clinical mastitis in primiparous Holstein cows

Transcription

1 Predictive variables for the occurrence of early clinical mastitis in primiparous Holstein cows under field conditions in France Jacques Barnouin, Michelle Chassagne Abstract Holstein heifers from 47 dairy herds in France were enrolled in a field study to determine predictors for clinical mastitis within the first month of lactation. Precalving and calving variables (biochemical, hematological, hygienic, and disease indicators) were collected. Early clinical mastitis (ECM) predictive variables were analyzed by using a multiple logistic regression model (99 cows with ECM vs. 571 without clinical mastitis throughout the first lactation). Two variables were associated with a higher risk of ECM: a) difficult calving and b) medium and high white blood cell (WBC) counts in late gestation. Two prepartum indicators were associated with a lower ECM risk: a) medium and high serum concentrations of immunoglobulin GI (IgGI) and b) high percentage of eosinophils among white blood cells. Calving difficulty and certain biological blood parameters (IgGI, eosinophils) could represent predictors that would merit further experimental studies, with the aim of designing programs for reducing the risk of clinical mastitis in the first lactation. Resume Facteurs previsionnels de l'eventualite de mammites cliniques precoces chez des vaches Holstein primipares gardees au paturage en France. Des genisses Holstein provenant de 47 troupeaux laitieres de France ont ete inclues dans une etude sur le terrain afin de determiner les facteurs previsionnels de la mammite clinique survenant dans le premier mois de lactation. Des donnees de pre-velage et de velage (biochimiques, hematologiques, hygienmques et indicatrices de maladies) ont ete recueillies. Les facteurs previsionnels des mammites cliniques precoces (MCP) ont ete analyses selon un modele de regression logistique multiple (99 vaches avec MCP et 571 sans mammites cliniques tout au long de leur premiere lactation). Deux facteurs ont ete associes a un risque plus eleve de MCP: a) un velage difficile; b) compte moyen ou eleve de globules blancs en fin de gestation. Deux indicateurs prenataux ont ete associes avec un risque moins eleve de MCP: a) des concentrations seriques moyennes ou elevees de immunoglobulines GI; b) un pourcentage eleve d'eosinophiles. Des difficultes de velage et certains parametres biologiques sanguins (immunoglobulines GI, eosinophiles) pourraient etre des facteurs previsionnels necessitant des etudes experimentales additionnelles afin d'etablir des programmes de reduction des risques de mammites cliniques lors de la premiere lactation. (Traduit par Docteur Andre' Blouin) Can Vet J 2000;41:47-53 Introduction Looking for prepartum predictive variables for postpartum disease occurrence under field conditions is a meaningful, but difficult, objective. In commercial herds, the collection of biological samples to determine the value of markers that vary according to time and disease stage is especially difficult with reference to practical and ethical limitations, as well as the impossibility of predicting the calving time in advance. Nevertheless, cooperation between epidemiological and mechanistic mastitis research has great potential. Risk indicators obtained in observational studies from a sample of commercial herds can be evaluated in detail in pathogenetic studies conducted in an experimental herd (1). With reference to such strategies, improving biological Unite d'epidemiologie Animale, INRA, Centre de Recherches de Clermont-Ferrand-Theix, Saint Genes Champanelle, France. Address correspondence and reprint requests to Dr. Jacques Barnouin. relevance of the disease predictors is a basic need for the epidemiologist, despite feasibility problems. Elaborate statistical models from computer intensive methods are also necessary in epidemiological studies (2) in order to analyze the numerous factors causing variation, as well as the effects of confounding variables (year, age, herd) on disease occurrence (3,4). Mastitis is a worldwide multifactorial disease problem (5) that decreases the profitability of the dairy cow through discarded milk of mastitic cows, lowered milk production, treatment costs, and losses due to involuntary culling. The incidence rate of clinical mastitis has been reported to be associated with feeding, housing, milking machine, hygienic, and milk yield factors (6,7). Clinical mastitis has a peak of maximum occurrence in early lactation, particularly in dairy production systems in France (8). The first peripartum period represents the primary critical epidemiological period for clinical mastitis in the dairy cow. Consequently, mastitis in the dairy heifer is a significant field of research (9-11). Early lactation primiparous females have not had extended exposure to contagious pathogens transported Can Vet J Volume 42, January

