.e. , increased days open and decreased milk production), only costs resulting from therapy administered for these two diseases were included. Days open to conception was increased by 39 days (P = 0.06) in cows with twin calvings and was included as a cost of twinning. Twinning cows also produced 63.1 kg less milk than cows bearing single calves during the study lactation but this production loss was not significant (P = 0.92) and therefore not considered a direct cost of twinning. In this study, cows experiencing a nonspecific off feed condition had no lost income because little therapy was necessary and no significant decrease in milk production occurred. 151 TABLE 7.7: Financial analysis of twin calvings for cows supplemented with 500 mg of bST (Posilac“, Monsanto Co., St. Louis, M0) at 14—day intervals beginning at approximately 63 days of lactation until the end of lactation compared to untreated control cows. no significant decrease in milk production occurred. Associated Change associated with twins Cost of condition condition Calf value Increased calf value - decreased calf $30.56 survival (14.3% reduced survival ) Retained 18. 4% increase X $20. 00 / case (cost ($3.68) placenta of therapy only) Metritis 16.8% increase X $20.00 / case (cost ($3.36) of therapy only) Days open 39 d increase ($1.25 / (I spent open) ($48.75) Culling 25.3% increase X (dairy value + cull ($67.60) value) / 2 - cull value Total cost Calf value - total costs per case of per case of twins per cow ($92.83) twins aDiscounted Total cost per case discounted by cost per 10% for one period ($84.39) case Present value = future value / (1 + discount rate) "°' “W "°"°"‘ ' These estimated costs were incurred during the subsequent lactation (t = 2); therefore, they have been discounted by one period at a nominal interest rate. DISCUSSION Field trials are susceptible to an underreporting bias because of potential deficiencies in the recording of animal health events. Previous researchers have identified this source of error and the variation in compliance regarding health data recording that exists among farms (Bartlett et al., 152 1992). In the present study, all herds had disease- recording systems in place at the start of the trial that assisted in the accurate recording of disease events, subsequent treatments and periods of discarded milk. Twenty-two cows that were originally selected for the trial were in poor health at the time when treatment was to begin and therefore, were excluded from the analysis. This exclusion was in accordance with the product label that states cows must be determined healthy if bST treatment is to begin during the ninth week of lactation (Monsanto Company, 1993). Because less than 5 % of the cows selected for the trial were excluded, and that bST-treated and control cows were excluded based on the same criteria, it is unlikely that a sizable bias was introduced from this omission. In this study, pre-existing IMI was balanced between bST-treated and control cows (Judge et al., 1997). This is important because cows with IMI have been observed to produce less milk than cows without infection (Bartlett et al. , 1990; Deluyker et al. , 1993). Milk production in the pretrial period was well balanced between the study treatment groups. However, for the one farm (farm three) milk production was marginally different between the study groups, with the control cows producing 192.3 kg more milk in the pretrial period compared to the bST-treated cows. However, for all study cows milk production during the pretrial period differed by only 46.9 kg between the study groups. Reproductive efficiency has been cited as an important determinant of dairy farm profitability (Hady et al. , 1994). Because no differences existed between the study treatment groups for any reproductive parameter (average days to first service, average days open, calving interval, conception rate (first service and all services), heat detection rate, pregnancy rate and services per conception), no cost for reduced reproductive 153 efficiency was necessary in our partial budgets (Chapter 4). The difference in milk production between study treatment groups was not calculated and displayed for each study farm (Table 7.2). This omission was necessary because of a bias in these differences caused by unequal numbers of cows in each study group. The difference in production calculated within each farm represents a unweighted (biased) mean difference compared to the overall (across all four farms) mean (weighted). The overall (weighted) mean best estimates the difference in production between the study groups and therefore was the only difference statistically evaluated. The milk production response observed in this study was slightly less (8.5% vs. 12% daily production response) than observed in other studies (Bauman et al. , 1989; Huber et al., 1997; Remond et al., 1991). It has been suggested that actual response observed by commercial dairies may be up to 25 % less than was seen during experimental situations (Fallert et al., 1987). The reasons for this difference may include: 1) overcrowding of cows causing DMI to be less than maximal, 2) harsh environmental conditions (e.g., extreme heat causing reduced DMI), and 3) differences in rates of diseases that may limited milk production. I observed an approximately 22% , 23 % and 10% greater production response to bST in primiparous compared to multiparous cows during the study trial, post-trial and entire study (trial andpost-trial periods combined) periods, respectively. This result is in contrast to the increase in milk yield observed in other studies where multiparous cows responded to bST treatment more favorably than (Bauman et al. , 1989; Remond et al. , 1991; Skarda et al., 1992) or similar to (Bauman et al., 1989; Remond et al., 1991) 154 primiparous cows. However, another study observed that multiparous cows milked 3X responded less to bST treatment than those milked 2X when compared to similarly treated primiparous cows (production response of 5.2% for 3X compared with 11.4% for 2X milked multiparous cows (54% less response) versus a 10.2% production response for 3X and 16.1% for 2X milked primiparous cows (37 % less response» (Speicher et al. , 1994). These authors suggested that multiparous cows may not respond to bST as well when milked 3X compared to 2X because they already may have been producing near their genetic potential prior to the initiation of bST (Speicher et al. , 1994). .Also, in this study cows of differing parities responded similarly to increased milking frequency alone while other studies observed that primiparous cows responded more favorably to 3X milking than did multiparous cows (Allen et al. , 1986; Amos et al. , 1985). If cows in the present study had been milked 2X and responded similarly, a 4.9 kg / day (3.6 kg / day x 37% greater response) and 4.3 kg / day (2.8 kg / day x 54% greater response) response to bST during the trial period for primiparous and multiparous cows might have been observed, respectively. It should be noted that this estimate is based upon a study where a daily bST treatment regimen was used compared to the present study which utilized a 14-day sustained release preparation (Speicher et al. , 1994). The increase in N FI observed in the present study agrees with an estimate made in a preapproval study (Elbehri and Yonkers, 1995). However, because bST-treated cows milked 3X may respond proportionally less than those milked 2X, the change in NFI that was observed is possibly lower than would be expected if cows were milked 2X. If cows in this study had been milked 2X, the NFI per bST-treated cow would be estimated at 155 $143.33 and $113.23 for primiparous and multiparous cows, respectively, for the trial period if cows responded to bST at comparable levels to those observed in another study (Speicher et al., 1994). This estimate closely agrees with the change in N FI attributable to bST calculated by Tauer and Knoblauch of $120.00 per cow per bST-treated lactation' (T auer and Knoblauch, 1997). Therefore, while the results of the present study may accurately depict the expected increase in NFI for farms that milk cows 3X, it may underestimate economic returns that would be expected for farms that milk cows 2X. No returns to management or opportunity costs were included in these partial budgets. Discounting was determined to be necessary for cases of twinning because the financial loss resulting from twinning occurs during the subsequent lactation. For