Low Somatic Cell Count: a Risk Factor for Subsequent Clinical Mastitis in a Dairy Herd

Transcription

1 Low Somatic Cell Count: a Risk Factor for Subsequent Clinical Mastitis in a Dairy Herd W. Suriyasathaporn,*,1 Y. H. Schukken, M. Nielen, and A. Brand *Department of Farm Animal Health, Yalelaan 7, 3584 CL Utrecht, The Netherlands Department of Population Medicine and Diagnostic Sciences, Cornell University, Ithaca, NY Wageningen Agricultural University, Wageningen, The Netherlands InterACT Agrimanagement, Giethoorn, The Netherlands ABSTRACT A case-control study was conducted to evaluate factors measured at the udder inflammation-free state as risk factors for subsequent clinical mastitis. The factors including somatic cell count (SCC), body condition score, milk yield, percentages of milk fat and milk protein, and diseases were evaluated for their association with the results of udder inflammatory response. The results of the response were specified as presence (case) and absence (control) of clinical signs of mastitis. Data on Holstein Friesian cows calving from January 1984 to November 1996 from a commercial farm with low bulk milk SCC were used. Univariable and multivariable random-effect logistic models were used to evaluate the effect of those factors on the risk of clinical mastitis. The following variables were associated with increased odds of case versus control events in the univariable analysis: early lactation period, low SCC, high milk yield, high percentage of milk protein, high percentage of milk fat, low body condition score, retained placenta, and milk fever. For the final multivariable model of all variables used for analysis, only low SCC remained significantly associated with increased risk of subsequent clinical mastitis. The authors concluded that very low SCC during the udder inflammation-free state are associated with increased risk of clinical mastitis. (Key words: milk composition, clinical mastitis, body condition score, random-effect logistic model) Abbreviation key: BMSCC = bulk milk SCC, OR = odds ratio, SCC-PRE = SCC before UIR, UIR = udder inflammatory responses. Received June 21, Accepted January 26, Corresponding author: W. Suriyasathapron. w. suriyasathaporn@vet.uu.nl. 1 Current address: Department of Veterinary Physiology, Khon- Kaen University, 40002, Thailand. INTRODUCTION Mastitis is the most costly disease in the dairy industry, especially because of decreased production, cost of treatment, extra labor, and the increased rate of cow replacement (3). As with most infectious diseases, mastitis incidence depends on three components: exposure to microbes, cow defense mechanisms, and environmental and management factors. In the past decade, the standard mastitis control program has provided hygiene and management practices to control intramammary infection (15). A decrease in bulk milk SCC (BMSCC) is an indicator of the success of the control program. Although dairy producers who have herds with low BMSCC have been able to decrease the prevalence of mastitis with contagious pathogens, these herds still show a high incidence of clinical mastitis by environmental pathogens (11, 20). An increased incidence of clinical mastitis caused by environmental pathogens was reported to be directly associated with impairment of cow defense mechanisms (13, 28). Mastitis is caused by bacterial invasion into the udder. The small numbers of somatic cells in milk that are normally present attempt to resolve this IMI immediately. Both bacteria and leukocytes in the infected quarters release products, many of which are chemoattractants for leukocytes. Neutrophils move rapidly from the bloodstream into the secretions of infected quarters, causing an increase in somatic cells in milk. If the bacteria are destroyed, recruitment of neutrophils into the gland is ceased, and only a mild inflammatory episode will be required to restore health in the gland (12). Occasionally, the innate defense mechanisms of the infected mammary gland lose the battle with bacteria, and bacteria multiply. This leads to a prolonged immune response within the mammary gland. Various cell types in the udder produce abundant soluble factors, such as cytokines, which cause clinical signs of mastitis. Hence, the udder inflammatory responses (UIR) to IMI can result in an absence or a presence of clinical signs of mastitis J Dairy Sci 83:

2 LOW SOMATIC CELL COUNT AND CLINICAL MASTITIS 1249 Figure 2. Definition of case and control in the present study. Figure 1. Bulk milk SCC in a dairy herd between January 1984 and November Because leukocytes in the udder are present to resolve the IMI, once an intramammary challenge occurs, a very low SCC may predispose cows to a higher risk of clinical mastitis. In experimental studies, it was shown that factors such as ketosis, low peripheral leukocyte count, and low SCC were associated with a more severe mastitis response (13, 21, 24, 28). However, no data are available from field studies indicating which factors measured before UIR are related to an increased risk of clinical mastitis. Therefore, a case-control study was conducted to evaluate which factors were associated with an increased risk of clinical mastitis. MATERIALS AND METHODS To eliminate variability in exposure to pathogens and management factors, data from one commercial farm with low BMSCC were used (Figure 1). The data consisted of test-day records on milk components of Holstein Friesian cows calving from January 1984 to November The average herd size was 90 milking cows. Facilities and herd management were described previously (25). Data on cow identification, parity, date of calving, daily milk production, BCS, disease occurrence, and dates of culling were recorded. Milk production data including kilograms of milk production, percentage of milk fat, percentage of milk protein, and SCC, provided by the Royal Dutch Cattle Syndicate (Dutch equivalent of DHIA), were collected approximately every 2 wk. Clinical mastitis was detected and identified by the farmers during premilking udder hygiene procedures on the basis of clinical signs, including abnormal milk, a hard or swollen udder, or both. The standard procedures used in body condition scoring were as described by Edmonson et al. (8). Cows were scored for body condition every 3 wk by one of the two veterinary practitioners (YHS & AB) visiting the farm, and once at calving by the farmer who was trained by the practitioners. Definition of Case and Control A change from an udder inflammation-free state to udder inflammation state within 30 d was used to define cases and controls (Figure 2). An udder-inflammationfree state, as defined in the present study, is a state in which a cow had an SCC of < cells/ml. An udder-inflammation state was defined as either SCC cells/ml (17) or clinical mastitis. A control event was an event with an absence of clinical signs of mastitis but presence of an increase in SCC, and a case event was an event in which a cow showed signs of clinical mastitis. An increase in SCC can indicate the first response of defense mechanisms. Therefore, a quick change of increased SCC to clinical mastitis (within 10 d) was also defined as a case (Figure 2). Because no SCC data on an udder-inflammation-free state were available, all data on UIR, either increased SCC or clinical mastitis, originating from the first incidence postpartum would have been missing. Therefore, the UIR included the state in which cows experienced the first udder-inflammation state within 30 d postpartum without any data of an udder-inflammation-free state (Figure 2). Definition of Variables Risk factors considered in our model are listed in Table 1 and include parity, lactation period, last recorded SCC before UIR (SCC-PRE), milk yields, percentage of milk fat, percentage of milk protein, BCS, and diseases. Data on all risk factors were collected before UIR was determined, but no longer than 30 d prior to UIR. To provide an obvious effect of changing SCC, the SCC-PRE, which was SCC < cells/

3 1250 SURIYASATHAPORN ET AL. ml, was divided by 20,000, and were considered a continuous variable. Because BCS decreased sharply in early lactation, BCS as a risk factor for severity of mastitis occurring in the first month postpartum was only included when measured within 10 d before the case or control event occurred. Statistical Analysis Pearson s correlation coefficients were calculated to identify the associations between milk component factors, and chi-square tests were used to identify the associations between disease factors. Repeated measurement analysis was used to evaluate the risk factors of severity of mastitis. To provide a practical method with reasonable statistical efficiency to analyze discrete and correlated data, random-effect logistic models were used (10). The variance-covariance structure in a random logistic model is analogous to the exchangeable working correlation in the Generalized Estimating Equations method and no bias of regression estimates is present in this method (10). Therefore, the GENMOD procedure (16) with the exchangeable correlation ma- trix was used to fit correlated response in the present study. The observations in a single cluster are uniquely identified by parity within individual cows, so parity within cows is the clustering variable. The simple random-effect-logistic regression model can be written as: logit P(Y ij = 1) = β 0 + Σβ i x i + ε, where ε = R + e, R = correlation matrix, e = random error, the fixed part includes the intercept (β 0 ) and covariates x i and regressors for covariates (β i ). Model Building Because Escherichia coli is a pathogen that usually causes clinical mastitis in low BMSCC herds (2), a preliminary test was performed to evaluate the effect of pathogens on severity of mastitis. However, pathogen data on control events were hardly recorded (Table 2). Therefore, a comparison between the models with and without E. coli mastitis for the risk factor SCC-PRE was used to evaluate the effect of pathogens. A significant difference of odds ratio (OR) between both models was considered having the effect of pathogens. Table 1. Risk factors for the model based on individual cows used in the analysis of udder inflammatory responses (UIR). 