2 primarily by the milking machine. However, prepartum intramammary infections can persist into early lactation (12). Most intramammary infections in the dairy heifer are due to Staphylococcus spp., mainly coagulase-negative staphylococci, and Streptococcus spp., mainly S. uberis (9,13-15). Mastitis in first lactation can be a common occurrence in well-managed, productive herds with relatively low somatic cell counts (SCCs), and can result in a high frequency of treatment for mastitis by veterinarians (16). Few multivariate analyses have been conducted with data from commercial dairy herds to examine the combined effects of the individual precalving parameters, nutritional and environmental conditions, and the calving conditions on the risk of clinical mastitis after calving. The aim of the present field study was to investigate the association between variables, focusing on calving parameters, biochemical measurements, and immunological markers measured in late gestation that were hypothesized to predict occurrence of clinical mastitis within the first month of the first lactation in Holstein dairy heifers under breeding management, as practised in France. The potential predictive variables were analyzed concurrently by using a multiple logistic regression model while controlling for the effects of calving year, calving season, and herd. Materials and methods Survey and data collection A prospective epidemiological survey was carried out during a 4-year period at commercial dairy farms (n = 47) located in France to investigate herd and individual risk factors for clinical peripartum diseases, especially mastitis. The general study design was as published previously (17). All the farmers were members of Dairy Herd Improvement Associations (DHIA) and volunteered to participate in the survey. They were selected on their ability to detect diseases efficiently, as perceived by their veterinarian. All the cows in the survey were loose-housed and had not been vaccinated against pathogens causing mastitis. Each herd was tested for and found to be free of brucellosis and tuberculosis. Holstein was the main breed (98% of the cows). Milk yield averaged between 6000 and kg/cow/y. Designated surveyors visited the farms once every 6 wk to collect or measure individual data and samples. Data analyzed in the present study were cases of clinical mastitis within the first 30 d of lactation (early clinical mastitis, ECM), body condition scores, cleanliness scores, concentrate feeding in late gestation, biochemical and hematological measurements, calving variables, and early lactation milk parameters. Clinical mastitis was characterized by udder inflammation or abnormal milk, with or without general clinical signs. Participating farmers were given training on clinical mastitis to enable them to identify cases that were not observed by the veterinarian. The farmers would not agree to prepartum intramammary secretions from healthy heifers being systematically collected for bacteriological analysis. Moreover, the inability of dairymen to collect milk samples adequately suggested that it was not feasible to examine the milk bacteriologically 48 when mastitis was detected. Four-quarter composite milk samples were collected by the surveyors only when they visited the farms routinely in a subset of cows within the first 6 wk of lactation. Milk samples for bacteriology were immediately put in an ice cooler until they were analyzed, within 24 h after collection. In addition, DHIA supplied milk production data concerning DHIA monthly milk tests of the first lactation. The quantities of concentrates supplied to the heifers during the late gestation period were recorded. Body condition and cleanliness scores were measured once within the last 60 d of gestation. Blood was collected once within the last 60 d of gestation, between 0900 and 1000, to control for diurnal fluctuations in circulating biochemical and hematological parameters. Heparinized plasma samples were kept frozen until they were analyzed. Whole blood samples were maintained at 5 C until being analyzed, within 5 h after collection. Calving variables included length of gestation; calving difficulty, defined as a calving requiring the farmer's or veterinarian's assistance; number of calves born; retained placenta, or not; perinatal calf mortality; and calving season. Calving season was identified as winter (December, January, February), spring (March, April, May), summer (June, July, August), and fall (September, October, November). Selection of animals In order to standardize the study population with respect to breed and to physiological and health factors, all the primiparous cows were selected by using the following criteria: 1) Holstein breed; 2) calving a single calf; 3) blood sample collected within the last 2 mo before calving; 4) body condition score and dirtiness score assigned within the last 2 mo of gestation; 5) complete set of data concerning gestation length, precalving feeding, blood indicators, and calving descriptors; 6) absence of clinical mastitis or other udder disease prior to the first calving. Primiparous females were divided in 2 groups. The 1st group included the first lactation cows with ECM (ECM+) and the 2nd group included the cows without clinical mastitis throughout the first lactation (CM-). Consequently, first lactation females with a first clinical mastitis occurrence over 30 d in milk (n = 52) were not selected in the study. All CM- females from herds without any selected ECM+ females were removed from the statistical analysis to satisfy the requirements for the logistic regression procedure with adjustment for fixed herd effect. Variables studied The variables tested for capability to predict occurrence of ECM aimed to describe the management, nutritional, inflammatory, and immunological status in the precalving heifer. Biochemical predictors were circulating levels of glucose, 1-hydroxybutyrate, free fatty acids, urea, free amino acids, albumin, calcium, magnesium, y-glutamyl transferase, glutamate dehydrogenase, ceruloplasmin oxidase activity, and IgGIs. Hematological parameters were red blood cell counts, white blood cell (WBC) counts and their differential leucocyte counts, packed cell volume, and hemoglobin. All circulating Can Vet J Volume 42, January 2001

3 Table l.variables tested for capability to predict occurrence of early clinical mastitis (ECM) in primiparous Holstein cows. All predicting data were collected within the last 60 days of gestation or at calving (ECM+: females with early clinical mastitis; CM-: females without clinical mastitis throughout the lactation) Overall (n = 670) ECM+ (n = 99) CM- (n = 571) Variable Mean (s) or incidence(*) Mean (s) or incidence(*) Mean (s) or incidence(*) Nutrition (kg/cow: last 2 mo of gestation) Cereal concentrate: total quantity Soybean meal: total quantity Reproduction Age at first calving (d) Birth of a dead calf (% of cows) Calving difficulty (% of cows) Calving season: winter (% of calvings) Calving season: spring (% of calvings) Calving season: summer (% of calvings) Calving season: autumn (% of calvings) Gestation length (d) Milk fever (% of cows) Retained placenta (% of cows) Interval Scoring/blood collection-calving (d) Scores Body condition score (0-5 points) Dirtiness score (0-8 points) Biochemical markers 3-hydroxybutyrate (mmol/l) Albumin (g/l) Calcium (mmol/l) Ceruloplasmin (U/L) Free amino acids (mmol/l) Free fatty acids (mmol/l) y-glutamyl transferase (U/L) Glutamate dehydrogenase (U/L) Glucose (mmol/l) IgGl (g/l) Magnesium (mmol/l) Urea (mmol/l) Hematological markers WBC (109/L) Neutrophils (% WBC) Lymphocytes (% WBC) Basophils (% WBC) Eosinophils (% WBC) Monocytes (% WBC) Red blood cells (1012/L) Packed cell volume (L/L) Hemoglobin (g/l) s - standard deviation; WBC - white blood cells 166 (178) 126 (170) 896 (114) 6.4* 13.4* 36.2* 19.5* 12.2* 32.1 * 281 (5) 0.001* 8.2* 26.7 (14.9) 3.62 (0.45) 2.06 (0.13) 0.55 (0.29) 36.4 (2.7) 2.44 (0.16) 52.9 (5.1) 2.47 (0.31) 0.54 (0.43) 11.7 (4.1) 6.53 (5.80) 3.67 (0.42) 16.0 (3.9) 0.59 (0.06) 4.27 (1.78) 8.62 (2.16) 36.3 (14.0) 52.2 (15.0) 0.03 (0.19) 8.00 (5.72) 3.47 (2.04) 6.68 (0.80) 0.35 (0.03) 122 (13.9) 169 (191) 139 (185) 887 (112) 7.1 * 19.2* 38.4* 19.2* 8.1* 34.3* 281 (6) 0.0* 9.1 * 26.7 (15.0) 3.64 (0.48) 1.91 (0.12) 0.54 (0.21) 36.6 (2.7) 2.44 (0.13) 52.4 (4.4) 2.49 (0.31) 0.53 (0.36) 11.9 (5.1) 6.10 (3.42) 3.67 (0.33) 15.5 (4.6) 0.60 (0.05) 4.23 (1.75) 8.83 (1.83) 36.5 (14.3) 53.4 (15.2) 0.03 (0.19) 6.88 (5.41) 3.19 (1.75) 6.78 (0.85) 0.34 (0.03) 121 (10.8) 166 (175) 124 (167) 898 (115) 6.3* 12.4* 35.7* 19.6* 13.0* 31.7* 281 (5) 0.2* 8.1 * 26.6 (14.9) 3.62 (0.45) 2.08 (0.13) 0.56 (0.31) 36.4 (2.7) 2.44 (0.17) 53.0 (5.2) 2.46 (0.31) 0.54 (0.45) 11.6 (4.0) 6.60 (6.11) 3.67 (0.44) 16.1 (3.8) 0.59 (0.06) 4.28 (1.78) 8.58 (2.21) 36.3 (14.0) 51.9 (15.0) 0.04 (0.20) 3.53 (2.08) 8.23 (5.75) 6.67 (0.79) 0.35 (0.03) 122 (14.3) predictors were determined by using previously published methods (18,19) and, generally, could be measured in a field laboratory with low cost