‘ V . . . _. , , . , , ‘ . ‘ u llllllllllllllllllllllllllllll 333 « 007 88 1059 we 1 l LIBRARY Michigan State __ University This is to certify that the thesis entitled N-ACETYL-B-D-GLUCOSAMINIDASE AS A PREDICTOR OF MILK LOSS AND RECOVERY FOLLOWING CLINICAL MASTITIS presented by David John Wilson has been accepted towards fulfillment of the requirements for Master of Science degree m Large Animal Clinical Sc1ences Wéfim Major professor g/aa/to 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution PLACE IN REI' URN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. | DATE DUE DATE DUE DATE DUE MSU Is An Affirmative Action/Equal Opportunity Institution end-emana- NHACETILHBEDHGIUCDSAMINIDASElAS4A PREDICTURLOFIMIIKIIDSS ANDlRECUVERY FOILOWING CIINICAILMDSTITIS By David John Wilson A.THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MRSTER.OF‘SCIENCE Department of Large Animal Clinical Sciences 1990 I), l .3; or» I 4' we) .1 L00 MALT N-ACEI'YIrB-D-GIIKDSAMINIDASE AS A FREDICIOR OF MILK 1088 AND W mum MQLMS'I‘ITIS By David John Wilson N—acetyl-B-D—glucosaminidase (NAGase) is a lysoscmal enzyme fanfiimmilkanddtherbodyfluids. MilkNAGaselevelhas beenreportedtobeincreasedinclinicalandmlinical mastitis, primarily due to leakage fran damaged secretory epithelial oe115.1:2:8r1°r13 This study evaluated milk NAGase at clinical onset as a predictor of severity of mastitis. Milk sanples were collected at clinical unset frail 508 wisodes of mastitis on a 1700 cow Mid'iigan dairy farm. Dailymilkproductionarddiseaseeverrtswererecordedfor allcowsintheherd. High milk NAGase levels were significantly associated with: 1) increased duration of treatment, 2) increased duration of clinical signs, 3) decreased daily milk production, and 4) increased risk of being called because of mastitis. These associations were weak. 'Iherefore, NAGase explained very little of the variability among cases, and was a poor prognostic test for clinical mastitis. ‘Ihis thesis is dedicated to my wife, Lesa, without whose support it could never have been carpleted. iii I inmld like to express my appreciation to my graduate cannittee, Drs. Paul Bartlett, John Kirk, Roger Mellenberger, and Edward Mather for their assistance with this project. Indispersable help was also provided by Mr. Charles Green, Flow Labs, Inc., and Software City. iv MOFW IISTOF'IEBIE‘S LIST OF FIGJRES mmowcnou IITERAIUREREVIEW IM'ERIAISANDI‘EEIIDDS Ch RESUBI‘S 15 DISCUSSION 32 CINCIUSIQI 39 APPENDDt—SAS-PC Programs Used in Analysis LISTOFREFERENCES LISTOF'IRBIES 1—Means for all cases and for etiologic agent groups- 24 2—oillingamxgallcasesandbyagentgrwps 25 3—Oorrelatim of NAGase with outoane variables 26 4—Predictors of milk loss following mastitis---—27 5—Predictors of duration of. nastitis therapy—--—-27 6—Predictors of time milk withheld following mastitis 28 7—Mastitis therapy and culling by stage of lactation 28 8—)hstitis culling by seasaa of onset 29 9—Mastitis etiologic agents by season of onset 29 vi LISTOF FIGURES FIGURE 1—-NAGase means by lactation rumba-r FIGURE 2—Production loss by stage of lactation vii 30 31 Increasing Inportance of Clinical Mstitis: Overthelasttwentyyears,corrtrolmeasureshavebeen developed and inplemented for pathogenic gram-positive incidence has declined on well- managed farms, a greater peroerrtage of mastitis cases are clinical in nature, caused byenvirawrtal pathogenssudlasthecoliforuBLQQLi, W. W. and MAW” Signs may include swollen or painful quarters, clunps, flakes or watery appearance of milk, fever, hypothermia, inappetence, dehydration, depression, Way and death.20 Severetoxicmastitiscarprises 10t023% ofallcasesofL cgn and We mastitis.16r17 Clinical signs include anorexia, depression and dehydration in addition to abnormal appearame of the milk and affected quarter.16'17 or these cows, seventy percent beoane agalactic.17 Clinical cases without severe life-threatening signs may also resllt in decreased milk production for extended periods of time. Forty-two percent of coliform infections with mild clinical sigzspersistedforlto22mrths.15 ‘Ihephenanenonof chronic productim loss following clinical signs is not limited to cows which have been affected by toxic nastitis. 2 Cows With less apparent clinical signs of mastitis caused by W am. F... 9211. We. eerironmental 8“Kai’s01'(J'thBprathogenscanalsodevelopachroniclossof milk prodwtion.15r17r18:19,20 MastitisisthenostcostlydiseaseinU.S.animl agriallture.22'23 Total mtary loss in the United States due to clinical mastitis was estimated at 700 million dollars in 1977, eqnling 4% of all total farm milk sales.23 Cost of clinical mastitis includes decreased milk production (69% of total), milk discarded during treatment and antibiotic withdrawal periods (11%), loss of animals due to culling or death (13%), therapartic costs (3%), veterinary fees (2%), and labor (2%).22r23 mete is substantial variation amng cases of clinical mastitis in lost milk production, duration of therapy, duraticn of clinical signs, aniriskofwllimfrmtheherd. Atalsetofclinical signs, it is difficult to predict which cases will be most severe. Accurate early prognosis for mastitis has eoonanic value. Sanecowssuffermarkedlossofproductionaniare non-responsive despite prolonged therapy. Feed, labor, and medicationvmldbesavedifthesecasescouldbemicldy identified and culled. Other cases recover fran clinical sigrs rapidly with minor production loss. If dairymen are micertainthatcmpleteclinicalreooveryisimimnt, they often treat even mild clinical mastitis for prolonged 3 periods. Perhaps the duration of antibiotic therapy and resultantdiscardingofmilkcouldbereducedinthesemild casesiftheywereidentifiedearlyinthecourseof disease. 