fmancial' estimates of increased NFI resulting from bST use, no discounting was considered necessary because only a short time lag exists between the investment in bST and the payoff (Hady et al., 1994). Care was taken to use actual, farm and time period—specific (1994 - 1995) data in our partial budgets because the values of inputs and outputs vary among farms and over time (Lloyd and Hady, 1994; Lloyd et al. , 1987). In regard to bST cost, the base price of the commercial product ($5.86 per dose) was used in these calculations. However, most herds using bST on a regular basis participate in a “subscription” program offered by the Monsanto Co. which reduces the price per dose by as much as $0.55. This serves to make bST profitable at approximately 0.14 kg less production response per bST-treated cow per day and increased NFI by $9.35 per bST- treated cow for the trial period. In these partial budgets, increased DMI was assumed to occur immediately after bST initiation. However, it has been found that feed intake does 156 not increase for several weeks after bST initiation (Bauman et al. , 1989). Justification for including increased feed cost during the early bST treatment period is that bST-treated cows have been observed to have greater energy intakes during late lactation compared to control cows (Bauman et al. , 1989). This serves to replenish body reserves and offsets the lag in increased feed intake observed during the initial period of bST treatment. The Monsanto Company has released a computer program designed to assist producers in calculating financial gains in their herds resulting from bST use (Profit Planner“) (Monsanto Company, 1996). In this study, approximately 2.3 kg of production response was necessary to offset the increased cost incurred by bST use. This closely agrees with the results of a preapproval analysis (Elbehri and Yonkers, 1995). At this level of response, the milk price received must minimally be $21.86 / 100 kg in order to break-even using bST. The price of wholesale milk has not been this low for nearly 20 years (National Milk Producers Federation, 1997). Because the USDA has set a floor milk price of $21.78 / 100 kg, the price of milk currently cannot fall significantly below the break-even milk price calculated for the observed level of bST production response in this study. If producers received the support (floor) milk price, the production response observed during the trial period in our study would approximate the break-even production response for milk produced with bST at the USDA floor price. In this study, milk was priced at $28.09 / 100 kg, which was the average all-milk wholesale price for 3.5 % fat milk produced during 1994 and 1995 (National Milk Producers Federation, 1997). However, the study farms likely received a higher actual 157 (mailbox) milk price when premiums for fat and protein content, and bacteria or SCC counts were added to the Basic Formula Price. Therefore, the changes in N F1 reported here are likely to slightly underestimate actual increases realized by these study herds. For example, if a farm received a $0.10 / 100 kg premium for fat, N FI would be expected to increase by approximately 1.6% (from $43.01 to $43.70) from the estimate made for bST in this study. A large variation in NFI change attributable to bST was observed among the four herds in this study. This was primarily a result of differences in milk production response between the study herds because feed, bST and labor costs were not greatly different between the study herds. A large variation in production response to bST among herds has been observed by others (Thomas et al. , 1991). Because of the large effect production response has on the profitability of bST use, and because bST production response varies greatly among herds, the increase in N F1 from bST use would be expected to vary widely between herds. For a given input, the profitability of the investment depends on the relative prices of the variable inputs necessary to enact the change (Marsh et al. , 1988). In this study, the price paid for bST and feed did not have a large effect on the profitability of bST use. Based on the current pricing scheme for bST, the difference between the highest and lowest prices paid for bST only vary by about 9.5 %. Therefore, any change in price paid for bST will have only a modest impact on profitability. Because cows, compared to their total intake, experience only a small increase in DMI in response to bST (observed DMI increases approximately 4 to 6%) (Bauman et al. , 1989), changes in feed price have a 158 relatively minor impact on bST profitability. Nevertheless, it has been stated that bST use . will be least profitable when milk prices are low and feed prices are high (Schmidt, 1989). Overall, the profitability of bST use depends primarily on the level of milk production response and ultimately on overall milk yield per cow (Elbehri and Yonkers, 1995). Because of its major impact on farm profitability, the importance of frequent monitoring of milk yield per cow cannot be overemphasized. It is acknowledged that our study considers only a single lactation. It is possible that additional costs or revenues could be realized in subsequent lactations that could impact the economics of bST use and were not included this analysis (Marsh et al. , 1988). However, culling during the study lactation and milk production during the first 63 days of the next lactation were analyzed and found to be similar between bST-treated and control cows. Therefore, while it is unlikely that any significant costs or revenues would be carried over into subsequent lactations, the costs of inputs required for bST use (bST product, labor and feed) as well as the price of outputs (milk) may change over time and significantly effect the economics of bST use. It has been observed that the most common price fluctuation affecting dairy farms is feed price changing relative to the price of milk (Congleton et a1. , 1984). Based on the results of this study, bST-assisted milk production would be most profitable when the price received for milk is high, but an increase in the price paid for feed would not dramatically reduced bST profitability. In this study, cows experiencing a nonspecific off feed condition had no lost income because little therapy was necessary and no significant decrease in milk production occurred. However, a case of twins cost $84.39 after discounting the cost by 10% , which 159 was similar to the losses calculated by other workers (Beerepoot et al., 1991; Eddy et al. , 1991). Cows bearing twins produce more total calf value, but this was offset by increased calf mortality, greater rates of diseases (retained placenta and metritis), more days spent open and higher culling rates compared to non-twinning herdmates. In this study, bST-treated cows were more than twice as likely to conceive twins than were control cows (relative risk = 2.13). When this increased risk was applied to the incidence of twins observed prior to bST administration (6.0% lactational incidence) and the cost of a case of twins in this data, N FI resulting from bST use was decreased by $0.78 (1.8% of the expected benefit of bST) per bST-treated cow per lactation. - In the four herds that were studied, bST was not associated with an increase in clinical mastitis. Although lameness was not monitored, most effects of bST on lameness were evaluated indirectly through comparisons of culling and milk production. However, a meta-analysis conducted, in Canada combined the weighted means of many bST studies performed in several herds and concluded that bST increased the risk of clinical mastitis by 25% and lameness by 50% (Canadian Veterinary Medical Association Expert Panel, 1998). Clearly, herds in which bST may be associated with economic losses because of these diseases need to consider such losses as additional expenses when estimating the change in NFI from bST use in their herds. CONCLUSIONS Profitability of bST was quite variable between farms studied because many factors influence the change in NFI per cow that occurs as a result of bST use, most notably, 160 production response to bST. Although our study only included four farms, similar variation would be expected to occur in the larger population of dairy farms. However, the amount of variation between farms can not be accurately estimated from these data because of the small number of farms utilized in this analysis. Among farms studied, the change in NFI observed varied from a slight loss to an increase of approximately $90.00 per bST-treated cow. The level of production response and the price received for milk had the largest effects on the change in NFI. By contrast, prices paid for bST and feed price had only minimal effects on bST profitability. Therefore, dairy producers would be advised that bST use will be most profitable when the price received for milk is high. Also, the economics of bST use would not be expected to change greatly when either feed or bST prices vary moderately. This analysis emphasizes the importance of monitoring production response given its major effect on the change in N FI resulting from bST use. In the herds we studied were analyzed together, bST use was profitable when production response exceeded 2.3 kg per bST-treated cow per day and milk price received was above the USDA support price. Diseases that may be associated with bST may reduce the profitability of this product and need to be considered as a cost of bST use if present. 