1 Risk factor 2 Coding N (%) 3 Mean (SD) 4 General factors Parity 1 (1) 292 (21.4) 2 (2) 280 (20.5) 3( 3) 792 (58.1) Lactation period Early (1st mo pp) 272 (19.9) Mid (2nd to 5th mo pp) 499 (36.6) Late (> 5 mo pp) 593 (43.5) Milk component factors SCC ( 20,000 cells/ml) 3.95 (2.77) Milk yield (kg/d) (9.25) Percentage of milk fat 4.34 (0.68) Percentage of milk protein 3.55 (0.36) BCS factor BCS 5 before UIR 1 (1 to 1.75) 29 (3.0) 2 (2 to 2.75) 349 (36.3) 3 (3 to 3.75) 479 (49.8) 4 (4 to 5) 105 (10.9) Disease factors 6 Retained placenta 21 (1.5) Genital infection 7 36 (2.6) Cystic ovary 32 (2.3) Lameness 88 (6.5) Milk fever 30 (2.2) 1 UIR was indicated by either presence (clinical mastitis) or absence (SCC > cells/ml) of clinical mastitis. 2 Risk factors were collected before UIR (the risk factors data were collected within 30 d before UIR). 3 Number of observations and percentage of particular value on a total of observations were calculated for categorical variables. 4 Means and standard deviation (SD) were calculated for continuous variables. 5 Body Condition Score 1 = thin, and 5 = fat. 6 Disease factors were codes as 0 = no, 1 = yes. 7 Genital infection included acute endometritis, chronic endometritis, metritis, and pyometra.

4 LOW SOMATIC CELL COUNT AND CLINICAL MASTITIS 1251 Table 2. Distribution of pathogens causing clinical mastitis in defined case events in a dairy herd between January 1984 and November Number Percentage of Percentage of Pathogen of cases all observations sampled observations Streptococcus dysgalactiae Streptococcus agalactiae Streptococcus uberis Staphylococcus aureus Coagulase-negative staphylococci Escherichia coli Actinobacterium pyogenes Others Fungi Mixed culture Negative culture Contaminated Not sampled but recorded Univariable random-effect logistic models were developed for each risk factor. P values less than 0.05 were considered significant. Multivariable random-effect logistic models were used to evaluate the factors associated with the results of UIR. The free entering method was used, and the maximum log-likelihood test was calculated to identify significant levels. Stepwise forward regression analysis was used and a variable had to be significant at the 0.10 levels before it could enter the model. A variable remained in the model, when its P value was < For a categorical variable, all dummy variables were entered at the same time. Because of a large amount of missing data for BCS and milk component data, the analysis was performed with four data files. The first model used data of parity, lactation period, and disease factors, and included 1364 records. The second model was based on 926 observations, which included parity, lactation period, diseases, and milk component data as risk factors. The third model included 962 observations including parity, lactation period, diseases, and BCS factors. The last model used all variables to create the model and included 643 observations. RESULTS Data for 1520 lactations from 659 cows were used. A total of 969 cases of clinical mastitis (in 718 quarters from 464 lactations) of 295 cows were recorded during a total of 456,770 d at risk (lactation period ranged from 0 to 665 d). An overall incidence rate of this herd was 0.77 cases per 365 cow-days at risk, or 0.14 cases per 365 quarter-days at risk. The majority of cases of clinical mastitis were caused by E. coli (43% of sampled observations). In a total of 15,713 SCC data points, 1,715 observations were higher than cells/ml. A total of 1364 UIR events comprised 733 control events and 631 case events. Thirty-five observations were clas- sified as cases because clinical mastitis was seen within 10 d after increased SCC. To test the effects of pathogens on the result of UIR, data from 122 observations of clinical mastitis caused by E. coli were excluded in the model without E. coli mastitis. No difference was found between OR of SCC- PRE for the model with and without E. coli mastitis (OR = 0.71 and OR = 0.71). Results of Pearson s correlation coefficients showed significant correlation among SCCPRE, percentage of milk fat, percentage of milk protein, milk yield, and BCS (P < 0.01). Associations between disease factors showed the correlation between milk fever and retained placenta (P < 0.01), and between retained placenta and genital infection (P = 0.008). Univariable Analyses Proportions of case and control for SCC-PRE, milk yield, milk protein, and milk fat are shown in Figures 3, 4, 5, and 6, respectively. Odds of cases versus controls decreased with an increase in SCC-PRE (Figure 3), percentage of milk protein (Figure 5), and percentage of milk fat (Figure 6), but the odds of getting clinical mastitis increased with an increase in milk yield. Table 3 shows the variables that were significantly related to UIR in the univariable analyses. Of the original 13 variables considered, eight risk factors showed significant levels, including SCC-PRE, milk yield, percentage of milk fat, percentage of milk protein, lactation period, BCS, retained placenta, and milk fever. Cows in early lactation and midlactation ran higher risks of getting clinical mastitis than cows in late lactation. The UIR was negatively associated with SCC-PRE (P < ). The odds of getting a clinical case decreased by a factor 0.71 with each unit (= 20,000 cells/ml) increase in SCC- PRE (OR = 0.71). Also, cows that had a high percentage of milk fat or milk protein had less chance of having