'lhe main objective of this study was to evaluate N—acetyl-B- D-glucosaminidase (NAGase) as a prognostic test for clinical mstitis at onset of signs. mecifically, we investigated therelatimshipofNAGasetomilkprochctimdlarge, duratimofantibiotictherapy, timemrtilretmrntothe lactating herd, and risk of culling. WW: NAGaseisalysosanalauymefourdinmilk,ser1m,arriother bodyfluids. NAGaselevelhasbeenreportedtoincreasein milk of cows with clinical and nonclinical mastitis. MagnitrieoftheincreasemaybeasMasonermrflred fold.1r2r3r4'5r3r1°vl3 Part of the increase in NAGase in mastiticmilkisduetodanagetosecretoryepithelialcells mudlreslltsinlealageofmzynefrmuleoellular cmpartment into milk.1r2r8v10r13 This has led to specllatimthatmsasecanbeusedasaneaslreofseverity ofanepisodeofmstitisincontrasttotestssmhas Sanatic Cell Count (SCC) which measures inflamnatory W to mastitis.2r3r4v5r9r1°r13 ‘Ihere appears to be a 4 strong relaticnship betveen SCI: and NAGase. Correlation betweenNAGasearriSCCwasreportedhigherinclinicalcases (r=.97) than in nonclinical mastitis (r=.62) .14v32v33 'IhepercentageofNAGaseincreaseinmastitiswhichisdue todanagedsecretorycellsisamatterofsmedebate, with estimates ranging fran 90%toabart45%.112I8110113I24 'Ihis isaninportantquestimrincreaseinmilkmsaselevelasa neasureofseverityofmstitiswufldbemostmeaningful if it was highly specific for danage to milk producing tissue. Age of the cow, stage of lactatim, and parity affect the baselirelcvelofmcaseinmilk. BasalNAGaseishighest during the first 30 days of lactation, declines sharply, and risesthrulghaltlactationmrtildaysinmilkeaweedszso (Fig. 3),4.5.25.33 NAGase baseline also increases with the lactationrnmberofthecow,withthegreatestincm occurring between the first and second lactaticrs (Fig. 4),4.5.10 There are considerable differences in baseline NAGase anong m.4'5 'mese sources of variation nust be taken into account when interpreting NAGase in mastitic quarters. 'Iodate, threemethodsofattatptirgtominimize the effects of basal NAGase variatim have been described. the interquarter ratio is calwlated by dividing the mstiticquarterNAGasebythelowestNAGasevalueofthe otterquartersfranthesamecow(thereferenoequarter).405 5 Ombinations of the natural log of mastitic NAGase (log NAGase)arritheinterquarterratiohavealsobeen evaluated.5r12 Difference in NAGase activity between the mastitic and reference quarter has also been used.13 These transformations are reported to minimize background differences and facilitate comparison of NAGase levels among different (116.4'5'12'13 Incmparismwithothertestsmnastiticmilk, NAGasehas beenreportedtobeasuperiorpredictorofassociated lost milk production. Log Nagase was more negatively correlated with daily milk pmductim (r2=-.39, p<.0001) than milk antitrypsin or plasnin levels.9 Difference in NAGase ammo infectedandreferencequarterswasabetterirflicatorof lost milk prediction (r2=-.4l, p<.001) than milk lactose, chloride, leukocyte percentages, bacterial status, or Log soc.13 'Ihesereslltsarefransuriyofnornlinicalcases ally. Arntterinportantdmracteristicofmcaseisitsrapid increaseduring clinicalmastitis. Within9t012hours afterquartersbeoaneinfected, interquarterNAGasehas alreadyrisentoalevelatornearitspeakva11.1e.14v15 'Batisassoonasanysigrssmflmasswellingandhardness ofquartersoratmornalappearancesofmilkaredetectable. Peaklevelismaintainedfor 24t036 110113.14'15 ‘Ihese 6 damcteristics have led to speculation that the NAGase assay may be developed as a rapid test for predicting severity of clinical mastitis at the time of onset, yet to airlomledgemprevimsreseard’ihasbeencamicted specifically addressing this question.2r3r4r5v9r10r13 MA'I'ERIAISANDMEBDIB Sanple Collection: Milk sanples were collected at clinical alset fran every episodeofclinicalmastitismidloowrredduringaémrth period in a 1,700-cow Holstein dairy herd in central of a hot, hard, or mllen quarter and/or abnormal flakes, color or consistency of the milk. Clinical mastitis was further confirmed with the California Mastitis Test (041'). Any gel formation following addition of the cur reagent (+1 or greater) was considered positive. All four quarters of clinical cows were stripped mice and then individually sanpled for later NAGase determination and Wisconsin Mastitis Test (VHF). For clinical quarters only, teat ends were swathsdusing 70% isopropyl alcohol andforenilkwas aseptically collected for subsequent microbiological examination. All sanplee were frozen and returned to our laboratory. 7 Sanples were obtained frun 716 episodes of clinical mastitis between January 29 and July 27, 1988. Four milk sanples leakedbeforetheycouldbefrozen. Another42 sanpleswere excluded because of visible dirt or manure contamination of the microbiology sanple tubes. An additional 71 samples weremtslhnittedmletoinconsistentlabellingofthe mstiticquarterontheNAGasesanplingtubesardthe microbiology tubes. (he hundred seventeen (117) sanples were lost; 599 retained. Oneepisode of clinical nastitisooclrredccmcurrently with aleftdisplacedabarasm (Ink), andonecowhadclinical mstitisinallll garters, arrithereforehadmcontrol qaarter. mesecdoepisodeswereexcluded. Five miseries werelosttofollow—upduetothecowmovingtoashowbarn manotherfarmwheremdailymilkrecordsarekept. Case Definition. Smecowshachlinicalepisodesinthesamequarterwithin afewdaysofeadiother. Anysuchepisodethatoccnrred withinSdaysofrecovery(ret1m1franthehospitalbarn) followingtheearlier episodewasnotconsideredanewcase of mastitis. Anysuchepisode thatooalrredwithin 14 days ofreooveryfrantheearlierepisodewasmtconsidereda newcaseofmastitisifthesameetiologicagentwas 8 isolated fran both episodes. 'Ihese exclusions eliminated 84 mastitis episodes franthestudybecausetheywerejudgedto beacontinlationofapreviouscase. 'Ihisleft508cases for inclusicninthestudy. Milksanples franeadlcase were mailed to our microbiology laboratory, but 8 sanples were spilled in shipnent. Microbiology: Bovinebloodagararflmwmkeyagarwere inocllatedwith .1 m1frunead1asepticmilksanpleardim1batedat37Cfor 48 hours. Colony norphology, reanalysis, and lactose fernentatim were used for the first level of diagnosis and calfirmatory tests were used for final identificatim of pathogens.21 Wisconsin Mastitis Test: “Ewasperfornedmallcpartersanples. 'I'womlofmilk fruneadtsanpleuereplacedinaflfl'testtubeardmixed wichmlofWMI'reagent. Brasscapswithstandardsized holeswereplacedonthetubes. AfterlOsecondsofmixing, thetubeswereinvertedforlSsecaris,placeduprightforl mirute of settling, arri measured for height (m) of column ranaining inthetube.29 maseMeasuranent: NAGase concentration was determined using a commercially available test kit.7:12 Using an 8-channel micropipette, 10 ulofmilkpersanplewereplacedintoeachtestwellofa 96 wall plate. Substrate (50 ul 4-Methylunbelliferyl-N- aoetyl-B-D-gluoosaminide) was added to each test well. Sanpleswereincubated forlSminutesat 25thilebeing agitated by a plate shaker. 