161 REFERENCES Allen, D. B., E. J. DePeters, and R. C. Labin. 1986. Three times a day milking: effects on milk production, reproductive efficiency and udder health. J. Dairy Sci. 69:1441. Amos, H. E., T. Kiser, and M. Loewenstein. 1985. Influence of milk frequency on productive and reproductive efficiencies of dairy cows. J. Dairy Sci. 68:732. Bartlett, P. C., G. Y. Miller, C. R. Anderson, and J. H. Kirk. 1990. Milk production and somatic cell count in Michigan dairy herds. J. Dairy Sci. 73:2794-2800. Bartlett, P. C., G. Y. Miller, S. E. lance, and L. E. Heider. 1992. Clinical mastitis and intramammary infections on Ohio dairy farms. Prev. Vet. Med. 12:59-71. Bartlett, P. C., J. Van Wijk. 1991. Temporal patterns of lost milk production following clinical mastitis in a large Michigan Holstein herd. J. Dairy Sci. 74: 1561-1572. Bauman, D. E. 1992. Bovine Somatotropin: Review of an emerging animal technology. J. Dairy Sci. 75:3432-3451. Bauman, D. E., D. L. Hard, B. A. Crooker, M. S. Partridge, K. Garrick, L. D. Sandles, H. N. Erb, S. E. Franson, G. F. Hartnell, and R. L. Hintz. 1989. Long-term evaluation of a prolonged-release formulation of N-methionyl bovine somatotropin in lactating dairy cows. J. Dairy Sci. 72:642-651. Beerepoot, G. M. M., A. A. Dykhuizen, M. Nielen, and Y. H. Schukken. 1991. The economics of naturally occurring twinning in dairy cattle. J. Dairy Sci. 75:1044- 1051. Brown, M. 1., J. W. Lloyd, J. B. Kaneene, and J. Staatz. 1993. Theoretical financial benefit of internal parasite control for Michigan dairy farms using National Animal Health Monitoring System data. Prev. Vet. Med. 17:47-56. Canadian Veterinary Medical Association Expert Panel. 1998. Report of the Canadian Veterinary Medical Association Expert Panel on rBST. Ottawa, Canada. Centers for Epidemiology and Animal Health. 1996. National Animal Health Monitoring System, Dairy ‘96. Fort Collins, CO. 162 Congleton, W. R. , Jr. and L. W. King. 1984. Profitability of dairy cow herd life. J. Dairy Sci. 67:661-674. Dairy Comp 305“, Version 3.1. 1990. Valley Agricultural Software. Tulare CA. Deluyker, J. M., and L. D. Weaver. 1993. Interrelationships of somatic cell count, mastitis, and milk yield in a low somatic cell count herd. J. Dairy Sci. 76:3445- 3452. Dickrell, J. Where’s the Milk? 1999. Dairy Today. 2:16-17. Eddy, R. G., 0. Davies, and C. David. 1991. An economic assessment of twin births in British dairy herds. Vet. Rec. 129:526-529. Elbehri, A., and R. D. Yonkers. 1995. Economic analysis of the impacts of bovine somatotropin on the profitability of representative dairy farms in the northeast. Ag. and Res. Econ. Rev. 4:88-100. Excel“, Version 5.0. 1993. Microsoft Corporation, Redmond, WA. Enting, H., D. Kooij, A. A. Dijkhuizen, R. B. M. Huirne, and E N. Noordhuizen- Stassen. 1997. Economic losses due to clinical lameness in dairy cattle. Livest. Prod. Sci. 49:259-267. Fallert, R. F. , T. McGuckin, C. Betts, and G. Bruner. 1987. Bovine somatotropin and the US dairy industry: a national, regional and farm-level analysis. USDA Econ. Res. Serv., Agric. Econ. Rep. 579. USDA, Washington, DC. Farm Financial Standards Council. 1997. Financial guidelines for agricultural producers. Farm Financial Standards Council, Naperville, IL. Hady, P. J ., J. W. Lloyd, J. B. Kaneene, and A. L. Skidmore. 1994. Partial budget model for reproductive programs of dairy farm businesses. J. Dairy Sci. 77:482-491. Harsh, S. B., L. J. Connor, and G. D. Schwab. 1981. Managing the farm business. Prentice-Hall, Inc., Englewood Cliffs, NJ. Hoblet, K. H., G. D. Schnitkey, D. Arbaugh, J. S. Hogan, K. L. Smith, P. S. Schoenberger, D. A. Todhunter, W. D. Hueston, D. E. Pritchard, G. L. Bowman, L. E. Heider, B. L. Brockett, and H. R. Conrad. 1991. Costs associated with selected preventive practices and with episodes of clinical mastitis in nine herds with low somatic cell counts. JAVMA. 199(2):190—196. 163 Huber, J. T., Z. Wu, C. Fontes, Jr., J. L. Sullivan, R. G. Hoffman, and G. F. Hartnell. 1997. Administration of recombinant bovine somatotropin to dairy cows for four consecutive lactations. J. Dairy Sci. 80:2355-2360. Judge, L. J ., P. C. Bartlett, J. W. Lloyd and R. J. Erskine. 1999. Recombinant bovine somatotropin: association with reproductive performance in dairy cows. Therio. 52:481-496. Judge, L. J ., R. J. Erskine and P. C. Bartlett. 1997. Recombinant bovine somatotropin and clinical mastitis: incidence, discarded milk following therapy, and culling. J. Dairy Sci. 80:3212-3218. Lloyd, J. W., and P. J. Hady. 1994. Partial Budgeting as a dairy consultation tool. The Bovine Pract. 26:84-87. Lloyd, J. W., J. B. Kaneene, and S. B. Harsh. 1987. Toward responsible farm-level economic analysis. JAVMA. 191(2):195-199. Marion, B. W. , and R. L. Wills. 1990. A prospective assessment of the impacts of bovine somatotropin: a case study of Wisconsin. Am. J. Agr. Econ. 5:326-336. Marsh, W. E. , D. T. Galligan, and W. Chalupa. 1988. Economics of recombinant bovine somatotropin us in individual dairy herds. J. Dairy Sci. 71:2944-2958. Miller, G. Y., P. C. Bartlett, S. E. lance, J. Anderson, and L. E. Heider. 1993. Cost of clinical mastitis and mastitis prevention in dairy herds. JAVMA. 202(8):1230- 1236. National Milk Producers Federation. 1997. Prices received by farmers for 3.5% milk, 1920-1995. Arlington, VA. National Research Council. 1988. Nutrient requirements of dairy cattle. Sixth rev. ed. Natl. Acad. Press., Washington, DC. Posilac“ product label. 1993. Monsanto Co., St. Louis, MO. Posilac“ Profit Planner, Version 1.2.4. 1996. Protiva, a unit of Monsanto Co. , St. Louis, MO. Remond, B., M. Cisse, A. Ollier, and Y. Chilliard. 1991. Slow release somatotropin in dairy heifers and cows fed two levels of energy concentrate. 1. Performance and body condition. J. Dairy Sci. 74:1370-1381. 164 SAS System for Windows“, Version 6.10, 1994. SAS Institute, Cary, NC. Schmidt, G. H.. 1989. Economics of using bovine somatotropin in dairy cows and potential impact on the US dairy industry. J. Dairy Sci. 72:737-745. Shoeffling, J. R., R. C. Angus, D. V. Armstrong, and J. T. Huber. 1991. Economic implications of bovine somatotropin use for the Arizona dairy industry. J. Dairy Sci. 74:2347-2352. Skarda, J ., E. Markalous, J. Slaba, P. Krejci, O. Skardova, and J. Zednik. 1992. Effect of methionyl bovine somatotropin in a prolonged-release vehicle on milk production, hormone profiles and health in dairy cattle. J. Dairy Res. 59:499-506. Spartan Dairy Ration Evaluator/Balancer, Version 2.01. 1992. Cooperative Extension Service and the Agricultural Experiment Station. Michigan State University, E. Lansing, MI. Speicher, J. A., H. A. Tucker, R. W. Ashley, E. P. Stanisiewski, J. F. Boucher, and C. J. Sniffen. 1994. Production responses of cows to recombinantly derived bovine somatotropin and to frequency of milking. J. Dairy Sci. 77:2509-2517. Tauer, L. W. , and W. A. Knoblauch. 1997 . The empirical impact of bovine somatotrOpin on New York dairy farms. J. Dairy Sci. 80:1092-1097. Thomas, J. W., R. A. Erdman, D. M. Galton, R. C. lamb, M. J. Arambel, J. D. Olson, K. S. Madsen, W. A. Samuels, C. J. Peel, and G. A. Green. 1991. Responses by lactating cows in commercial dairy herds to recombinant bovine somatotropin. J. Dairy Sci. 74:945-964. United States Department of Agriculture, National Agricultural Statistical Service. 