methods. Body condition and dirtiness scores were determined just before blood collection, as described by Bazin (20) and Faye and Barnouin (21). The milk parameters used to measure the consequences of ECM included milk yield, milk fat content, milk protein content, and SCC at the first milk test of the current lactation (or at the first milk test after ECM for ECM+ cows). Fluoro-opto-electronic cell counting (22) was used to measure the milk SCC. Statistical analyses The cow was the unit of study. The dependent variable was the presence or absence of ECM in the selected females. Statistical analyses were run by using a statistical software (Statistical Analysis Systems; SAS Institute, Cary, North Carolina, USA). The procedures Can Vet J Volume 42, January 2001 performed to compute descriptive data and compare milk variables of the first lactation between ECM+ and ECM- animals were NPAR1WAY and FREQ. The ECM potential predictive variables were analyzed in 2 steps. In a first univariate step, continuous variables were classified according to quartiles (lst quartile, 2nd-3rd quartiles, 4th quartile) and frequency distributions were calculated for all predictive variables from the UNIVARIATE procedure of SAS. When distributions between groups with and without the occurrence of ECM differed at P < 0.25, the variables were considered as potential candidates for further multiple logistic regressions. Regarding their biological importance, 4 key variables (calving year, calving season, blood collectionto-calving interval, herd) were added to the logistic models, without consideration of their contribution, in order to control for possible confounding and to increase statistical validity (3). The blood collection-to-calving 49

4 Table 2. Milk yield variables used to measure early clinical mastitis (ECM) consequences at the first milk test (or at the first test after ECM) in primiparous Holstein cows (ECM+: females with early clinical mastitis; CM-: females without clinical mastitis throughout the lactation) Overall (n = 670) ECM+ (n = 99) CM-(n = 571) Variable Mean (s) Mean (s) Mean (s) Daily milk yield (kg) 23.9 (4.3) 23.1 (5.1) 24.0 (4.2) Fat content (g/l) 41.8 (6.0) 41.1 (6.3) 41.9 (6.0) Protein content (g/l) 29.6 (3.0) 29.8 (3.2) 29.6 (3.0) Somatic cell count (103/mL) 219 (577) 392 (1065) 189 (435) s = standard deviation Table 3. Distributiona and univariate analysis of the risk factors offered to the final regression model for primiparous Holstein cows with and without early clinical mastitis Not diseased Diseased Risk factor Level % (n) % (n) P-Value Difficult calving No 86.2 (500) 13.8 (80) 0.07b Yes 78.9(71) 21.1 (19) Glucose (mmol/l) < (140) 11.4 (18) 0. lb (316) 17.3 (66) > (115) 11.5 (15) Mg (mol/l) < (147) 10.4 (17) 0.18b (296) 15.9 (56) > (128) 16.9 (26) IgG I (g/l) < (132) 21.0 (35) 0.02c (290) 14.0 (47) > (149) 10.2 (17) WBC (109/L) < (141) 8.4 (13) 0.02c (298) 15.6(55) > (132) 19.0 (31) Eosinophils (% WBC) < (128) 17.4 (27) 0.06b (301) 16.2 (58) > (142) 9.0 (14) Monocytes (% WBC) < (104) 16.1 (20) 0.07b (369) 16.1 (71) > (98) 7.5 (8) adistributions of the confounding variables are not shown bp-value < 0.25 CP-value < 0.05 interval was aimed to control for physiological changes in circulating metabolites in late gestation. Model building employed a stepwise algorithm that assessed multicolinearity and 2-way interaction at each step. The criterion for inclusion in the models was set at P = 0.25 (3), while removal was assessed at P > Assessment of how the models fitted the data was determined by using the goodness-of-fit test (23). Results Descriptive data Ninety-nine ECM+ and 571 CM- primiparous Holstein females were studied from 33 herds. The subset of females with a milk sample for bacteriology comprised 61 ECM+ and 183 CM- animals. Tables 1 and 2 provide summary results for predictive variables and early lactation milk parameters. The mean interval between blood collection and calving was 26.7 d (range, 1 to 60 d) and did not statistically differ according to group (the mean value was 26.7 d in the ECM+ group vs d in the CM- group). In the so ECM+ group, 30.3% of the females were collected between 1 and 15 d before calving (bc), 26.3% between 16 and 30 d bc, 32.3% between 31 and 45 d bc, and 11.1% between 46 and 60 d bc. In the CM- group, the respective percentages were 27.2%, 33.6%, 27.0%, and 12.2%. In the ECM+ group, the average interval between calving and ECM was 7.2 d; 67.7% of the ECM cases occurred between 0 and 7 d, 10.1% between 8 and 14 d, 11.1% between 15 and 21 d, and 11.1% between 22 and 30 d. Thirteen ECM+ females experienced a 2nd clinical mastitis within the 1st lactation (recurrence rate = 13.1%). From the study population, 65.6% of clinical mastitis occurred in the primiparous cow within the 1st month of lactation, 6.0% occurred within the 2nd month, 4% within the 3rd month, 6% within the 4th month, and 18.4% from the 5th month to the end of the lactation. In early lactation, the SCC was higher in ECM+ (392 X 103 cells/ml) than in ECM- females (189 X 103 cells/ml) (linear scores SCC: 5.04, s = 1.19 vs. 4.42, s = 1.09, P < ), while daily milk yield, milk fat Can Vet J Volume 42, January 2001