'Ihen 100 1.11 of stopping buffer wereaddedtoead1well.'n1eannmtofNAGaseinthemilk sample is proportional to the amount of 4- nethyltm'belliferone (4111) released fran the substrate per unittime. Becausectmisafluorescentcmpamd,the NAGase concentration in each sanple test well was quantified bymeaslringthefluorescenoeusingafluorrmeter. ‘Ihe fluoraneter was set at excitation wavelength 355 m and anission wavelength 480 m. Calibration of the fluoraneter, including correction for roan tarperature, was performed before every test plate was sunnitted. NAGase activity was expressed as uMoles/min/L. NAGasewasneasuredforead'iclinicalquarterardforone referenceanterfrunthesamecow. ‘Iheqlarterusedfor referenoeNAGasewasthequarterwiththelowestVMI'value. Interquarter ratio was calculated by dividing the nestitic 10 quarterNAGasebytheNAGasevalueofthereferenoequarter franthesamecow. The following abbreviations were used for Mase values: MforNAGaseinthemastiticquarter,NAGDforthe differencebetweenNAQdardreferernecparterNAGase,and Mforrefererneantermisase. NAGaseresultsare rqaortedasnean+/-standarddeviatiminmitsof uMoles/min/L. Therapy: Oniceoftherapyforclinicalmastitiscaseswasbasedupm appearance of milk, palpatim of affected quarters, and rectal taperature. Ifrectaltarperaturewasgreaterthan 39.7 c (103.5 F), 4 g oxytetracycline was administered intravenously (IV) once daily, while cephapirin was administered intramanmary (11M) twice daily. If rectal taperaulrewaslessthan39.7 C, buttheanterwas swollen or milk was grossly abnormal, IBM cephapirin only was administered. If milk appeared normal and quarter swellingwasabsent, oowswerere’olmedtothemilkingherd following a 3-day antibiotic withdrawal period. Cases that were evaluated and not treated at all with antibiotics were inmediately returned to the lactating herd. 11 Recording of Data: Milkproductimwasrecordedeveryday forallcowsatall milkings. All records of daily milk production were converted to twice daily milking (2X) basis using conversion factors according to the lactation nunber of the cows.28 For conversion of three times daily (3X) milk weights to 2X, factors of .833, .855, and .869 were used for let lactation, 2nd lactation, and 3rd and greater lactation (Brd-t) cows, rewectively. Forallcows, reoordswerekeptofIVandIntmastitis therapy, imludingtotalmmberoftreatmentsardnmberof daysuntilreturntothemilkingherd. Wiveevents, diagnosisandtherapyofdiseasesbesides mastitis, deaths, Gills,andreasonforcullingwerealsorecorded. Stageof lactation was measured using the continlous variable DIM (days in milk at onset), and also usitg the categorical variable DGRP (days in milkgrmp). Each case was placed into one of 5 DGRP categories: 1-30 om, 31-90 om, 91-150 DIM, 151-250 DIM, >250 DIM at mset. Cases were also classified according to parity (1,2, or 3+ lactations) . 'niesuxlyperiodmsdividedintothreeseasors: 1. Winter msdeffiedastheperiodfranJamaryZ9 (firstdayof sanple collection) until mind: 31. 2. Spring was fran April 12 luntilMayBl. 3. SmuerwasfranJmelmrtilJulyZ? (last day of sampling). Caseswere categorized acoordingto seasmofmset, soeachcasewasassignedtooneseason using date of onset of clinical sigrs. DatamreenteredintotheDBASEIIdeatabaseprogram, W031 Measures of Disease Severity: Milkproductimdiargewasusedasamaaslreofseverityof clinical nastitis. Meandailyproduction forthelast7 daysbeforeonset(BASE)wassubtractedfrmmeandailymilk production for the first 14 days following return to the milkingherdfrunthehospitalbamforeadiepisodeof clinical mstitis. mismanmilkproductimdiargeis calledDIFF. DIFFismeasuredinkg.,andisnegativeif productim declines relative to BASE. ForacasetobeircludedincalcllatimofDIFF, certain criteriahadtobemet. At least4catpletedailymilk mightshadtoberecordedduringthe7dayBASEperiod,and no treatment or withholding fran the lactating herd for any diseasecouldbepresentduringthisperiod. 'Ihereason thatscmemissingmilkmightswereallmedduringBASEwas 13 that sanecows contractedmastitis early in lactation. ‘Ihey weremlyinthelactatingherdlongernlghtohave4daily milkweights before onset. Forthefirst 14 daysafter recovery, catplete daily milk weights were required for at leastthefirst7daysafterreturntothelactatingherd. 'metworeasonsforincludingcaseswithsaremissingpost- recovery daily milk weights in calwlatim of DIFF were: 1) thecowwasdriedoffafter7daysor2)thecow's transpmder was not read properly and therefore her productim valuesweremissim. Trarmonders randanly failedforS%ofallmilkingsead1day. Reoordswere eacluiedifanycornlrrentdiseasemspresentbeforeor afterthecase. 'IhereforeDIFFcalldrntbecalwlatedfor 164casesbecause:10werenotfreshlongernlghtohavea validBASEbeforeonset,46hadmastitisduringthe7days before thecase, 66 contracted mastitis again within? days afterreturnfranthemastitisbarn,9weredriedoff, 32 weresold,and1hadanlll§afterclinicalmastitis. Another measure of clinical and eoonanic severity of mstitis is the duration of antibiotic therapy before wssation of clinical signs or ailling. ‘Ihe variable IIJR represents the duration (in days) of antibiotic therapy for each clinical mastitis case. [HR was transformed for sure statistical tests to the natural log(wR +1) to achieve a l4 “metimeacow'smilkmswithheldfransalewasalsoused asameasureofmastitisseverity. 'Ihevariablean'isthe nmberofdaysmitilreturntothemilkingherdormtil culling following each case of clinical mastitis. (111' was trarsformed to the natural log(CIJT +1) for some statistical teststoadiieveamremrmaldistrihrtim. Cullirg was used as another measure of mastitis severity. CULL and CIJLIM are categorical variables. are possible outcanes foraJlLare"yes"or"no"regardingwhetherthe cowlefttheherdbeforetheendoflactatimforany reason. (mm (yes or no) refers specifically to culling due to mastitis. GeneralLinearModels: Twenty nine potential predictor (independent) variables were evaluated (ore-my ANOVA) for their relationship with each of the continuous outcane (dependent) variables 111R, (111', and DIFF. 'nresesanepotentialpredictorswerealsocarparedtothe categorical outcane (dependent) variables CIJLL and CIJLIM using ore-way ANOVA or Chi-square analysis. 15 111R, CIJT, andDIFFweretheneachusedasthedependent variable in separate general linear models. 'Ihesemodels wereca'structedtouseinformatimobtainableatcnsetofa caseof clinical mstitistopredict severity of that case. All significant predictors were identified and retained in eachfinallinearmodel. NAGase thresholds: Previalsreportshaveslggestedthattheremaybethresholds of NAGase level useful in detecting mastitis.3 It is micertainwhetherNAGasethresholdscanbeusedtogroup clinical cases according to severity. 