5 1252 SURIYASATHAPORN ET AL. Figure 3. Odds of case versus control events for SCC at last record before case-control was defined, in a dairy herd between January 1984 and November Figure 5. Odds of case versus control events for percentage of milk protein at last record before case-control was defined, in a dairy herd between January 1984 and November clinical signs of mastitis, but cows with higher milk yields ran more risk of getting clinical mastitis (Table 3). Cows at BCS = 1 (BCS 1 to 1.75) had the highest probability of getting clinical mastitis (OR = 3.09 compared with cows at BCS 3 to 3.75). Both retained placenta and milk fever were associated with increased odds of getting clinical mastitis. Cows with retained placenta had a four-times higher risk to have clinical mastitis, and OR for milk fever cows was Multivariable Analyses Results of four models from four data files are shown in Table 4. The final model of model 1 included lactation period, retained placenta, and milk fever. Cows in early and midlactation periods were at about a 1.5 times higher risk of getting clinical mastitis (P < 0.01). Retained placenta and milk fever increased the risk of getting clinical mastitis. Results of model 2 showed that SCC-PRE and genital infection were significantly associated with severity of mastitis. SCC-PRE was entered in the model with a high value of the likelihood ratio test (LR = , P < ). Cows with higher SCC-PRE were less susceptible to clinical mastitis. Cows with genital infection were about 3.7 times more at risk of getting clinical mastitis. The final model of dataset 3 shows that BCS, lactation period, and milk Figure 4. Odds of case versus control events for milk yields at last record before case-control was defined, in a dairy herd between January 1984 and November Figure 6. Odds of case versus control events for percentage of milk fat at last record before case-control was defined, in a dairy herd between January 1984 and November 1996.

6 LOW SOMATIC CELL COUNT AND CLINICAL MASTITIS 1253 Table 3. Univariable random-effect logistic models for significant risk factors for case and control events in a dairy herd between January 1984 and November Risk 95% confidence Variable Level 1 Coefficient P ratio interval Lactation period (n = 1364) Early 0.64 < Mid 0.44 < Late comparison 1.00 SCC-PRE 2 (n = 926) 0.34 < % milk fat (n = 926) % milk protein (n = 926) 0.71 < Milk yield 3 (n = 926) 0.16 < BCS (n = 962) 1.0 to to to 3.75 comparison to Retained placenta (n = 1364) Milk fever (n = 1364) For categorical variables, each level is compared with the reference value which is the value that has a coefficient equal to SCC measured before the udder inflammatory responses, either case (clinical mastitis) or control (SCC > 200,000 cells/ml), was defined. fever were significantly associated with severity of mastitis. As with results of the univariable model, thin cows (BCS = 1 to 1.75) were at increased risk of clinical mastitis (OR = 3.1, P < 0.01), and cows in early and midlactation had higher risk to get clinical mastitis. The last model that used all variables to create the model (model 4) included only SCC-PRE in the final model. Genital infection had entered in the model with P < 0.10, but it could not stay in the model because the P level was slightly higher than DISCUSSION The BMSCC data show that the herd in our study was a low BMSCC herd (Figure 1). During 13 yr, the average annual BMSCC was lower than 200,000 cells/ ml. Although four observations of BMSCC were higher than 250,000 cells/ml, 90% of cows during this period had an SCC of < 250,000 cells/ml. This indicated that high BMSCC during the specified periods originated from only a few cows. The incidence rate of all cases of Table 4. Summary of multivariable random logistic regression analyses for case and control events of four datasets in a herd between January 1984 and November Risk 95% confidence Variable Level 1 Coefficient P ratio interval Model 1 (n = 1364) Lactation period Early Mid 0.44 < Late comparison 1.00 Milk fever Retained placenta Model 2 (n = 926) SCC-PRE 0.34 < Genital infection Model 3 (n = 962) BCS 1.0 to to to 3.75 comparison to Lactation period Early 0.77 < Mid Late comparison 1.00 Milk fever Model 4 (n = 624) SCC-PRE 0.33 < For categorical variables, each level is compared with the reference value which is the value that has a coefficient equal to 0.00.