1997. Final Estimates, 1993-97. Washington, DC. Wells, S. J ., A. M. Trent, R. S. Collier, and W. J. Cole. 1995. Effect of long-terrn administration of a prolonged release formulation of bovine somatotropin (sometribove) on clinical lameness in dairy cows. Am. J Vet. Res. 56(8):992-996. Willett, G. S., R. Blauwiekel, D. C. Grusenmeyer, and H. R. Hinman. A worksheet for analyzing the economic of bovine somatotropin adoption. 1994. Wash. State Univ. Coop. Ext. Serv., Ext. Bulletin 1777. Pullman, Wash. 165 SUMMARY While likely not the final word on bST-related knowledge, some important findings did come to light as a result of conducting this study. First and foremost, the lack of effect of bST on the incidence of clinical mastitis and other common diseases of dairy cows was reinforced. With the exception of nonspecific off feed episodes, no other studied diseases were found to be associated with bST use in this study. These off feed episodes were not associated with significant economic losses and therefore only exerted a slight “nuisance” effect on these study. farms. Also, no effect of bST was observed for any reproductive indices evaluated in this analysis. However, the sample size of this study limited the statistical power for the analysis of most diseases and reproductive parameters. Power ranged from 0.05 to 0.48 for these calculations, with higher values found for disease incidence compared to reproductive parameters. Nevertheless, this study likely would have detected any strong effect associated with bST use on reproductive parameters had one been present in the study population. Twin conceptions that occurred during the study lactation were significantly greater in cows treated with bST. While this increased risk was found in this study, the FDA allowed the removal of a cautionary statement to this effect from the label of the available bST product. Their decisions was based on data provided by The Monsanto Company from the Post-approval Monitoring Project. In this study, this effect on twinning resulted in a slight reduction in the estimated profitability of bST. Because twin calvings may 166 increase as milk production rises, it is likely that any production enhancement technology will be associated with increased rates of twinning, independent of any effect such products may have on the reproductive physiology of cows. In this study, bST was not found to be associated with any change in culling rates during the period of this study (one lactation plus the first 60 days of the subsequent lactation). However, because cows were only followed for approximately 1.2 lactations, the effect of long term treatment with bST (multiple lactations) could not be determined from this data. It was observed that culling for reason of low milk production was significantly less in bST-treated cows. This would be expected because of the enhanced production caused by bST results in herd managers prolonging the retention of these cows in the herd. It is likely that increased herd longevity serves to increase the profitability of a cows lifetime through increased milk production and decreased demand for replacement animals. In these data, it was observed that commercially-available bST was associated with a decreased accuracy of DHIA lactation milk production projections for a period of approximately 100 days after the commencement of treatment (i.e. , the first three monthly test periods). As such, milk yield estimates made by DHIA significantly underestimated production in bST-treated cows until the time when the third administration of bST was made during the study lactation. The significant inaccuracy observed during this period has implications for dairy herd managers and consults who use DHIA estimates for making decisions (e.g. , feed requirements and culling decisions) for cows and herds that are treated with bST. 167 Finally, treatment with bST in this study increased the estimated profitability of cows by $42.23 per bST treated cow after subtracting a cost of $0.78 per bST treated cow because of the increased incidence of twin calvings that were associated with bST use. While this estimate is somewhat less than calculated by other workers, it still demonstrates the ability of bST to enhance the profitability of cows. There was tremendous variability observed between the four study herds that participated in this project. This was because production response varied greatly and by sensitivity analysis was found to be one of the largest effects influencing the change in profitability resulting from bST use. 168 CONCLUSIONS The success of any scientific study should be based on its ability to resolve the questions posed at the time when the project was designed. For this study, the answers to five study objectives were sought: The first objective of this study was to determine if bST was associated with the clinical mastitis, both in terms of incidence and days of production withheld from market resulting from therapy of clinical cases. In this study, incidence was similar between the study treatment groups on all but one farm (resulting from the reoccurrence of chronic infections). A non-significant trend of less days of production withheld from market because of therapy for. clinical mastitis occurring in bST-treated cows. This would support the work of Burvenich et al. (1993) who observed that cows treated with bST and then experimentally infected with E. coli (to produced clinical mastitis) recovered to milk production significantly faster than those not treated with bST. Further work is needed to determine if this effect is repeatable in cows with naturally occurring infections. The second objective was to investigate the association of bST and reproductive function of cows, including any association with diseases that may affect reproduction. Aside from being associated with more nonspecific off feed conditions during the study lactation and greater twin births at the subsequent calving, bST did not significantly alter reproductive performance of cows compared to controls. However, because of the difficulty in detecting small differences in rare disease rates and reproductive parameters, future studies will need to utilize more cows than were included in this trial. As dairy 169 farms grow larger and the availability of computerized records of disease and reproductive data increases, such studies will become more feasible. Still, the recording of such data by study farms must be monitored carefully to ensure. that significant reporting bias does not lead to spurious results and conclusions. Completion of the third study objective confirms the work of Ruegg et a1 (1998) regarding the effect of bST on culling of cows. No effect of bST was observed on culling or death rates, but the amount of culling for reason of low milk production was found to be less for cows treated with bST. Much has been written in the past on this topic, often blaming bST for causing cows to “burn out” (i.e. , develop debilitating disease that resulted in premature culling or death). It would seem, based on the results of these two studies, representing 36 Midwestern herds and over 6,000 cows, that these fears have little basis. In fact, the opposite effect of bST was observed in this study because fewer bST-treated cows were culled for reason of low milk production compared to controls and culling for other reasons was not significantly different between the study treatment groups. The fourth objective analysis was the most novel and important finding of this work. A negative association was observed between bST and the accuracy of DHIA lactation milk estimates in this study. From a review of the literature, it appears that this was the first study to examine this relationship. Based on the results of this study, dairy herd managers should delay culling decisions based on DHIA lactation milk predictions until after receiving the third estimated lactation production projection following bST treatment initiation. This advice is based on the observed result that cows treated with bST will produce nearly 1,000 pounds more milk by 305 days of lactation than is predicted by 170 DHIA at the first milk test after treatment is initiated. Because DHIA recommends the culling of cows who are predicted to produce 2,000 pounds of milk less than herdmates, this inaccuracy could impact the decisions of herd managers based. As a result of this study, work is needed to recalculate lactation extension factors used by DHIA for cows that are treated with bST, primarily during the three-month period following the initiation of treatment. Finally, completion of the fifth study objective confirmed the ability of bST to increase