'Iherefore, various levelsofNAGaseweretestedtoseeifcasesabovethat levelsufferedworseartccmesthanthosebelwthelevel. Descriptive Epidemiology: Etiologic agents isolated included 5; m (13.8%), m m (8.6%), Dipltheroids (A; m) (2.8%), m. non- ag's (7-0’3), Colifoms (El 921.1. Eel-siege. m, and W) (22.4%), Psaflawnads (5.6%), m, Wig, 2% art} Yeast (2.8%), Oaltaminated (11.2%), 16 m (1.2%), Mixed Major Pathogens (4.8%), and No Growth (19.8%). Mean NAG! and NAGD were significantly different among agents (p<.01, one-way ANOVA). As shown in Table 1, NAQJ was lowest for Mixed Major Pathogais and highest for m non- ag's. NAGD followed the same pattern (Table 1). NAGM mean for all 508 sanples was 9.03 +/- 10.16 LN and mean NAGD was 7.85 +/- 10.03 uM (Table 1). m did not differ significantly among different agents. Tablelalsostnvsproductimdiangeafterclinicalmastitis as measured by DIFF. Mean value of DIFF was -2.2 +/- 3.4 kg for 344 cases. DIFF mean varied among etiologic agarts (p <.05, one-way ANOVA). Mixed Major Pathogens had the greatestlossofanyagentgroup. leastmilklosswas associated with 52:13:13, M, We, and Yeast (Table 1). Diration of therapy for clinical cases was measured as 111R (indays). Fora11508cases,themeanvalueowawa32.0 +/— 3.5 days, rangingfranoto26 days of treatment. [11R was not significantly different anong etiologic agents. Tineofmilkwithholdingduetomastitiswasneasmedas daysCIJTforeachcase. MeanvalueofCIJTforSOBcaseswas 17 5.7 +/- 7.9dayswitharangeof0to48days. 'Iherewere no significant differences anong etiologic agents in days (IJT. ‘I‘ablez showstherates ofculling forallcasesandfor different etiologic agents. Culling for mastitis or for all reasons was not significantly different ammg agents. NAGase was associated with parity (p<.05, ore-way ANOVA). MardNAGDwerelowestiandlactatimcmrsardhighest in 3rd plus lactation cows (Figure 1). NAGR was not associated with parity (Figure 1) . Simple linear regression wasusedtotestforeffectofstageof lactationonNAGase. m and man were not significantly affected by stage of lactatim (test for slope-0) . Analytical Epidemiology: NAGaseasapredictorofmilkloss (DIFF): Irrzreasedeasassociatedwithgreatermilkloss forthe 344casesthatret1m1edtothelactatingherd. Table3 stursthecorrelatimofNAQtwithDIFF forallcnsesand for different agents. NAGD was not significantly associated with DIFF (sinple regressim) . 18 NAGaseasapredictoroftreatmentrospalse (IIJRandCUI‘): Increased NAGase was associated with longer treaunent for mastitis, as shown by the correlation coefficients for NAGM and NAGD with IIJR for all 508 cases (F.23, r=.21, respectively, p<.01) . Table 3 show the relationship betweeiNAQlarriIIIRforeadlpathogen. HigherNAGasewas mated with longer milk withholding following mastitis. NAG! air! NAGD were positively correlated with CIJT for all 508 cases (r=.24, r=.22, respectively, p<.01). Table 3 stmstheconelatimbetweenNAGMarriaJrforeadlagent. NAGaseasapredictorofculling (CIJILardOJLIM): IrmeasesinMardNAGDwereassociatedwithincreased culling due to mastitis (CUIIM) (r=.13, r-—-.12, respectively, p<.01). Risk of calling for all ream ((1111,) was not associated with differences in NAGase level. CowsallledduetomastitishadahighermeanNAGaselevel at onset than thosenot culled (p<.01, Student'st test). Mdidnotenceedlfitflfor 67 (13.2%) ofthecases,and noneofthosecaseswerecllled. ‘mereweremNAGase levels above which a majority of cases were culled. 19 Allcasesvvereassigrxedtoorxeof4Mgru1ps, franthe bcttan25% oftheNPQ‘IvaluestothetopZSft. Wranges and their respective mastitis culling rates were as follows: < 1.99, 1.5%, 1.99- 5.10, 6.4%, 5.11- 13.61, 7.1%, > 13.61, 12.5%. The increased risk of milling due to mastitis was statistically significant (p<.01, Chi-square) . EValuation of NPBase Predictive Ability in Conjunction With Other Predictive Variables: 'mebestpredictivegeneral lirearmdelformilkloss (DIFF) used 5 variables obtainable at day of «set. In order of statistin significance these were: BASE productim, NAGM, DGRP (stage of lactation categories), laboratory batch of NAGase determinatim (see below), and NAGR. This model was highly significant (p<.0001) but 79% ofthevariationanrmgcasesinDIFFwasmtexplained (R2 2.21) (Table 4). As NAQJ, NAGR, and BASE increased, there was a tadacytwardgreaterprodctim loss (lower DIFF). Oowsfran91tol§0daysinmilklostthemstproductim (DIFF -3.3 +/- 3.8 kg). Production was best maintained by cows in the first 30 days of lactation (DIFF -.05 +/- 4.1 kg) (Figure 2). ‘Ihereappearedtobeaomsistentproblanwiththemilk stardardsusedwiththecaunercialNAGasetestkit. 20 Following their recmstitution, the NAGase activity of the milkstarriardsircreasedsteadilybegimlirgwithinafew hours. Therefore, NAGasesanplesnmmdifferenttestwell plates could not be accurately oampared by adjusting for NAGaseactivityofthemilkstaniards. ‘Ihisraisedthe question of whether there were differences among laboratory batdiesinmeasuranentofNAGase. Milksanplesweretested forNAGaseindironologicalorderin7batdies,ttms laboratorybatchincludestheeffectsofseasmofonset. A strongmlationshipbetweenbatdiardseasalofonsetis evidencedbytneoii-squareteet (x2=732, p<.01). Ninety threeperoentofthevariationinNAGaseamngsanpleswas notduetobatoh. ulringoorstruztionofgenerallinear models, season and batch were never statistically significant inthesamemdel. Forcedinclusion ofboth variablesincreasedthetypeIIIpvaluesforboth. 'Ihis statsthatbatdlexertedupmNAGasetheeffectsofseasonal as well as laboratory differences. As laboratory batch and seasmofonsetprogreesedfranwintertospringtoammer, less milk production was lost (more positive DIFF) . The final linear model to predict um (duration of treatnent) included in order of statistical significance: m, DIM (daysinmilkatcnset), andseason. 'Iheerrtire model was statistically significant (p<.0001) , but overall predictive ability was weak (192.11) (Table 5). A trend was 21 obeervedthatasmircreased,uRwasalsogreater. Mastitis earlier in lactationwas treated longer. As season progressedfranwintertospringtoammer, wRdecreased. 'mefinal linearmodel forpredictingan' (tinemltilretnlrn to the lactating herd) used m, 0114, and parity, in order of statistical significance. 'Ihe overall relationship was statistically significant (p<.0001). However, like the othermodels, thismdelleftmlohofthevariabilityamng cases unexplained (122:.09) (Table 6). 