5 Table 4. Final logistic regression modela for early clinical mastitis (ECM) occurrence in primiparous Holstein cows 95% Confidence Risk factor Level Odds ratio (OR) interval for the OR Difficult calving No 1 Yes 2.20b IgGI (g/l) < > c WBC (109/L) < lb > c Eosinophils (% of WBC) < > c aherd, blood collection-to calving interval, calving year and calving season herd effects were adjusted for in all analyses, but the results concerning these confounding variables are not shown. bp-value < 0.05 CP-value < 0.01 content, and milk protein content did not differ between the 2 groups of primiparous cows. Moreover, microorganisms were isolated in 57.4% of ECM+ milk samples vs. 7. 1% of CM- samples. In ECM+ cows with a milk bacteriological analysis, the most prevalent microorganisms were coagulase-negative staphylococci (21.3%), Streptococcus uberis (21.3%), and Escherichia coli (4.9%). In CM- cows, the most prevalent milk bacteria were coagulase-negative staphylococci (3.9%), Streptococcus uberis (1.7%), and Bacillus spp. (1.7%). Univariate analysis Table 3 summarizes the selected variables offered to multiple logistic regression. From univariate analysis, ECM cases were more frequent when the circulating IgGI concentrations were within the lowest values (P = 0.02) and when the WBC counts were within the highest values (P = 0.02). Multivariate analysis In the final logistic regression (score = , P = 0.016, 47 df) (Table 4), after control by calving year, calving season, blood collection-calving interval, and herd, 2 variables were significant risk factors for ECM in the primiparous Holstein: difficult calving [OR = 2.20, P = 0.018], as well as medium (OR = 2.1 1, P = 0.034) and high (OR = 3.05, P = 0.005) WBC counts in the late gestation. Moreover, 2 prepartum biological indicators were significantly associated with a lower ECM risk: medium (OR = 0.49, P = 0.014) and high (OR = 0.30, P = 0.001) serum IgGI concentrations and a high percentage of eosinophils among WBC (OR = 0.34, P = 0.008). No significant interaction was found between the significant variables. The goodness-of-fit test found that the data fit the model well (P = 0.98). Discussion In the present study, 67.7% of ECM cases occurred within the 1st week of lactation. This finding is consistent with a previous survey conducted in France, in which 63.2% of the cases were observed in the 1st 7 d after calving (8). Higher SCC in ECM+ cows agrees with previous results in which mean SCCs at first milk control were 359 X 103 cells/ml for mastitic primiparous females and 202 X 103 cells/ml for control cows, respectively (16). Although the SCC was higher in ECM+ females, milk yield did not differ between the ECM+ and CM- groups. Consequently, mastitic primiparous females could have higher genetic potential for milk production (16), as SCC was negatively correlated with herd and cow milk yield (7,24). A positive association between difficult calving and clinical mastitis during the first lactation period had been previously demonstrated in primiparous Swedish cattle (25). In French dairy herds, difficult calving was associated with clinical mastitis, but only when females in 1 st, 2nd, and 3rd lactations were considered together (26). Although the reasons for the dystocias in the present study are not known, 4 etiological mechanisms could favor ECM occurrence with dystocia in primiparous cows: 1) infection of the udder from uterine lesions related to calving management and bacteremia via intestinal bacteria, especially coliforms (26); 2) poor hygiene at calving, associated with an increased contact time with the bedding due to the dystocia, and the subsequent higher risk of udder infection; 3) impairment of immune function related to more serious negative energy and protein imbalances via a decreased appetite in cows with dystocia (27); and 4) a stress-induced increase of plasma cortisol via dystocia in primiparous females (susceptible to stressful conditions), with a subsequent decrease in phagocytosis (28-30). Milk bacteriological status was not determined at mastitis occurrence in the present study. Thus, it is not possible to determine if the first mechanism was the major one. However, coliform intramammary infections are generally infrequent in the heifer (13,14). Concerning the 2nd and the 3rd hypotheses, mean dirtiness scores were not different in late gestation between the 2 groups, and nutritional parameters (biochemical markers and body condition score) did not discriminate between ECM + and CM - animals. However, these measurements were not made at calving. A stress-dependent alteration in the immune responsiveness of the primiparous females with dystocia could be a significant detrimental factor regarding the Can Vet J Volume 42, January