7 1254 SURIYASATHAPORN ET AL. clinical mastitis was 0.77 cases per 365 cow-days at risk in this herd and that was high in comparison with rates in previous studies (2, 20, 30). Because this herd has been a high-producing herd (25) with low BMSCC, the high incidence rate of clinical mastitis was not unexpected (2, 19). The predominant microorganism isolated in this herd was E. coli (42.8% of sampled cases). E. coli was reported to be the major cause of clinical mastitis on low SCC farms (2, 19). Clearly, our results are from a single herd with a low prevalence of subclinical mastitis. Therefore, these results need to be interpreted carefully. In the present study, models of UIR were proposed instead of models of log-transformed SCC or clinical mastitis incidence, as in a previous study (17). In the present study, SCC was used as a proxy to define either noninfected or infected state. In general the log transformation of SCC (1) and repeated bacteriological samplings over time are highly recommended methods of identifying the infection status of a quarter (22, 23). An increase in SCC from <200,000 to 200,000 cells/ml was optimal for the prediction of new IMI (6, 17). Results from the univariable analyses showed that SCC-PRE was strongly associated with UIR (Table 3). In addition, results of the multivariable analysis, either model 2 or model 4, showed that SCC-PRE was solely associated with increased risk of clinical mastitis as UIR (Table 4). The probability of clinical mastitis versus increased SCC increased when SCC-PRE decreased (Figure 3). This is in agreement with results of experimental infection studies, which reported that SCC prior to experimental IMI was negatively correlated with severity of mastitis (21, 24, 27, 28). In contrast, high SCC has been established in some studies as a predisposing factor for clinical mastitis (5, 7). However, those studies evaluated all levels of SCC including SCC > cells/ml as a risk factor for incidence of clinical mastitis. In herds in which Staphylococcus aureus or other pathogens that also cause subclinical mastitis play a dominant role, high SCC-PRE may be associated with more clinical mastitis. In this study we specifically focused on very low SCC-PRE levels. We also evaluated the effect of pathogen as a possible confounder for SCC-PRE in relationship to UIR. Results of models with and without E. coli mastitis showed that no differences between OR of both models were observed. Schukken et al. (21) demonstrated that a S. aureus challenge of cows that resisted an infection had a higher preinfection SCC compared with cows that became infected. In experimental E. coli IMI, cows with high SCC before challenge showed a lower severity of mastitis, indicated by the area under the curve of bacterial count (28) or peak bacterial concentration (24). Milk somatic cells are primarily leukocytes, which include macrophages, lymphocytes, and polymorphonuclear leukocytes. A somewhat higher number of SCC- PRE would indicate more leukocytes in milk. A higher number of leukocytes may be able to kill microbes by themselves or initiate inflammatory response better than a low number of leukocytes in SCC-PRE. In an E. coli induced mastitis study, prechallenge exposure milk SCC was correlated with a stimulated CD18 expression, and both measures correlated inversely with bacterial growth rate (24). This may suggest that a very low SCC- PRE indicates a somewhat lower immune efficiency in the mammary gland. Results of univariable analyses showed that milk yield, percentage of milk protein, and percentage of milk fat were strongly associated with the presence of clinical signs of mastitis (Table 3). However, entering of SCC-PRE into the model resulted in loss of significance levels of milk yield, percentage of milk fat, and percentage of milk protein. The cause of this phenomenon may be explained by the power issues, a correlation between SCC and other milk components, or both. We found that SCC was negatively correlated with