'niere was a trend thatasNAQdandparityincreased,dayszirureased. Cases occurring earlier in lactation did not recover as quickly. Evaluaticn of Using a NAGase 'Ihredlold for Prognosis: ‘nle21oasesabovetheMthresholdof 32 uMhadgreater mm (4.9 +/- 4.7 days) and an (12.5 +/- 9.9 days) than cases below 32 uM (p<.01, one way ANNA). Production dropped for all 21 cases. DIFF was -6.8 +/— 6.1 kg for cases above the threshold, a significantly greater drcp in production than for cases below the 32 m4 threshold (p<.01, one-wayAMNA). Casesabovethethresholdhada 14.3%CIJIIM rate ompared to a 6.37% rate for those below it. 11115 difference was not statistically significant (p=.15, Chi- Square), possiblyduetosmll numbers. 'Iherewasno 22 increased risk of culling for all reasons for the high NAGaseoows. Influence of Parity, Stage of lactation, and Season on Mastitis: wR increased with parity (p<.01, ore-way ANOVA) . For cases in lactation 1, UR mean was 1.1 +/- 1.6 days, lactation 2, 1.7 +/- 2.6 days, lactatim 3+, 2.5 +/— 4.2 days. Table 7 shows that IIJR varied with stage of lactation (Ix-RP) (p<.05, one-wayAmVA). 'I'reatmentwaslongest foroowsfran91to 150 daysinmilk, andshortestforoowsgreaterthanzso days in milk. Parity did not significantly affect milk production loss following mastitis (me-way ANOVA) . (mm rate was significantly associated with IIERP (p<.01, Chi-Square) and season (p<.05, Chi-Square). No cows greater than 250 days in milk were allled for mastitis, while 14.1% ofthose 31 to 90 days inmilkwere culled (Table 7). CULIM ratewashighest forcasesinthespringandlowestinthe W (Table 8) . 'Iheperoentageofcasesattrihltedtodifferentagentswas significantly different among seasors (p<.01, Chi—Square) . m spp. and diphtheroids oarprised a higher percentage of cases in winter, while ooliforns, m non-ag's (m 23 Miss and M M). and Psalms were increasedinspringandsmmerCI‘ableQ). ‘Iheoverall rate of new clinical infections was not significantly affected by season (Chi-Square) . N .5 TABLE 1—Means for all cases and for etiologic agent groups Days Days Prod. Agent W ** NAGDb ** of Rxc outd dlfe* All cases 9.0 7.8 2.0 5.7 -2.2 gm 7.9 6.8 1.6 4.5 -2.5 Staph. spp. 6.9 5.7 1.7 4.4 -l.6 No Growth 11.5 10.4 2.3 6.2 -l.9 Diphthercids f 5.7 4.9 2.8 6.6 —1.2 g, non-ag's 9 13.9 12.6 2.2 6.5 -3.4 Coliforms h 9.3 8.3 2.2 6.2 -2.6 pseudonmeds 7.5 6.0 1.3 3.8 -1.5 genetic and Others 1 10.7 9.9 2.9 6.4 -0.5 Contaminated 7.0 5.5 2.2 5.7 -l.8 genius 9.0 7.0 0.8 3.3 -2.2 Mixed mjor Pathogens 3 5.6 4.1 2.6 9.3 —4.9 (a) NACM = N-acetyl-B-D-glucosaminidase (NAGase) in the mastitic quarter expressed in uMoles/min./L. (b) NAGD=MastiticquarterNAGasemimsreferencecparter NAGase. (c) Days of Rx = Days of antibiotic therapy for each case of mastitis. ((1) Days out=Daysunti1 returntothelactatingherdfor each case of mastitis. (e) Prod. dif. = Change in daily milk production following mastitis expressed in kg. (f) Formerly We: W. (9) mi Mlactiae and Seep.- men's. (h) _E_. coli, Klebsiella, Citrcbacter, and Enterobagrg. (i) Serratia or Salmonella or Proteus, and Yeast. (j) M major pathogens isolated fran sanple. * significantly different among agents, p<.05, one-way ANOVA. ** significantly different among agents, p<.01, one-way ANOVA. 25 TABLE 2-Culling'among’a11 cases and.by'agent.groups am mice all am All cases 6.7% (34/508) 33.2% (169/508) fiyggrggg 4.4% (3/69) 31.9% (22/69) m spp. 4.7% (2/43) 32.6% (14/43) No Growth 6.1% (6/99) 30.3% (30/99) Diphthercids 7.1% (1/14) 21.4% (3/14) s; non-ag's 2.9% (1/35) 28.6% (IQ/35) Ooliforms 9.8% (ll/112) 33.0% (37/112) Pseudmnnads 0.0% (0/28) 28.6% (8/28) mm and Others 21.4% (3/14) 57.1% (8/14) Contaminated 5.4% (3/56) 33.9% (19/56) minus 16.7% (1/6) 83.3% (5/6) NfixedgMajor Pathogens 12.5% (3/24) 41.7% (10/24) See'Table 1 fOr key. 26 ME 3—Oorrelation of NAGase with outcane variables Agent Correlation coefficients for NAGM with: Days Days Prod of Rx Out Diff. All cases .23 ** .24 ** -.10 * m .20 .19 -.16 m spp. .35 * .28 -.26 No Growth .15 .22 * .04 Dipltheroids .06 .10 -.15 g. non-ag's .27 .27 -.40 * Ooliforms . 13 . 19 * . 19 Psexlanonads .75 ** .74 ** -.55 ** m; and .58 9* .48 -.18 Others Omtaminated .50 ** .33 * -.22 goings .64 .30 .86 Mixed Major Pathogens -.02 .07 .17 * statistically significant at p<.05. ** statistically significant at. p<.01. See Table 1 for key. 27 TABLE 4—Predictors of milk loss following mastitis Type III Source of variation DF Type III SS Prob. Base production a 1 1750 .0001 m 1 502 .0010 Stage of lactation 4 1684 .0001 laboratory batch ‘0 6 829 .0066 NAG-R C 1 129 .0926 Error 330 14990 (a) mily milk production for 7 days before clinical onset. (b) Laboratory batch for NAGase test. (tested in chronological order so includes influence of season of onset). (c) NAG-R = Reference quarter NAGase. R2 = .21, p<.0001. Grand mean for productim loss = -2.2 lg. See Table 1 for explanation of NAGM. TABLE 5—Predictors of duration of mastitis therapy Type III Source of variation DF Type III SS Prob. NPR! 1 19 . 0001 0m 3 1 8 . 0002 Season b 2 2 . 0123 Error 503 272 (a) DIM = Days in milk at onset of clinical mastitis. (b) Season of onset of clinical mastitis. = .11, p<.0001. Grand mean for duration of therapy = 2.0 days. See Table 1 for explanation of NAGM. 28 TABLE 6—Predictors of time milk withheld following nastitis Type III Source of variation DF Type III SS Prob. - m 1 34 .0001 DIM 1 8 .0062 lactation.a 2 6 .0588 Error 503 535 (a) First, second, or third-plus lactation. = .09, p<.0001. Grandmeanfortimeuntil recovery=5.7 days. See Table 1 for explanation of NAGM, Table 5 for explanation of DIM. 'mBIE 7—Mastitis therapy and culling by stage of lactation Days in milk Duration of Ollling for at onset Therapy a * Mastitis u 1-30 2 3 6.1% (3/49) 31-90 2 2 14.2% (15/106) 91-150 2.9 7.9% (8/101) 151-250 1 8 4.9% (8/163) >250 1 2 0.0% (0/89) (a) Days of antibiotic therapy for each case of mastitis. * Significantly different among stages of lactation, p<.05, one-way ANOVA. ** Significantly different among stages of lactation, p<.01, Chi-Square. 29 ME 8—Mastitis culling by season of onset Seasm of Onset Culling for Mastitis * Winter a 6.6% (12/183) Spring 10 9.8% (18/184) Sunnerc 2.8% (4/141) (a) Jan. 29 until Mar. 31. (b) April 1 until May 31. (c) Jim 1 until July 27. * Significantly different among seasons, p<.05, Chi-Square. ME 9—Mastitis etiologic agents by season of onset m All Agent Winter Spring Sumner Cases All cases 36.4% 35.8% 27.8% 100.0% m 14.3% 14.5% 12.2% 13.8% SEQ; spp. 11.5% 6.7% 7.2% 8.6% No Growth 20.9% 19.0% 19.4% 19.8% Diphther'oids 5.0% 0.6% 2.9% 2.8% §_,_ rrJn-ag's 5.0% 7.8% 8.6% 7.0% Coliforms 20 . 