6 risk of mastitis, although we did not estimate the level of stress or the level of immune suppression. We did not have a precalving measure of intramammary infection, so it is possible that subclinical infections could have occurred before blood collection and have had an impact on biochemical markers. But this is doubtful, since the mean IgGI concentrations were precisely lower in ECM+ heifers than in control animals (without clinical mastitis and with low SCC in the early lactation). Potential udder infections in ECM + heifers after blood collection could have contributed to increase (not decrease) IgGI concentrations as, in several previous studies, intramammary infections or immunizations with major pathogens increased the IgGl antibody titer or concentration (31-34). Moreover, plasma ceruloplasmin oxidase activity did not differ between ECM + and CM - females and was within the normal values in both groups. Ceruloplasmin is a marker of inflammation, which, in one study, was highly correlated with circulating copper (35). As serum copper and milk SCC were positively correlated in the dairy cow (36), it can be hypothesized that udder inflammation was not present in both ECM+ and CM- heifers. Nevertheless, infections with minor pathogens could have occurred in some animals. It is not expected that these infections would have had significant biological effects, although additional research is required for a better understanding of their role in udder immune defense (37,38). In the present work, high plasma IgGI concentrations in late gestation were significantly associated with a lower ECM risk in the subsequent lactation. Immunoglobulins are components of the humoral defense against intramammary infections (39), and IgGI is a major immunoglobulin in mammary gland secretions during lactation (40). Previous studies have revealed that circulating IgG concentrations decreased at calving and that low values or low antibody responses around calving were associated with a higher risk of clinical mastitis (16,41,42). Moreover, calves of dams diagnosed with early mastitis had lower mean serum IgG concentrations at 10 h after birth (43). In a field trial with an experimental vaccine against Staphylococcus aureus, milk concentration of IgGl, but not IgG2, was increased in the vaccinated cows (33). Cows with low serum IgGI titers recognizing Escherichia coli experienced a 5-fold increase in the risk of coliform mastitis (44). Phenotypic and genetic associations were reported between the mastitis measures, the assays that measure the innate capability of neutrophils to respond to an infection, and the circulating immunoglobulins (45). Consequently, a standardized serum IgGI test might be useful to predict clinical mastitis in the heifer. However, further experimental studies should be conducted to estimate the specificity and predictive value of such a test in field conditions. A high WBC count before calving was a risk factor for ECM. Furthermore, a percentage of eosinophils among the WBC higher than 11 % was associated with a decreased risk of ECM (average % of eosinophils: 8.2 in CM- cows and 6.9 in ECM+ cows). There are very few studies about prepartum leukocyte counts and functions before ECM (27). Bosinophils have a great importance on the respiratory burst activity of polymorphonuclear leukocytes (46), which partially supports bactericidal activity of phagocytic cells. Bovine eosinophils produce much higher amounts of reactive oxygen species than do polymorphonuclear leukocytes under basal conditions and after stimulation (47). A correlation between eosinophils and udder infection-related inflammatory markers (PGE2) was recently demonstrated, but its significance is uncertain (48). In conclusion, whether hormone and immune system fluctuations, and interrelationships between them, during the periparturient period are responsible for ECM remains unclear, despite considerable research. Our field study points to health and circulating blood parameters that represent potential etiological pathways for