the profitability of dairy cows: However, tremendous variability was observed between the four farms studied, with two study farms enjoying approximately $80.00 of increased NFI per bST treated cow while the other two study herds only had enough production response from bST use to just cover costs associated with the use of the product. It was determined that the increase in profitability from bST use was most sensitive to changes in production response to bST and price received for milk while being less sensitive to changes in the prices paid for feed and bST itself. However, future studies are need to examine the factors that are associated with a herds ability to respond to bST treatment because of its importance in determining the profitability of the products use. Such studies are need to determine what factors are associated with the great variability in production response that was observed between herds in this trial and additionally how the identified factors can be managed to increase the production response resulting from use of bST. Overall, this study adequately addressed the questions posed and concerns raised by the investigators. The major shortcoming of this work was lack of adequate sample size 171 to give great statistical power to some of the observed results. As has been previously mentioned, studies that could address this concern are feasible and will be able to utilize 4 sufficient numbers of herds and cows to increase the power of future analyses. While studies hopefully answer questions, they often raise even more questions through their conduct. As a result of the conduct of this study and a review of the results of the analyses performed, the following are areas where further study is needed. First, field studies of the effect of bST on the severity of clinical diseases (especially clinical mastitis). While the incidence of disease was not greatly affected by bST, it is possible that the severity (as measured by duration of illness and amount of milk production lost) may be impacted by use of this product. The method of Bartlett and Van Wijk et al (1994) could be employed to estimate the lost production resulting from various diseases (especially clinical mastitis). Because lost production (either milk discarded because of antibiotic therapy or simply never produced) accounts for nearly 90% of the economic loss resulting from clinical mastitis (Bartlett et al., 1990), such an effect, if found, might significantly impact the profitability of bST use. Second, study is needed of potential differences in production response from bST treatment between cows of differing parities and milked at varying frequencies. While it was not possible. to ascertain the effect of milking frequency from the present study because all cows were milked three times per day, it was possible to determine difference in response between prirni- and multiparous cows. In this study, primiparous cows had greater response to bST treatment compared to multiparous cows. This result has been previously, but infrequently observed. However, independent of bST treatment, is has been found that primiparous cows 172 responded to increased milking frequency at greater levels than did multiparous cows (Amos et al., 1985; Allen et al., 1986). Further work is needed to determine both the influence of milking frequency and its interaction with parity in regard to bST production response. Thirdly, studies will be needed in the future when other production enhancement products become available (approved by the FDA) for use in lactating dairy cows. It is likely that such products will become commonplace in the future and each will need to be evaluated in light of potential interaction (either synergy or blockage of effect) that may occur. While the FDA throughly evaluates each new product for both safety and efficacy, interaction of the proposed new product with existing products already in use is not a factor that is either studied nor necessary for approval. This task will be reserved for other workers after the product has passed the necessary testing that is part of the FDA’s approval process. On the immediate horizon is one such product: Rumensin“ (monensin tartrate from Elanco Animal Health, Greenfield, IN) will likely by the first product of this type to be approved by the FDA for use in lactating dairy cows. It will be necessary to conduct field trials after this product is granted FDA approval to determine what interaction, if any, exists been monensin and bST. Because of the improvement in energy conversion resulting from monensin treatment, it is possible that this effect may alter production response from bST. This is because as a repartitioning agent bST, directs more nutrients to the mammary gland (thereby enhancing milk production) and theoretically, the availability of more nutrients to be diverted to the mammary gland may support the production of even more milk than occurs when cows are treated with either monensin or bST alone. Reports of data derived from herds located in Mexico tend to 173 support this proposed mechanism. In conclusion, while the conduct of this study increased knowledge of the effect that bST has on dairy cows and farms, more questions were raised that need to be addressed by future work. It is only through the conduct of research that inference is gained that helps us understand the world we live in and discern what is real from that which we wish were “real.” 174 APPENDIX A 175 APPENDIX A Definitions of diseases used to diagnosis diseased cows in all study herds. Displaced abomasum: Either left or right displacement of the abomasum as diagnosed by a veterinary practitioner. Cystic ovarian disease: The presence of a follicle > 2.5 cm on the ovary which was diagnosed by veterinary palpation. Twin birth: More than one calf born to a particular cow at calving or the time of abortion. Nonspecific diarrhea: A period of abnormally loose fecal material which was sufficiently severe to cause a noticeable decrease in milk production and/ or feed intake which required treatment. No diagnostic tests were performed on most of these cases. Nonspecific off feed condition: A period of significantly reduced feed intake which was sufficiently severe to cause a noticeable decrease in feed intake and/ or milk production. Abortion: Either observed or unobserved loss of a diagnosed (palpated) pregnancy which occurred at less than 250 days of gestation. Clinical mastitis: The observation of abnormal milk expressed throughout milking in a quarter; the cow was not required to be systemically ill. Milk fever: A cow observed to be ataxic or unable to rise near the time of parturition which required treatment with calcium to regain normal function. Ketosis: A cow observed with reduced feed intake and/ or milk production in the early post- parturm period and confirmed by a positive urine Ketostix“ test (Bayer Corp. , Elkhart, IN). Dystocia: A calving sufficiently difficult to require manual assistance to deliver the calf. Retained placenta: The presence of fetal membranes observed to be retained in the cow greater than 24 h post-calving. Metritis: The presence of foul-smelling uterine discharge in a post-partum cow; the cow was not required to be systemically ill. 176 BIBLIOGRAPHY 177 BIBLIOGRAPHY Allen, D. B., E. J. DePeters, and R. C. Labin. 1986. Three times a day milking: effects on milk production, reproductive efficiency and udder health. J. Dairy Sci. 69:14-41. Amos, H. E., T. Kiser, and M. Loewenstein. 1985. Influence of milk frequency on productive and reproductive efficiencies of dairy cows. J. Dairy Sci. 68:732. Arkins, S., R. Dantzer, and K. W. Kelley. 1993. Somatolactogens, somatomedins, and irnmumity. J. Dairy Sci. 76:2437-2450. Bartlett, P. C., G. Y. Miller, C. R. Anderson, and J. H. Kirk. 1990. Milk production and somatic cell count in Michigan dairy herds. J. Dairy Sci. 73:2794-2800. Bartlett, P. C., G. Y. Miller, S. E. Lance, and L. E. Heider. 1992. Clinical mastitis and intramammary infections on Ohio dairy farms. Prev. Vet. Med. 12:59-71. Bartlett, P. C ., J. Van Wijk. 1991. Temporal patterns of lost milk production following clinical mastitis in a large Michigan Holstein herd. J. Dairy Sci. 74:1561-1572. Bauman, D. E. 1992. Bovine Somatotropin: Review of an emerging animal technology. J. Dairy Sci. 75:3432-3451. Bauman, D. E., R. W. Everett, W. H. Weiland, and R. J. Collier. 