9% 25. 7% 20. 1% 22 . 4% Pseuiomonads 4.4% 6.7% 5.8% 5.6% Sen—“am and Others 2.8% 2.2% 3.6% 2.8% Contaminated 4.4% 14.0% 16.6% 11.2% Millus 1.7% 0.0% 2.2% 1.2% Mixed Major Pathogens 9.3% 2.8% 1.4% 4.8% Agents statistically differ among seasons, p<.01, cli- Square. See Table 1 for explanation of agents, Table 8 for explanation of seasore. NAGase (u m/min/ L) 12 10 Figure 1. NAGase means by lactation number El Mastltlc I Reference 2 . 3 a .1 Lactation number 30 Production Change (kg) Figure 2 Production loss by stage of lactation 1-30 31-90 91-150 151-250 >250 Days In Milk at Onset 31 DISCIJSSIW : MilkProductionaniNAGase: Increased milk NAGase was associated with decreased milk production following clinical mastitis. This was as expected, ththecorrelatianwastoosmalltobeofnuch biological significance. NAGase by itself was not a strong predictor of last milk, and for sane etiologic agents W and coliforne) , the relatiaiship was reversed. Although statistically nan-significant, these trench suggest that for califarm and w cases, the likelihood of retirntcproductimisgreaterwhenNAGaseisgreater. 'mis alggeststhatNAGasenaynotbespecific farsecretory celldanege, orat leastmtforchrenicinpairwrtofmilk secretion capability. Duration of Mastitis and NAGase: Mastitiscaseswithhighermsasetendedtobetreatedwith antibiotics and withheld fran the lactating herd for longer periods. As with production change, the relationship was weak. UsedaloneNAGasewasapoorpredictorofchranicity of clinical signs. Apparently, damage to glarters necessitating more extensive treatment and longer 32 33 canvalascencebeforerecoveryfrunclinical signsisnot specifically indicated by NAGase activity. Culling and NAGase: Caseswith elevatedNAGasewerereportedwlledbecausecf mastitis at a significantly greater rate, but the associationwasweak. Farcasaewithinhigherrangesaf NAG! values, nestitis culling percentage was greater. Despitethistrend, therewasncthresholdafNAGaselevel abovewhidlamajarityafthecaseswereallleddueto mastitis. mismaybefurtherevidencethatdecreased capability for lactation is not specifically indicated by highermilkNAGaselevelsinclinicalmastitis. Another reason for difficulty in predicting milling is that culling decisiare are sanewhat subjective and depend upon judgement by dairy farmers. While ecananically inportant, culling is not an objective measure of severity of disease. NAGase Ombined With Other Predictors: With reliable predictive models, cows most likely to rebound rapidlyfranclinicalmastitis, thoseexpectedtahave moderate production loss, and those likely to suffer major production loss could be rapidly identified on day of onset. Managanent decisions could be made utilizing records of 34 previous productim, reproductive status, and previous diseaseaswellasthemastitispragnosisgeneratedbythe model. Oows whose previous performnce and current prognosis (determinedmingareliablemadel)werebothpoar could be treated to mximize survival for slaughter. alpportive therapy for clinical sigrs without use of antibiotics could be irstituted so that they could be rapidlyallled. Intramamarytherapyormtherapycouldbe usedforcaseswithalikelihoadofminimalpradrctianloss. Forhighlyvaluedcowsthatwerepredictedtahaveasevere casebythemodel,intenseintravermstherapywithfluids and antibiotics vmld be indicated even if initial clinical signs were mild. Even in cmbinatim with other information obtainable at (meet of clinical signs, NAGase did not consistently identify cases that wild ultimately have the greatest production losses. Nevertheless, NAGase was a better predictoraf lostmilkthanmilkcultureorm. ‘Iherewas atrendthathigherpraducingcowswithhighermilkNAGase at onset suffered the mast past-mastitic production loss. This agrees with our hypothesis and previous speculation about NAGase and clinical severity 2.3.4.6.9.10.13, but the mdelinitsartiretywasapoorpredictorofproductian diange. Biological significance and practical usefulness of the nodal are doubtful. 35 NAGasecmbinedwithotherinformtionfrandayofaiset could not accurately predict duration of therapy. NAGase wasaiperiartovurandmilkcultureforpredictingum, however. AtrerdwasabservedthatcaseswithhigherNAGase weretreatedforalangertime,aswasexpected,butthe predictive ability was weak. Parsistence of clinical signs could not be reliably predicted using NAGase combined with otherpredictorvariables. Again,NAGasewasabetter predictorofwrthanmilkalltureorm. Caseswith higherNAGasetendedtotake langertarecover. This surportedourhypothesimbutthemdelleftmlchofthe variabilityammgcasesmiexplainedandtlmsappearstahave littlepracticaluse. Effect of Parity arri Stage of lactatim m NAGase: lbstiticanterNAGasewaslowestiandlactatiancowsand higiestincowsintheir3rdorgreaterlactatian. This differs fran previous reports that NAGase elevation with mastitis increases with each lactatim.4r5:1° NAGase was not significantly affected by stage of lactaticn in the stay cows, in contrast to previous reports.4v5I25 It appearethatthevariabilityamangcowsinNAGaseresparee tomstitisisinfluencedmrebyotherfactorethanbydays inmilkandparity. Useafinterquartercmparisanaf NAGase activity within cows may be preferable to attempting 36 to adjust for differences in parity and stage of lactation. mite blood cells arxi msecretory ductular epithelial cells haverecentlybeenpostulatedtobemjarsourcesofNAGase in mastitic milk.24 The degree of white cell infiltration intheearlystagesofclinicalnestitisasmeasuredbyscc ishighlyvariablebetweencows. 'Ihismaybepartofthe reasonwhyfactorsotherthandestructianafnamnarycells appeartobeinportantcartributorstototalNAGaselevelin clinicalcases. Clinical Mastitis and Factors in Addition to NAGase: Treatment duration increased with parity. This may be due to more pranmced clinical signs or more chronicity of mastitis in older, high producing cows. It may also indicate greater perceived value of older cows such that they warrant longer treatment. OowsgreaterthanZSOdaysinmilkatansetweretreatedfor theshortesttimeformastitis. Nacowsinthisgraupwere Gilledformastitis. Cowsinadvancedlactatimdonothave thepotentialtolaseasmxzhdailymilkproductimor becane agalactic far as nuch of their lactation as cows with mastitis in earlier lactation. Advanced lactation cows usually recover rapidly with minimal mastitis treatment or 37 they are dried off earlier than originally planned. Clinical mastitis fran 91 to 150 days of lactation resulted inthelangesttreatment. 