further focused studies. Such investigations could determine the significance of the present results under experimental conditions, including kinetic biochemical and bacteriological prepartum analyses, which were not always feasible in a survey conducted in commercial dairy herds. Nevertheless, according to our epidemiological study, difficult calving increases the risk of mastitis within the first month of lactation in the primiparous Holstein. Consequently, ECM preventive measures aimed at lowering the risk of difficult calving in heifers via appropriate genetic measures, as well as controlling stressful conditions and ensuring a high standard of hygiene at parturition, particularly when a dystocia occurs, should reduce the risk of ECM. cvj References 1. Van Werven T, Schukken HY, Noorduizhen-Stassen E. Epidemiology and prediction of severity of Escherichia coli mastitis. Flem Vet J 1997: Richardson S. Developpements recents de la biostatistique. Rev Epidemiol Sante Publique 1996;44: Bouyer M. La regression logistique en 6pidemiologie. Partie II. Rev Epidem Sante Publ 1990;39: McDermott JJ, Schukken YH. A review of methods used to adjust for cluster effects in explanatory epidemiological studies of animal populations. Prev Vet Med 1994;18: Barnouin J, Fayet JC, Brochart M. Enquete 6copathologique continue: facteurs de risque des mammites de la vache laitiere I. Analyses multidimensionnelles sur donnees d'elevage. Can Vet J 1986;27: Barkema HW, Schukken YH, Lam TJ, Beiboer ML, Benedictus G, Brand A. Management practices associated with the incidence rate of clinical mastitis. J Dairy Sci 1999;82: Barnouin J, Geromegnace N, Chassagne M, Dorr N, Sabatier P. Facteurs structurels de variation des niveaux de comptage cellulaire et de frequence des mammites cliniques dans 560 6levages bovins repartis dans 21 d6partements fran,ais. INRA Prod Anim 1999;12: Faye B, Fayet JC. Enquete Ecopathologique Continue. II. Evolution des frequences pathologiques en 6levage bovin laitier en fonction du stade de lactation. Ann Rech Vet 1986;17: Aarestrup FM, Jensen NE. Prevalence and duration of intramammary infection in Danish heifers during the peripartum period. J Dairy Sci 1997;80: Robertson JR, Fox LK, Hancock DD, Gay J, Besser TE. Sources of intramammary infections from Staphylococcus aureus in dairy heifers at first parturition. J Dairy Sci 1998;81: Oliver SP, Lewis MJ, Gillepsie BE, Dowlen HH. Antibiotic residues and prevalence of mastitis pathogen isolation in heifers during early lactation following prepartum antibiotic therapy. J Vet Med B 1997;4: Oliver SP, Sordillo LM. Udder health in the periparturient period. J Dairy Sci 1988;71: Da Costa EQ, Melville PA, Ribeiro AR, Watanabe E, Viani FC, White CR. Prevalence of intramammary infections in primigravid Brazilian dairy heifers. Prey Vet Med 1996;29: Can Vet J Volume 42, January 2001

7 14. Matthews KR, Harmon RJ, Langlois BE. Prevalence of Staphylococcus species during the periparturient period in primiparous and multiparous cows. J Dairy Sci 1992;75: Pankey JW, Pankey PB, Barker R, Williamson JH, Woolford MW. The prevalence of mastitis in primiparous heifers in eleven Waikato dairy herds. N Z Vet J 1996;44: Myllis V, Rautala H. Characterization of clinical mastitis in primiparous heifers. J Dairy Sci 1995;78: Barnouin J, Chassagne M. Herd factors associated with clinical mastitis incidence in French dairy herds during late gestation and early lactation. Vet Res 1998;29: Aissaoui C. Enzymes circulantes (OCT, GLDH, GGT) a localisation h6patique chez la vache laitiere. Facteurs de variation physiologiques, nutritionnels et relation avec la pathologie [These de doctorat], Montpellier, France: Universit6 de Montpellier II, Barnouin J, Chacornac JP, Aissaoui C, El Idilbi N, Mazur A. Comment d6pister les desequilibres biologiques et les troubles de sante chez la vache laitiere dans le cadre d'etudes ecopathologiques? Vet Res 1994;25: Bazin S. Grille de notation de l'etat d'engraissement des vaches pie-noires, le 6d., Paris: Institut Technique de l'elevage Bovin, 1984: 31p. 21. Faye B, Barnouin J. Objectivation de la proprete des vaches laitieres et des stabulations. L'indice de propret6. Bull Tech du Centre de Recherches Zootechniques et Vet6rinaires de Theix 1985;59: Schmid Madsen P. Fluoro-opto-electronic cell-counting on milk. J Dairy Sci 1975;42: Hosmer DW, Lemeshow S. Applied Logistic Regression. New York: Wiley Interscience, 1989: 307p. 24. Bartlett PC, Miller GY, Anderson CR, Kirk JH. Milk production and somatic cell count in Michigan dairy herds. J Dairy Sci 1990;73: Oltenacu PA, Frick A, Lindhe B. Epidemiological study of several clinical diseases, reproductive