1999. Production Responses to Bovine Somatotropin in Northeast Dairy Herds. J. Dairy Sci. 82:2564-2573. Bauman, D. E., D. L. Hard, B. A. Crooker, M. S. Partridge, K. Garrick, L. D. Sandles, H. N. Erb, S. E. Franson, G. F. Hartnell, and R. L. Hintz. 1989. Long-term evaluation of a prolonged-release formulation of N-methionyl bovine somatotropin in lactating dairy cows. J. Dairy Sci. 72:642-651. Beaudeau, F., V. Ducrocq, C. Fourichon, and H. Seegers.1995. Effect of disease on length of productive life of French Holstein dairy cows assessed by survival analysis. J. Dairy Sci. 78:103-117. Beerepoot, G. M. M., A. A. Dykhuizen, M. Nielen, and Y. H. Schukken. 1991. The economics of naturally occurring twinning in dairy cattle. J. Dairy Sci. 75:1044- 1051. 178 Brown, C. A., P. T. Chandler, and J. B. Holter. 1977. Developement of Predictive equations for milk yield and dry matter intake in lactating cows. J. Dairy Sci. 60:1739-1754. Brown, M. I., J. W. Lloyd, J. B. Kaneene, and J. Staatz. 1993. Theoretical financial benefit of internal parasite control for Michigan dairy farms using National Animal Health Monitoring System data. Prev. Vet. Med. 17:47-56. Burton, J. H,G. K. MacLeod, B. W. McBride, J. L. Burton, K. Bateman, I. McMillan, and R. G. Eggert . 1990. Overall efficacy of chronically administered recombinant bovine somatotropin to lactating dairy cows. J Dairy Sci. 73:2157- 2167. Burton, J. L., B. W. McBride, J. H. Burton, and R. G. Effert. 1990. Health and reproductive performance of dairy cows treated for up to two consecutive lactations with bovine somatotropin. J Dairy Sci. 73:3258-3265. Butler, W. R., and R. D. Smith. 1989. Interrelationships between energy balance and postpartum reproductive function in dairy cattle. J Dairy Sci. 72:767-783. Canadian Veterinary Medical Association Expert Panel. 1998. Report of the Canadian Veterinary Medical Association Expert Panel on rBST. Ottawa, Canada. Ceelen, H .J. 1995. Bovine somatotoropin and cow health - what are the facts? Can Vet J. 36:25-27. Centers for Epidemiology and Animal Health. 1996. National Animal Health Monitoring System, Dairy ‘96. Fort Collins, CO. Cole, J. W., P. J. Eppard, B. G. Boysen, K. S. Madsen, R. H. Sorbet, M. A. Miller, R. L. Hintz, W. E. Ribelin, B. G. Hammond, R. J. Collier, and G. M. Lanza. 1992. Response of dairy cows to high doses of a sustained-release bovine somatotropin administered during two lactations. 2. Health and reproduction. J. Dairy Sci. 75:11-123. Cole, W. J ., K. S. Madsen, R. L. Hintz, and R. J. Collier. 1991. Effect of recombinantly- derived bovine somatotropin on reproductive performance of dairy cattle. Theriogenology. 36:573-595. Collet, D.1994. Modelling survival data in medical research. Chapman and Hall. New York. 179 Congleton, W. R., Jr., and R. W. Everett. 1980. Application of the incomplete gamma function to predict cumulative milk production. J. Dairy Sci. 63:109-119. Congleton, W. R. , Jr. and L. W. King. 1984. Profitability of dairy cow herd life. J. Dairy Sci. 67:661-674. Cox, D. R., and D. Oakes. 1984. Analysis of survival data. Chapman and Hall, London. Crooker, B. A.and D. E. Otterby. 1991. Management of the dairy herd treated with bovine somatotropin. Vet. Clin. of North Am. Food Anim. Pract. 7(2):417. Cullor, J. S. 1991. Mastitis in dairy cows: does it hinder reproductive performance? Vet Med. 86: 830-832. Dairy Comp 305“, Version 3.1. Valley Agricultural Software. Tulare CA. 1990. De la Sota, R. L., M. C. Lucy,C. R. Staples, and W. W. Thatcher. 1993. Effects of recombinant bovine somatotropin (sometribove) on ovarian function in lactating and nonlactating dairy cows. J Dairy Sci. 76:1002-1013. Deluyker, H. A., J. M. Gay, L. D. Weaver, A. S. Axari. 1991. Change in milk yield with clinical diseases for a high producing dairy herd. J Dairy Sci. 74:463-445. Deluyker, J. M. , and L. D. Weaver. 1993. Interrelationships of somatic cell count, mastitis, and milk yield in a low somatic cell count herd. J. Dairy Sci. 76:3445- 3452. Dickrell, J. Where’s the Milk? 1999. Dairy Today. 2:16—17. Downer, J. V., D. L. Patterson, D. W. Rock, W. V. Chalupa, R. M. Cleale, J. L. Firkins, G. L. Lynch, J. H. Clark, B. O. Brodie, B. F. Jenny, and R. De Gregorio. 1993. Dose titration of sustained-release recombinant bovine somatotropin in lactating dairy cows. J. Dairy Sci. 76:1125-1136. Eddy, R. G., 0. Davies, and C. David. 1991. An economic assessment of twin births in British dairy herds. Vet. Rec. 129:526-529. Eicker, S. W., Y. ‘T. Grohn, and J. A. Hertl. 1996. The association between cumulative milk yield, days Open, and days to first breeding in New York Holstein cows. J. Dairy Sci. 79:235-241. 180 Elbehri, A., and R. D. Yonkers. 1995. Economic analysis of the impacts of bovine somatotropin on the profitability of representative dairy farms in the northeast. Ag. and Res. Econ. Rev. 4:88—100. Elsasser, T. H., M. Richards, R. Collier, and G. F. Hartnell. 1996. Physiological responses to repeated endotoxin challenge are selectively affected by recombinant bovine somatotropin administration to calves. Dom. An. Endo. 13(1):91-103. Enting, H., D. Kooij, A. A. Dijkhuizen, R. B. M. Huirne, and E N. Noordhuizen- Stassen. 1997. Economic losses due to clinical lameness in dairy cattle. Livest. Prod. Sci. 49:259-267. Epi Info“, Version 6.03, 1996. Centers for Disease Control, Atlanta, GA. Eppard, R. J ., S. Hudson, W. J. Cole, R. L. Hintz, G. F. Hartnell, T. W. Hunter, L. E. Metzger, A. R. Torkelson, B. G. Hammond, R. J. Collier, and G. M. Ianza. 1991. Response of dairy cows to high doses of a sustained-release bovine somatotropin administered during two lactations. 1. Production response. J. Dairy Sci. 74:3807-3821. Erdman, R. A., and M. Varner. 1995. Fixed yield responses to increased milking frequency. J. Dairy Sci. 78:119-1203. Erskine, R. J ., R. J. Eberhart, L. J. Hutchinson, S. B. Spencer, and M. A. Campbell. 1988. Incidence and types of clinical mastitis in dairy herds with high and low somatic cell counts. JAVMA l90(11):1411-l416. Essl, A. 1998. Longevity in dairy cattle breeding: a review. Livest. Prod. Sci. 57:79-89. Esslemont, R. J ., and M. A. Kossaibati. 1997. Culling in 50 dairy herds in England. Vet Rec. 140:36-39. Esteban, E., P. H. Kass, L. D. Weaver, J. D. Rowe, C. A. Holrnberg, C. E. Franti, and H. F. Trout. 1994. Interval from calving to conception in high producing dairy cows treated with recombinant bovine somatotropin. J. Dairy Sci. 77:2549-2561. Esteban, E., P. H., Kass, L. D. Weaver, J. D. Rowe, C. A. Holrnberg, C. E. Franti, and H. F. Trout. 1994. Pregnancy incidence in high producing dairy cows treated with recombinant bovine somatotropin. J Dairy Sci. 77:468-481. Esteban, E., P. H. Kass, L. D. Weaver, J. D. Rowe, C. A. Holrnberg, C. E. Franti, and H. F. Trout. 1994. Reproductive performance in high producing dairy cows treated with recombinant bovine somatotropin. J Dairy Sci. 77:3371-3381. 181 Etherington, W. G., M. L. Kinsel, and W. E. Marsh. 1996. Relationship of production . to reproductive performance in Ontario dairy cows: herd level and individual animal descriptive statistics. Theriogenology. 46:935-959. Excel“, Version 5.0. 1993. Microsoft Corporation, Redmond, WA. Fallert, R. F., T. McGuckin, C. Betts, and G. Bruner. 1987. Bovine somatotropin and the US dairy industry: a national, regional and farm-level analysis. USDA Econ. Res. Serv., Agric. Econ. Rep. 579. USDA, Washington, DC. Farm Financial Standards Council. 1997. Financial guidelines for agricultural producers. Farm Financial Standards Council, Naperville, IL. Ferguson, J. D. 1996. Diet, production, and reproduction in dairy cows. An Feed Sci Tech. 59:173-184. Firat, M. Z. 1993. An investigation into the effects of clinical mastitis on milk yield in dairy cows. Livest. Prod. Sci. 36:311-321. Forar, A. L., J. M. Gay, and D. D. Hancock. 1995. The frequency of endemic fetal loss in dairy cattle: a review. Theriogenology. 43:989-1000. Galligan, D. M. 1997. Extended calving intervals with the use of bST. Protiva Reproductive Management Symposium (in conjunction with the Tri-State Dairy Nutrition Conference). 11-17. Galligan, D. T., J. D. Ferguson, W. Chalupa, and C. F. Ramberg. 1992. Quantifying changes in milk production: don’t let someone throw you a curve. Compend. l4(4):549-553. Gallo, G. F., and E. Block. 1991. Effects of recombinant bovine somatotropin on hypophyseal and ovarian functions of lactating dairy cows. Can. J. Anim. Sci. 71:343-353. Galton, D. M., R. W. Everett, M. E. Van Amburgh, D. L. Bauman, and W. A. Knoblauch. 