'Ihisisprabablyduetomarked clinical signs and high value attributed to these cows. Mastareproducingwellenaughthattheyaremtculled, and theyaretooearlyinthelactatianarrirepraductiancycle to be dried off. Fifteencf106casesfrcm31tc90daysinmilkresultedin culling farmstitisrthiswasthehighestcill rate farany stage of lactation. If marked drop in milk production affectscowsinthispeakstageafthelactatiancurve,cr theyareperceivedasdlranicmastitiscows,theyareata stageaf lactatianwheretheyrepresentamajorecarmic lass. Manyofthanarenatpregnant, ardtheexpenseof rebreeding and maintaining these cows is often not justified: therefore more of them are culled. Mm.anddifilthercidscausedlessmstitisinthe springanismmer,whilemastitiscausedby§tm,_ Wammm, colifarns, andpsaxiatrmads increasedinspringandslmmer. Diringthesttxiyperiod, thereweremanagenentdlangesirstitutedmthefarmasa resultafincreasedawarenassafmastitis. Washingand drying teats more thoroughly and segregation of mastitic cows (including milkingthenlast) were irstituted during 38 May (spring). ‘Ihereductianindiphtheraidandflm spp. nastitisastheprojectprogressedwasprabablyaresultaf trespringardsnmerareenviramentalpathogensthat primarilyimradethemamnaryglarfifrancows'mrrumdings between milkings. Increased environmental mastitis is camarduringwetandmmyconditiansafwarmerseasuis. NAGase‘IhresholdasaPrognasticTest: WefoundathresholdafNAQdabovewhichcaseshadabad prognosis. 'IhosecowswithNAngreaterthanntflwere more likely to suffer greater production loss and require a longercourseoftherapy. Itslmldberenenbered,tnrever, thatcasesbelaw32uMNAQJcannotbeassumedtabemild cases. Perhapsthrasholdscouldbeusefulinmastitis prognosis. Furtherstudymotherfarmsvmldberequired before specific thresholds could be recamended for widespread use. Milk Standards for NAGase Test: The milk standards in the cmmercial NAGase test kit increasedmarkedlyinNAGaseactivitywithinafewhcursof reconstitution. FailureafthemilkNAGasestandardswasa 39 concern. UseafcamercialmilkstandardNAGaseactivityto adjust for differences annng test plates should be done with caution. (INCHJSIQ‘I : Ourresults indicate thatNAGaseisabetterpredictaraf severalmeasuresafclinical severityofmastitisthanother information currently available at onset, such as age, stage of lactation, season of cmset, level of milk production, milkculture, andWII'. But, NAGasealane arcatbinedwith other variables could not consistently predict the sequelae of clinical cases. Most of the variability in outcane among clinical mastitis cases retained meorplained. APPENDIX SAS-PC Prograns Used in Analysis 'IhisistheSAS-PCprogramusedtocanvertthedatasetinto its final form. Transformations of variables, daily milk weight adjustments, and mathematical calculations are performed by this program: DATA SAVJIVD: ssr SAV.GRN: IDU=IDG(001~+1): Imr-IDGUIJR-t-l): DAT=FRan+Dm: IF DAT<=821 'IHEN SEAs=1; IF DAI>821 AND m1~<=882 'niEN IE90ANDDDIGI'30'II-mmb2; LEISOANDDIMGI'90'IHENII5RP=3; IEZSOANDDIMGI‘ISO'IHENIERH: GT2 IF W=L98 mm m1; IF NAQDL98 AND NAGM<=5.10 mm W2; IF macaw AND NAQK=IB£I mm W3; IF W116i mm W4: IF NAGM<=NAQ2 an NAQD=NAGR W 1W: Immense: IF mecca OR NAQi>=NAGR mm W; mm; mm; IF NAGD<=1.13 mm NAch=1; IF NAGD>1.13 AND NAGD<=3.40 mm NAGDCr—Z; IF NAGD>3.40 AND NAGD<=10.77 mm NAsDc=3: IF NAGD>10.77 mm manor-4: INAQ)=IDG(NAGD+10.87): IF mm 80 0 mm W1; IF mm no 0 mm VMIR=13 Ime<=2nmnvmm=nIFma>2ANDm<=10m W2: IF mm>10 AND W=26 mm W3; IF W26 mm W: IF mm OR W m meter/mm: WIOGMHM): IF mm OR W mm W; 40 41 HWWWI;IFWOANDMD<=SIHENW23 IFWSANDWIQ'IHENWB; IFVMID>19THEN W; WWW-F35)? IFH4II€US7GE777CI1HIDUS7EDOMHIINUS7=J IFHm‘JUSGGE777GiHmJS6mO'H-IENHENUSS=.; IFHIDRJSSGE777G2WUS5EQOMHENUS-.; IFHdINUS4GE777GiRflNUS4EDO'H-IENHENUS’J IFHflNUS3GE777G23flNUS3EDOMHENUS3=J IFHm‘IUSZGE777mH‘flNUS2EDO'H-IENHENUSZ=.; IFHIENUSlGE777CRHdINUS1m0mml=J P1GE777GIP1EDOWP1=J P2GE777GQP2EDOWP2=J P3GE777G1P3EDO'IHENP3=.3 P46E777mP4m0'nIENP42J P5GE777QRP5EDO'IHENP5-J P6CE777CRP6EDOWP6=J P7CE777CRF7EQO'BIENP7=.3 P8GE777GZP8EDO'IHENP884 P96E777G2P9ED011IENP9=J P10 77 (RPIO O'IHENP10=.3 P11 77(11P11 O'IHENP11=.3 77mm O'IHENP12=.3 3 77mPI3 O'D-IENP13=.; 4 77WP14 O'IHENPllk-J THEN m7=nm137*.833: 'DIEN WUSG=HMJSG*.8337 mm HflNIJSS=m5*.833; 'IHEN mmwmmew: mm EMINUS3=HdINUS3*.833; ‘IHEN HIINUSZ=HWINUSZ*.833; THEN HENUSl=HIIMJSlM8333 THEN P1=P1*.833; THEN P2=P2*.833; THEN P3=P3*.8333 ‘IHEN P4=P4*.833; THEN P5=P5*.833: THEN P6=P6*.833; 'IHEN F7=P7*.833; THEN P8=P8*.833; THEN P9=P9*.833; 'IHEN P10=P10*.833; THEN P11=P11*.833; THEN P12=P12*.833: 'IHEN P13=P13*.833; 'IHEN P14=P14*.833; THEN HflNUS7=IMINUS7*.854; THEN H’IINUS6=HIENUS6*.854; THEN IMINUSS=H4INUSS*.854; 'IHEN HENUS4=IMINUS4*.854; THEN HENUSB=IMINUS3*.854; 'IHEN HENUS2=HIINUS2*.854: QH’JQQHC}HGQHQQQQQGQHQQGGQGHGUQQHQUHQGHHQHH aaaaaaaaaaamaaaaaafifiaaaaaaaasg $5E55555§5§§55Egggggggggggg 85888855888588588858888888888888 888888888888888888888888888a8999 EEEEEEEEEEEEEEEEEg‘éEEEEEEEE 42 THEN HENUSI=HENUS1* . 854 ; THEN P1=P1* . 854 3 THEN P2=P2* . 854 3 THEN P3=P3* . 854; THEN P4=P4* . 854; 'IHEN P5=P5* . 854 3 THEN P6=P6* . 854 3 THEN P7=P7* . 854 3 THEN P8=P8* . 854 3 'IHEN P9==P9* . 854; THEN P10=P10* . 854 3 THEN P11=P11* . 854 3 THEN P12=P12* . 854; 'IHEN P13=P13* . 854; THEN P14=P14* . 854; 'IHEN HENUS7=H4INUS7* . 869 3 'H‘IEN HENUS6=HENUS6* . 869 3 'IHEN MISS=REIJIIS5* . 869 : 'H-IEN WW‘F . 869; THEN HENUSB=HIENUS3* . 869 ; THEN HflZNUS2=FMINUS2* . 869 3 THEN HdINUSl=H4INUSI* . 869 3 THEN P1=P1* . 869 3 THEN P2=P2* . 869; 'IHEN P3=P3* . 869 ; 'IHEN P4=P4* . 869 : THEN P5=P5*.8693 THEN P6=P6* . 869 3 THEN P7=P7* . 869 3 THEN P8=P8* . 8693 WHEN P9=P9* . 869 3 'IHFN P10=P10* . 869 ; THEN P11=P11*.869; THEN P12=P12* . 869 3 THEN P13=P13* . 869 3 'IHEN P14=P14* . 