milk yield, and positively correlated with percentage of milk fat, and percentage of milk protein (P < 0.05). In this study, lactation period was also associated with UIR, but we could not find an effect of parity on UIR. As in previous studies, most of the clinical cases were observed in early lactation (14, 29). We also found that cows with low BCS (BCS = 1 to 1.75) had higher risks of getting clinical mastitis as UIR in comparison with cows with BCS 3 to 3.75 (P < 0.01). BCS is a tool for estimating energy balance status (8), and Kremer and colleagues (13) have shown that cows with negative energy balance from feed restriction had a higher severity of experimental E. coli mastitis, which is not a surprising result. Previously, it was shown that cows with low BCS are more likely to be ketotic (9). In an experimental setting, we showed that ketone bodies decreased the chemotactic function of leukocytes (26), and therefore may put a cow at risk for severe clinical mastitis. Although in this study, we did not measure ketone bodies, the results are in agreement with the mechanism described here. We also found that cows with retained placenta and milk fever also showed increased risk of clinical mastitis (Table 3, Table 4). Schukken et al. (18) also showed that cows with retained placenta were at a higher risk of developing clinical mastitis. Cows with milk fever were at increased risk of retained placenta and clinical mastitis (4). CONCLUSIONS It can be concluded from the present study on a single low BMSCC herd that very low SCC during the nonin-

8 LOW SOMATIC CELL COUNT AND CLINICAL MASTITIS 1255 fectious state (SCC < 200,000 cells/ml) was negatively associated with an increased risk of clinical mastitis as UIR. Results of multivariable analyses also show that low SCC before UIR was the only remaining variable predicting the severity of UIR in this study. As a result of this, very low SCC might be used as the variable predicting the outcome of IMI as either presence or absence of clinical signs of mastitis in herds with low BMSCC. Although some immunological functions of leukocytes have been reported in relation with SCC, complete steps of the defense mechanism remain to be explored further. ACKNOWLEDGMENTS The authors thank J.J.M. Nieuwenhuizen and family for providing the data and study opportunity and H. Tj. Heeringa for preparing the VAMPP records. REFERENCES 1 Ali, A.K.A., and G. E. Shook An optimal transformation for somatic cell concentration in milk. J. Dairy Sci. 63: Barkema, H. W., Y. H. Schukken, T. J. Lam, M. L. Beiboer, H. Wilmink, G. Benedictus, and A. Brand Incidence of clinical mastitis in dairy herds grouped in three categories by bulk milk somatic cell countss. J. Dairy Sci. 81: Bartlett, P. C., G. Y. Miller, C. R. Anderson, and J. H. Kirk Milk production and somatic cell count in Michigan dairy herds. J. Dairy Sci. 73: Curtis, C. R., H. N. Erb, C. J. Sniffen, and R. D. Smith Path analysis of dry period nutrition, postpartum metabolic and reproductive disorders, and mastitis in Holstein cows. J. Dairy Sci.68: Deluyker, H. A., J. M. Gay, and L. D. Weaver Interrelationships of SCC, mastitis, and milk yield in a low somatic cell court herd. J. Dairy Sci. 76: Dohoo, I. R., and K. E. Leslie Evaluation of changes in somatic cell counts as indicators of new intramammary infections. Prev. Vet. Med. 10: Dohoo, I. R., A. H. Meek, and S. W. Martin Somatic cell counts in bovine milk: relationships to production and clinical episodes of mastitis. Can. J. Comp. Med. 48: Edmonson, A. J., I. J. Lean, L. D. Weaver, T. Farver, and G. Webster A body condition scoring chart of Holstein dairy cows. J. Dairy Sci. 72: Heuer, C., Y. H. Schukken, P. Dobbelaar Postpartum body condition score and results from the first test day milk as predictors of disease, fertility, yield, and culling in commercial dairy herds. J. Dairy Sci. 82: Hu, F. B., J. Goldberg, D. Hedeker, B. R. Flay, and M. A. Pentz Comparison of population-averaged and subject-specific approaches for analyzing repeated binary outcomes. Am. J. Epidemiol. 