1997. Extended calving intervals with the use of bST. Presented at the Tri-State Dairy Nutrition Conference, Fort Wayne, IN. Garverick, H. A. 1997. Ovarian follicular cysts in dairy cattle. J Dairy Sci. 80:995-1004. Gibson, J. P., B. W. McBride, J. H. Burton, 1. Politis, and X. Zhao. 1992. Effect on production traits of bovine somatotropin for up to three consecutive lactations. J. Dairy Sci. 75:837-846. 182 Gill, J. L. 1978. Design and Analysis of Experiments. The Iowa State University Press, Ames, IA. ‘ Gong, J. G., T. Bramley, and R. Webb. 1991. The effect of recombinant bovine somatotropin on ovarian function in heifers: follicular populations and peripheral hormones. Biol Reprod. 45:941-949. Grohn, Y. T., V. Ducrocq, and J. A. Hertl. 1997. Modeling the effect of a disease on culling: an illustration of the use of time-dependent covariates for survival analysis. J. Dairy Sci. 80:1755-1766. Grohn, Y. T., S. W. Eicker, V. Ducrocq, and J. A. Hertl. 1998. Effect of diseases on the culling of Holstein dairy cows in New York state. J. Dairy Sci. 81:966-978. Grohn,Y. T. , J. A. Hertl, and J. L. Harman. 1994. Effect of early lactation milk yield on reproductive disorders in dairy cows. Am J Vet Res. 55: 1521-1528. Hady, P. J. and J. W. Lloyd, J. B. Kaneene, and A. L. Skidmore. 1994. Partial budget model for reproductive programs of dairy farm businesses. J. Dairy Sci. 77:482- 491. Hansen, W. P., D. E. Otterby, and J. G. Linn. 1993. Multi-farm use of bovine somatotropin for two consecutive lactations and its effects on lactational performance, health, and reproduction. J Dairy Sci. 77:94-110. Hard, D. L., F. Adriaens, C. G. Prosser, J. A. Newbold, R. B. Heap, R. H. Phipps, and M. J. Hannah. 1992. Effects of recombinant bovine somatotropin (bST) and nutrition on galactopoiesis, hepatic bST receptors and plasma hormone concentrations in the dairy cow. J. Physiol. 446:171P. (Abstr.). Harman, J. L., Y. T. Grohn, H. N. Erb, and G. Casella. 1996. Event-time analysis of the effect of season on parturition, parity, and concurrent disease on parturition-to- conception interval in dairy cows. Am J Vet Res. 57(5):640—645. Harrison, R. O., S. P. Ford, J. W. Young, A. J. Conley, and A. B. Freeman. 1990. Increased milk production versus reproductive and energy status of high producing dairy cows. J Dairy Sci. 73:2749-2758. Harsh, S. B., L. J. Connor, and G. D. Schwab. 1981. Managing the farm business. Prentice-Hall, Inc., Englewood Cliffs, NJ. Hartnell, G. F. 1994. Bovine somatotropin in the dairy industry: a review. Prof. An. Sci.10:85-101. 183 Hartnell, G. F., S. E. Franson, D. E. Bauman, H. H. Head, J. T. Huber, R. C. Lamb, K. S. Madsen, W. J. Cole, and R. L. Hintz. 1991. Evaluation of sometribove in a prolonged-release system in lactating dairy cows - production responses. J Dairy Sci. 74:2645-2663. HeatWatch“, Version 1.0. DDx Inc.. Denver CO. 1992. Hemken RW, Harmon R.J, Silvia WJ, Tucker WB, Heersche G, Eggert RG. Effect of Dietary energy and previous bovine somatotropin on milk yield, mastitis, and reproduction in dairy cows. J Dairy Sci 1991; 74:4265-4272. Hoblet, K. H., G. D. Schnitkey, D. Arbaugh, J. S. Hogan, K. L. Smith, P. S. Schoenberger, D. A. Todhunter, W. D. Hueston, D. E. Pritchard, G. L. Bowman, L. E. Heider, B. L. Brockett, and H. R. Conrad. 1991. Costs associated with selected preventive practices and with episodes of clinical mastitis in nine herds with low somatic cell counts. JAVMA. 199(2):190-196. Houben, E. H. P., A. A. Dijkhuizen, J. A. M. Van Arendonk, and R. B. M. Huirne. 1993. Short and long-term production losses and repeatability of clinical mastitis in dairy cattle. J. Dairy Sci. 76:2561-2578. Huber, J. T., Z. Wu, C. Fontes, Jr., J. L. Sullivan, R. G. Hoffman, and G. F. Hartnell. 1997. Administration of recombinant bovine somatotropin to dairy cows for four consecutive lactations. J. Dairy Sci. 80:2355-2360. J agannatha, S. , J. F. Keown, and L. D. VanVleck. 1998. Estimation of relative economic value for herd life of dairy cattle from profile equations. J. Dairy Sci. 81:1702- 1708. Jones, T.. 1997. Empirical Bayes prediction of 305-day milk production. J. Dairy Sci. 80: 1060-1075. Judge, L. J ., P. C. Bartlett, J. W. Lloyd and R. J. Erskine. 1999. Recombinant bovine somatotropin: association with reproductive performance in dairy cows. Theriogenology. 52:481-496. Judge, L. J ., P. C. Bartlett, J. W. Lloyd and R. J. Erskine. The impact of recombinant somatotropin on net farm income: analysis of four Michigan dairy farms. Submitted for publication in J. Dairy Sci. Judge, L. J ., R. J. Erskine and P. C. Bartlett. 1997. Recombinant bovine somatotropin and clinical mastitis: incidence, discarded milk following therapy, and culling. J. Dairy Sci. 80:3212-3218. 184 Karuppanan, P., M. C. Thurmond, and I. A. Gardner. 1997. Survivorship approaches to measuring and comparing cull rates for dairies. Prev. Vet. Med. 30(171-179. Keown, J. F. , and L. D. Van Vleck. 1972. Extending lactation records in progress to 305- day equivalent. J. Dairy Sci. 56:1070-1079. Kleinbaum, D. G., L. L. Kupper, and H. Morgenstern. 1982. Epidemiologic researcthrinciples and quantitative methods. Lifetime Learning Publications. Belmont, CA. Kronfeld, D. S. 1994. Health management of dairy herds treated with bovine somatotropin. JAVMA 204(1):116-130. Lamb, R. C., M. J. Anderson, S. L. Henderson, J. W. Call, R. J. Callan, D. L Hard, and L. Kung, Jr. 1988. Production response of Holstein cows to sometribove USAN (recombinant methionyl bovine somatotropin) in a prolonged release system for one lactation. J. Dairy Sci. 71 (Suppl. l):208(Abstract). Lean, I. J. , J. C. Galland, and J. L. Scott. 1989. Relationships between fertility, peak milk yields and lactational persisitency in dairy cows. Theriogenology. 31:1093-1103. Lee, L. A., J. D. Ferguson, and D. T. Galligan. 1989. Effect of disease on days open assessed by survival analysis. J. Dairy Sci. 72:1020-1026. Lefebvre, D. M., and E. Block. 1992. Effect of recombinant bovine somatotropin on estradiol-induced estrus behavior in ovariectomized heifers. J Dairy Sci. 75:1461- 1464. Leitch, H. W., E. B. Burnside, and B. W. McBride. 1990. Treatment of dairy cows with recombinant bovine somatotropin: genetic and phenotypic aspects. J. Dairy Sci. 73:181-190. LeLorier, J ., G. Gregoire, A. Benhaddad, J. Lapierre, and F. Derderian. 1997. Discrepancies between meta-analyses and subsequent large randomized, controlled trials. The New England Journal of Medicine. 337(8):536-542. Leonard, M., M. Gallo, G. Gallo, and E. Block. 1990. Effects of a 28-day sustained- release formulation of recombinant bovine somatotropin (rbST) administration to cows over two consecutive lactations. Can J Anim Sci. 70:795-809. 185 Lissemore, K. D., K. E. Leslie, B. W. McBride, J. H. Burton, A. R. Willan, and K. G. Bateman. 1991. Observations. on intramammary infection and somatic cell counts in cows treated with recombinant bovine somatotropin. Can. J. Vet. Res. 55:196- 198. Lloyd, J. W., and P. J. Hady. 1994. Partial Budgeting as a dairy consultation tool. The Bovine Pract. 26:84-87. Lloyd, J. W., J. B. Kaneene, and S. B. Harsh. 1987. Toward responsible farm-level economic analysis. JAVMA. 191(2):195-199. Lucy, M. C., T. L. Curran, R. J. Collier, and W. J. Cole. 1994. Extended function of the corpus luteum and earlier development of the second follicular wave in heifers treated with bovine somatotropin. Theriogenology. 41:561-572. Lucy, M. C., J. D. Savio, L. Badinga, R. L. De La Sota, and W. W. Thatcher. 1992. Factors that affect ovarian follicular dynamics in cattle. J Anim Sci. 70:3615-3626. Marion, B. W. , and R. L. Wills. 1990. A prospective assessment of the impacts of bovine somatotropin: a case study of Wisconsin. Am. J. Agr. Econ. 5:326-336. Marsh, W. E. , D. T. Galligan, and W. Chalupa. 1988. Economics of recombinant bovine somatotropin us in individual dairy herds. J. Dairy Sci. 71 :2944-2958. McClary, D. S., R. K. McGuffey, and H. B. Green. 1990. Bovine somatotropin: Part 2. Agri-pract. 11:5-11. McDaniel, B. T., W. E. Bell, J. Fetrow, B. D. Harrington, and J. D. Rehman. 1990. Survival rates and reasons for removal of cows injected with rBST. J. Dairy Sci. 73(suppl 1):159. McGuire, M. A., D. E. Bauman, M. A. Miller, and G. F. Hartnell. 1992. Response of somatomedins (IGF-I and IGF-IT) in lactating cows to variations in dietary energy and protein and treatment with recombinant n-methionyl bovine somatotropin. J. Nutr. 122:128. Michigan Dairy Herd Improvement Association. 1989. Records Manual. p. 2-18. Miller, G. Y., P. C. Bartlett, S. E. lance, J. Anderson, and L. E. Heider. 1993. Cost of clinical mastitis and mastitis prevention in dairy herds. JAVMA. 