869 3 I a355555'sasgaagasaaagaaaasagggagasa §§§§§§§§§§§§§§§§§§§§§§§§§§§§EEE§§§ ANDIACI'm EH14t1I?ClQHQUQQHGHHQHQGHQHHHQHHHQQQC}Emmi} E‘EESZS{33525652562535fifififififififififififififififififififififififi 888888888358888588888888888888888888 % a g E E E E E E E; E I I I I I H V E E :8 a F a a 3 I I I I I I ); RET2=MEAN(P8,P9,P10,P11,P12,P13,P14): P6,P'7,P8,P9,P10,P11,P12,P13,P14); DIFF1=REr1-BASE: DIFF2=REI'2-BASE: DIFF=REI‘—BASE: IF W m AGENT=.; IFAGENI‘=1 WWI; mmzmmz; HWB mm m3;mmr=4m1me=4;mmr=smmp=s; HWWW? ORAGINHORWMW; IF mm W m7; IF AGENT-=11 ca AGENP=15 mm mm? mm mm MERP=9;IF W13 'mEN HERP=103 43 IF AGENT>15 THEN IISRP=113 IF TIME>=600 AND TIME<=14OO 'IHEN SHIFT=13 IF TIME>14OO AND TIME<=2200 THEN SHIFT=2; IF TIME>2200 (R TD4E<600 THEN SHIFT=33 IFSOID>=1‘IHEN(IJIL=1:IFSOID=O'IHENCIJIIF=O: IFCAR=1WCIJILM=1§IFCARNE1THENCIJIIM=O; KIN: misSAS-PCprogramoorstructsthefimlgeneral linear model for production difference (DIFF): HDCGLMDK'IZFSAVJDD; CLASSIIERPIABIRY; DDIEL DIFF=BASE m IIZRP m m: RUN; misSAS-PCprogramoonstructsthefinalgenerallinear model for log of duration of therapy (log of IIJR=IJR). HDC GIN MVJI‘be CLASS SEAS; mum. 1W DIM SEAS: 11th 'mjsSAS-PCprogramoonstnntsthefinalgenerallinear nodelforlogoftimetmtilreturntolactatingherdaogof arr-400). HUG GIM MINA“); CLASS IACI‘; mm W DIM IACI‘; KIN; IISTOFREFERENCES Kitdm, B.J., Middleton, G. and Salmon, M. Bovine milk N-acetyl -beta-D-glucosaminidase and its significance in the detection of abnormal udder secretions. J. DairyRes. 45:15-20, 1978. Obara, Y. and Kanatsu, M. Relationship between N- acetyl-beta-D-glucosaminidase activity and cell count, lactose, chloride, or lactoferrin in cow milk. J. Dairy Sci. 67:1043-1046, 1984. Kitchen, B.J., Knee, W.S., Middleton, G. and Andrews, R.J. Relationship between the level of N- acetyl-beta—D—glucosaminidase (NAGase) in bovine milkandthepresenceofmastitispathogens. J. Dairy Res. 51:11-16, 1984. Mattila, 'I‘., Syvajarvi, J., Jensen, N.E. and Sandholm, M. N-acetyl-beta-D-glucosaminidase and antitrypsin in subclinically infected quarter-milk samples: effect of bacteria and hemolysins, lactation stage, and lactation umber. Am. J. Vet. ms. 47:139-142, 1986. Mattila, T. and Sandholm, M. Antitrypsin and N- acetyl-beta-D-glucosaminidase as marker of mastitis inaherdofAyrshirecows. Am. J. Vet. Res. 46:2453-2456, 1985. Kitchen, B.J. Enzymic methods for estimatim of the somatic cell count in bovine milk. 1. Developnent ofassayted‘miquesardasmdyoftheirusefulrms in evaluating the sanatic cell content of milk. J. miry Res. 43:251-258, 1976. Lirflto—l’bppam, S. and Makinen, M. A microtiter pl ate assay for N-acetyl-beta-D—glucosaminidase using a fluorogenic substrate. Analyt. Biochem. 148:50-53, 1985. Fox, L.K. and Schultz, L.H. Effect of infection status on quarter milk production and carposition following anitted milking. J. Dairy Sci. 68:418- 423, 1985. Mattila, 'I'., Syvajarvi, J. and Sandholm, M. Milk antitrypsin, NAGase, plasmin and bacterial replication rate in whey. Effects of lactation 44 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 45 stage, parity and daily milk yield. Zentralbl- Velterinarmed. (B) 33:462-470, 1986. Schaar, J. and Funke, H. Effect of subclinical nastitis on milk plaaninogen and plasmin cmpared with that on sodium, antitrypsin and N—acetyl-beta- D-glucosaminidase. J. Dairy m5. 53:515-528, 1986. Mattila, 'I'., Pyorala, S. and Sandholm, M. Omparison of milk antitrypsin, albmnjh, n-acetyl- beta-D-glucosaminidase, somatic cells and bacteriological analysis as indicators of bovine subclinical mastitis. Vet. Res. Oamun. 10:113-124, 1986. Milk NAGase Test Instruction Manual. Flow laboratories Inc. McLean, Va., 1987. Fox, L.K., Shook, 6.13., and Schultz, L.H. Factors related to milk loss in quarters with low sanatic cell counts. J. Dairy Sci. 68:2100-2107, 1987. Runkel, J.R., Bushnell, R.B., Cullor, J., and Alecng, J. Studies on induced Ml; mastitis with relationships among N-acetyl-B-D- glucosaminidase, bacterial and sanatic cell counts. Cornell Vet. 77:225-234, 1987. Fitz-Gerald, C.M., Dearth, H-C, Kitchen, B.J. The relationship between the levels of free fatty acids, lipoprotein lipase, carboocylaterase, N-acetyl-beta- D-glucosaminidase, sanatic cell count and other mastitis indiws in bovine milk. J. Dairy Res. 48:253-265, 1981. mrphy, J.M., Harson, J.J. Infection of the bovine udder with coliform bacteria. Cornell Vet. 33:61, 1943. Eberhardt, R.J., Natzke, R.P., Newbould, F.H.S., Nonnecke, B., and ‘Ihatpson, P. Colifom mastitis- a review. J. [hiry Sci. 6:21-22, 1979. Schalm, 0.W., Carroll, R.J., and Jain, N.C. Bovine mastitis, pp. 254-255, 1971. Anstutz, H.E. Bovine medicine and surgery, pp. 1066-1067 , 1980. Blood, D.C., Radostits, C.M., and Henderson, J.A. Veterinary Medicine, pp. 466-473, 1983. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 46 Laboratory and field handbook (11 bovine nastitis. National Mastitis Oamcil, 1987. Blosser, 'I'.H. Ecamic losses fran and the national researchprogrammmastitisintheUnitedStates. J. [niry Sci. 62:119-127, 1979. Jasper, D.E., Wald, J.S., Dbd‘lrie, R.B., Philpot, W.N., Farnsworth, R.J., and Spencer, S.B. Bovine mastitis research needs, funding, and sources of support. Proc. 21st National Mastitis Council, pp. 184-193, 1982. Kaartinen, L., Kuosa, P.L., Veijalainen, K., and Sandholm, M. Ompar‘tmentalizatim of milk N-acetyl- B-D-glucosaminidase (NAGase) . Release of NAGase franthecellularcmpartmentbystorage, freezing and thawing, detergent and using cell stinnlants. Unpublished mamscript. Mattila, 'I‘., O'chle, D., and Frost, A.J. Distribution of lysosanal enzymes within the non- secretory epithelium of the bovine namnary glarrl. J. [3in Rafi. 55:339-344, 1988. Nagahata, H., Noda, K., and Abe, '1‘. Changes in blastogenic response of bovine lynphocyta during acute clinical mastitis. Jpn. J. Vet. Sci. 49(6) :1148-1150, 1987. Rothbauer, D.L., Buchner, E.C., Wells, S.J. Effect of vaccination with E. coli (13 the incidence of acute mastitis. Proc. let Annual Conference American Association of Bovine Practitioners, 112- 115,1988. Wiggans, G.R., Dickinson, F.N. Starriardization of NCIHIP dairy cattle lactation records. National Cooperative Dairy Herd Inprovement Program Fact Sheet G-2, 1985. Wiscorsin Mastitis Test Kit. 2. D. Rowdy, Orem, Utah. DBase III. Ashton-Tate, Culver City, Cal. Statistical Analysis m for Parscnal Ompxters (SAS-PC). SAS, Cary, N. C. Callieri, C.A. , Oorbellini, C.N. , mick, A.M. , Martero, H.L., Murphy, M.I., and Muset, G.B. Carparison of aquatic cell count and N—acetyl-B-D- 33. 47 glucosaminidase activity of quarter milk sanples . Rev. Med. Vet., Argentina 68:69-75, 1987. Miller, R.B., Paape, M.J. Effects of parity, bacteriological status, stage of lactation, and dry period on N-acetyl-B-D—glucosaminidase activity of milk and dry secretim. J. Dairy Sci. 71:2508-2512, 1988. HI IGR N STATE UNIV. LIBRARIES WWI”WWWWIIIUIWIWIWHI 1293007881059 23 if