performance and culling in primiparous Swedish cattle. Prev Vet Med 1990;9: Faye B, Fayet JC, Brochart M, Barnouin J, Paccard P. Enquete ecopathologique continue: 5. Mise en evidence des associations pathologiques en elevage bovin laitier: donn6es individuelles. Ann Rech Vet 1986;17: Kehrli ME, Nonnecke BJ, Roth JA. Alterations in bovine neutrophil function during the peripartum period. Am J Vet Res 1989;50: Cai TQ, Weston PG, Lund LA, Brodie B, McKenna DJ, Wagner WC. Association between neutrophil functions and periparturient disorders in cows. Am J Vet Res 1994;55: Mallard BA, Dekkers JC, Ireland MJ, et al. Alteration in immune responsiveness during the peripartum period and its ramification on dairy cow and calf health. J Dairy Sci 1998;81: Persson K, Carlsson A, Hambleton C, Guidry AJ. Immunoglobulins, lysozyme and lactoferrin in the teat and udder of the dry cow during endotoxin-induced inflammation. J Vet Med B 1992;39: Caffin JP, Poutrel B. Physiological and pathological factors influencing bovine immunoglobulin concentration in milk. J Dairy Sci 1988;71 : Nasher TO, Stokes CR, Cripps PJ. Immune response to intramammary infusion with soluble (ovalbumin) and particulate (S. uberis) antigens in the preparturient bovine udder. Res Vet Sci 1991;50: Nordhaug ML, Nesse LL, Norcross NL, Gudding R. A field trial with an experimental vaccine against Staphylococcus aureus mastitis in cattle. 2. Antibody response. J Dairy Sci 1994;77: Tomita GM, Nickerson SC, Owens WE, Wren B. Influence of route of vaccine administration against experimental intramammary infection caused by Escherichia coli. J Dairy Sci 1998;81: Chacornac JP, Barnouin J, Raboisson T. Micro-dosage automatise de la c6ruloplasmine plasmatique par mesure de l'activite oxydasique chez les bovins et les ovins. Reprod Nutr Dev 1986;26: Kellog DW, Rakes JM, Gliedt DW. Effect of zinc methionine supplementation on performance and selected blood parameters of lactating dairy cows. Nutr Rep Intern 1989;40: Kirk J, Wright JC, Berry SL, Reynolds JP, Maas JP, Ahmadi A. Relationships of milk culture status at calving with somatic cell counts and milk production of dairy heifers during early lactation on a Californian dairy. Prev Vet Med 1996;28: Oliver SP, Jayarao BM. Coagulase-negative staphylococcal intramammary infections in cows and heifers during the nonlactating and periparturient periods. J Vet Med B 1997;44: Detilleux JC, Kerlhi ME, Stabel JR, Freeman LK, Kellet DH. Study of immunological dysfunction in periparturient Holstein cattle selected for high and average milk production. Vet Immunol Immunopathol 1995;44: Komine KI, Asai KI, Itagaki M, et al. Characterization of biodefense factors in the bovine mammary gland secretions and their variation during the periparturient period. Anim Sci J 1999;70: J169-J Mallard BA, Wagter LC, Ireland MJ, Dekkers JCM. Effects of growth hormone, insulin-like growth factor I, and cortisol on periparturient antibody response profiles of dairy cattle. Vet Immunol Immunopathol 1997;60: Slobodyanik VI, Parikov VA, Smirnova LV, Sapozhnikova NA. Prophylaxis of mastitis in cows. Zootekhniya 1995;10: Perino LJ, Wittum TE. Effect of various risk factors on plasma protein and serum immunoglobin concentrations of calves at postpartum 10 hours and 24. Am J Vet Res 1995;56: Cullor JS. J5 Escherichia coli: a core antigen vaccine for coliform mastitis. Proc Symp Coliform Mastitis, University of California, Davis, 1993: Kelm SC, Detilleux JC, Freeman AE, et al. Genetic associations between parameters of innate immunity and measures of mastitis in periparturient Holstein cattle. J Dairy Sci 1997;80: Heyneman R, Burvenich C, Vercauteren R. Interaction between the respiratory burst activity of neutrophil leukocytes and experimentally induced Escherichia coli mastitis in cows. J Dairy Sci 1990;73: Kremer WDJ, Noordhuizen-Stassen EN, Grommers J, Daemen AJM, Henricks PAJ, Brand J. Preinfection chemotactic response of blood polymorphonuclear leukocytes to predict severity of Escherichia coli mastitis. J Dairy Sci 1993;76: Atroshi F, Parantainen J, Sankari S, Jarvinen M, Lindberg LA, Saloniemi H. Changes in inflammation-related blood constituents of mastitic cows. Vet Res 1996;27: : ~~~~~~~~~~~~~~~~~f...:... CORRIECTION,~-~, Eq:..uinjhegoatnophilic..enterocolltls ls.-.can Vet.:J. 2 00aoP41:871 The~~~~~~ ssbc0le h eut ftebos wr eevd t I'M-c... h tlinwsgvna,4 gk bod et weight :was (BW) dose of dexam ethdecne~i (QxMetone''i'5; Vetqunol Joite ubc) fwẉhc as increased :1to8d e~t gab,2h, h.et 4"sol aend fe h;slso..he biops wer The dl attention po iz t thereaders for any misun.derstanding itmay.hav.e cavs::i;.d: ^:... Can Vet J Volume 42, January