202(8):1230- 1236. 186 Mohammed, H. O. , M. E. White, and N. LaFaunce. 1991. Multivariate analysis of factors associated with calving rate in dairy cows. Theriogenology. 35(2):443-449. National Mastitis Council. 1987. Field Handbook on Bovine Mastitis. W. D. Hoards and Sons Co., Fort Atkinson, WI. National Milk Producers Federation. 1997. Prices received by farmers for 3.5% milk, 1920-1995. Arlington, VA. National Research Council. 1988. Nutrient requirements of dairy cattle. Sixth rev. ed. Natl. Acad. Press., Washington, DC. Neter, J ., M. H. Kutner, C. J. Nachtsheirn, and W. Wasserman. 1996. Applied Linear Statistical Models. Irwin, Chicago and Boston. Nielen, M. , Y. H. Schukken, D. T. Scholl. 1989. Twinning in dairy cattle: a study of risk factors and effects. Theriogenology. 32:845-862. Norman, H. D., P. M. VanRaden, J. R. Wright, and J. S. Coy. 1999. Comparison of test interval and best prediction methods for estimation of lactation yield from monthly, a.m.-p.m., and trirnonthly testing. J. Dairy Sci. 82:438-444. Oldenbroek, J. K., and G. J. Garsen. 1993. Effects of treatment of dairy cows with recombinant bovine somatotropin over three or four lactations. J Dairy Sci. 76:453- 467 . Oltenacu, P. A., and I. Ekesbo. 1994. Epidemiological study of clinical mastitis in dairy cattle. Vet. Res. 25:208-212. Oresnik, A. 1995. Effect of health and reproductive disorders on milk yield and fertility in dairy cows. Bovine Pract. 29:43-45. Parmar, K. B. , and D. Machin. 1995. Survival analysis: a practical approach. John Wiley and sons. New York. Peel, C. J ., D. E. Bauman. 1987. Somatotropin and lactation. J Dairy Sci. 70:474-486. Pelissier, C. L., L. J. Koong, and L. F. Bennett. 1978. Influence of 3X milking on milk and milkfat production. J. Dairy Sci. (Suppl.):132. (Abstr.). 187 Pell, A. N., D. S. Tsang, B. A. Howlett, M. T. Huyler, V. K. Meserole, W. A. Samuels, G. F. Hartnell, and R. L. Hintz. 1992. Effects of a prolonged-release formulation of sometribove (n-methionyl bovine somatotropin) on Jersey cows. J. Dairy Sci. 75:3416-3431. Posilac“ product label. 1993. Monsanto Co., St. Louis, MO. Posilac“ Profit Planner, Version 1.2.4. 1996. Protiva, a unit of Monsanto Co. , St. Louis, MO. Post-approval Monitoring Program (PAMP). Protiva, a division of Monsanto C0.. St. Louis MO. 1997. Remond, B., M. Cisse, A. Ollier, and Y. Chilliard. 1991. Slow release somatotropin in dairy heifers and cows fed two levels of energy concentrate. 1. Performance and body condition. J. Dairy Sci. 74:1370-1381. Remond, B., M. Cisse, A. Ollier, and Y. Chilliard. 1991. Slow release somatotropin in dairy heifers and cows fed two levels of energy concentrate. 1. Performance and body condition. J. Dairy Sci. 74:1370-1381. Remond, B., M. Cisse, A. Ollier, and Y. Chilliard. 1991. Slow release somatotropin in dairy heifers and cows fed two levels of energy concentrate. 1. performance and body condition. J Dairy Sci. 74:1370-1381. Rijpkema, Y. S., L. van Reeuwijk, and D. L. Hard. 1990. Responses of dairy cows to treatment with sometribove (r-BST) during three consecutive years. Livest. Prod. Sci. 26:193-216. Ross, W. A., J. B. Kaneene, and J. C. Gardiner. 1998. Survival analysis of risk factors associated with the occurrence of lameness in a Michigan horse population. Am J Vet Res. 59(1):23-29. Rothman, KJ. 1986. Interactions between causes. Modern Epidemiology. Boston: Little, Brown and Company. Ruegg, P. L., A. Fabellar, and R. L. Hintz. 1998. Effect of use of bovine somatotropin on culling practices in thirty-two dairy herds in Indiana, Michigan, and Ohio. J. Dairy Sci. 81:1262-1266. SAS System for Windows“, Version 6.10, 1994. SAS Institute, Cary, NC. 188 Schemm, S. R., D. R. Deaver, D. R. Griel, and L. D. Muller. 1990. Effects of recombinant bovine somatotropin on luteinizing hormone and ovarian function in lactating dairy cows. Biol Reprod. 42:815-821. Schmidt, G. H. 1989. Economics of using bovine somatotropin in dairy cows and potential impact on the US dairy industry. J. Dairy Sci. 72:737-745. Schmidt, G. H. 1989. Effect of length of calving intervals on income over feed and variable costs. J Dairy Sci. 72:1605-1611. Selvin, S. 1996. Statistical analysis of epidemiologic data. Oxford University Press. New York. Shoeffling, J. R., R. C. Angus, D. V. Armstrong, and J. T. Huber. 1991. Economic implications of bovine somatotropin use for the Arizona dairy industry. J. Dairy Sci. 74:2347-2352. Skarda, J ., E. Markalous, J. Slaba, P. Krejci, O. Skardova, and J. Zednik. 1992. Effect of methionyl bovine somatotropin in a prolonged-release vehicle on milk production, hormone profiles and health in dairy cattle. J. Dairy Res. 59:499-506. Spartan Dairy Ration Evaluator/Balancer, Version 2.01. 1992. Cooperative Extension Service and the Agricultural Experiment Station. Michigan State University, E. Lansing, MI. Speicher, J. A., H. A. Tucker, R. W. Ashley, E. P. Stanisiewski, J. F. Boucher, and C. J. Sniffen. 1994. Production responses of cows to recombinantly derived bovine somatotropin and to frequency of milking. J. Dairy Sci. 77:2509-2517. Statistix“, Version 4.1, 1994. Analytical Software, St. Paul, MN. Tauer, L. W. , and W. A. Knoblauch. 1997. The empirical impact of bovine somatotropin on New York dairy farms. J. Dairy Sci. 80:1092-1097. Thomas, J. W., R. A. Erdman, D. M. Galton, R. C. Lamb, M. J. Arambel, J. D. Olson, K. S. Madsen, W. A. Samuels, C. J. Peel, and G. A. Green. 1991. Responses by lactating cows in commercial dairy herds to recombinant bovine somatotropin. J. Dairy Sci. 74:945-964. United States Department of Agriculture, National Agricultural Statistical Service. 1997. Final Estimates, 1993-97. Washington, DC. 189 United States Department of Agriculture, Science and Education Administration, Dairy Herd Improvement Letter. 1980. Projection factors for milk and fat lactation records. 56(1), March. Van Arendonk, J. A. M. 1991. Use of profit equations to determine relative economic value of dairy cattle herd life and production from field data. J Dairy Sci 74:1101- 1107. Vandeputte—Van Messon, G. , and C. Burvenich. 1993. Effect of somatotropin on changes in milk production and composition during coliform mastitis in periparturient cows. J. Dairy Sci. 76:3727-3741. Vila-Godoy, A., T. L. Hughes, R. S. Emery, L. T. Chapin, and R. L. Fogwell. 1988. Association between energy balance and luteal function in lactating dairy cows. J Dairy Sci. 71:1063-1072. Vollema, A. R., and A. F. Groen. 1998. A comparison of breeding value predictors and longevity using a linear model and survival analysis. J Dairy Sci 81:3315-3320. Waterman, D. F., W. J. Silvia, R. W. Hemken, G. Heersche Jr., T. S. Swenson, and R. G. Eggert. 1993. Effect of bovine somatotropin on reproductive function in lactating dairy cattle. Theriogenology. 40:1015-1028. Weigel, K. A., B. A. Craig, T. R. Bidwell, and D. M. Bates. 1992. Comparison of alternative diphasic lactation curve models under bovine somatotropin administration. J. Dairy Sci. 75:580-589. Weigler, B. J., D. W. Hird, W. M. Sischo, J. C. Holmes, C. Danaye-Elmi, C. W. Palmer, and W. W. Utterback. 1990. Veterinary and nonveterinary costs of disease in 29 California dairies participating in the national animal health monitoring system from 1988 to 1989. JAVMA 196(12):1945-1949. Wells, S. J ., A. M. Trent, R. S. Collier, and W. J. Cole. 1995. Effect of long-term administration of a prolonged release formulation of bovine somatotropin (sometribove) on clinical lameness in dairy cows. Am. J Vet. Res. 56(8):992-996. Whitaker, D. A., E. J. Smith, J. M. Kelly, and L. S. Hodgson-Jones. 1988. Health, welfare and fertility implications of the use of bovine somatotropin in dairy cattle. Vet. Rec. 122:503-505. White, T. C., et. al. 1994. Clinical mastitis in cows treated with sometribove (recombinant bovine somatotropin) and its relationship to milk yield. J. Dairy Sci. 77:2249-2260. 190 Willeberg, P. 1993. Bovine somatotropin and clinical mastitis: epidemiological assessment of the welfare risk. Livest Prod Sci. 36:55-66. Willett, G. S., R. Blauwiekel, D. C. Grusenmeyer, and H. R. Hinman. A worksheet for analyzing the economic of bovine somatotropin adoption. 1994. Wash. State Univ. Coop. Ext. Serv., Ext. Bulletin 1777. Pullman, Wash. Zhao, X., J. H. Burton, and B. W. McBride. 1992. Lactation, health, and reproduction of dairy cows receiving daily injectable or sustained-released somatotropin. J Dairy Sci. 75:3122-3130. 191 IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII llll(Pilllljll("Plum