147: Lam, T. J.G.M., J. H. Van Vliet, Y. H. Schukken, F. J. Grommers, A. Van Velden-Russcher, H. W. Barkema, and A. Brand The effect of discontinuation of postmilking teat disinfection in low SCC herds. 1. Incidence of clinical mastitis. Vet. Quart. 19: Kehrli, M. E., and D. E. Shuster Factors affecting milk somatic cells and their role in health of the bovine mammary gland. J. Dairy Sci. 77: Kremer, W.D.J., E. N. NoordhuizenStassen, F. J. Grommers, Y. H. Schukken, R. Heeringa, A. Brand, and C. Burvenich Severity of experimental Escherichia coli mastitis in ketonemic and nonketonemic dairy cows. J. Dairy Sci. 76: Miltenburg, J. D., D. de Lange, A. P. Crauwels, J. H. Bongers, M. J. Tielen, Y. H. Schukken, and A. R. Elbers Incidence of clinical mastitis in a random sample of dairy herds in the southern Netherlands. Vet. Rec. 139: Neave, F. K., F. H. Dodd, R. G. Kingwill, and D. R. Westgarth Control of mastitis in the dairy herd by hygiene and management. J. Dairy Sci. 52: SAS Institute, Inc SAS/STAT software: changes and enhancements through release Cary, NC. 17 Schepers, A. J., T. J. Lam, Y. H. Schukken, J. B. Wilmink, and W. J. Hanekamp Estimation of variance components for somatic cell counts to determine thresholds for uninfected quarters. J. Dairy Sci. 80: Schukken, Y. H., H. N. Erb, and R. D. Smith The relationship between mastitis and retained placenta in a commercial population of Holstein dairy cows. Prev. Vet. Med. 5: Schukken, Y. H., F. J. Grommers, D. v. d. Geer, and A. Brand Incidence of clinical mastitis on farms with low SCCs in bulk milk. Vet. Rec. 125: Schukken, Y. H., F. J. Grommers, D. v. d. Geer, H. N. Erb, and A. Brand Risk factors for clinical mastitis in herds with a low bulk milk SCC. 1. Data and risk factors for all cases. J. Dairy. Sci. 73: Schukken, Y. H., B. A. Mallard, J. C. M. Dekkers, K. E. Leslie, and M. J. Stear Genetic impact on the risk of intramammary infection following Staphylococcus aureus challenge. J. Dairy Sci. 77: Sears, P. M., B. S. Smith, P. B. English, P. S. Herer, and R. N. Gonzalez Shedding pattern of Staphylococcus aureus from bovine intramammary infections. J. Dairy Sci. 73: Sears, P. M., D. J. Wilson, R. N. Gonzalez, and D. D. Hancock Microbiological results from milk samples obtained premilking and postmilking for the diagnosis of bovine intramammary infections. J. Dairy Sci. 74: Shuster, D. E., E. K. Lee, and M. E. Kehrli, Jr Bacterial growth, inflammatory cytokine production, and neutrophil recruitment during coliform mastitis in cows within tend after calving, compared with cows at midlactation. Am. J. Vet. Res. 57: Suriyasathaporn, W., M. Nielen, S. J. Dieleman, A. Brand, E. N. Noordhuizen-Stassen, and Y. H. Schukken A Cox proportional hazards model with timedependent covariates to evaluate the relationship between bodycondition score and the risks of first insemination and pregnancy in a highproducing dairy herd. Prev. Vet. Med. 37: Suriyasathaporn, W, A.J.J.M. Daemen, E. N. NoordhuizenStassen, M. Nielen, S. J. Dieleman, and Y. H. Schukken. _Hydroxybutyrate levels in peripheral blood and ketone bodies supplemented in culture media affect the in vitro chemotaxis of bovine leukocytes. Vet. Immunol. Immunopathol. (in press) 27 van Werven, T., E. N. Noordhuizen-Stassen, A.J.J.M. Daemen, Y. H. Schukken, and C. Burvenich Preinfection in vitro chemotaxis, phagocytosis, oxidative burst, and expression of CD11/CD18 receptors and their predictive capacity on the outcome of mastitis induced in dairy cows with Escherichia coli. J. Dairy Sci. 80: van Werven, T The role of leukocytes in bovine Escherichia coli mastitis. Ph.D. thesis. University Utrecht. 29 Weller, J. I., A. Saran, and Y. Zeliger Genetic and environmental relationships among SCC, bacterial infection, and clinical mastitis. J. Dairy Sci. 75: Wilesmith J. W., P. G. Francis, and C. D. Wilson Incidence of clinical mastitis in a cohort of British dairy herds. Vet. Rec. 118: