# PAPER SELECTED

## A method of prevalence estimation of intramammary Staphylococcus aureus and Streptococcus agalactiae infections in herds by examination of bulk milk samples

BENDA P, VYLETELOVA M, TICHACEK A

Veterinarni Medicina 42, 1997, 101-109

Parallel quantitative determination executed in 92 herds at 107 sampling dates was focused on the counts of major pathogens (Staphylococcus aureus, Streptococcus agalactiae) in bulk milk samples and on the prevalence of the above pathogens in a herd by examination of individual milk samples. The counts of main pathogens were also determined in terms of quantity in rinsing water before milking and in bulk milk samples in 5 herds with pipeline milking. Tab. I shows the qualitative analysis of the relation. Sensitivity of the method is satisfactory for the pathogens observed (95% and 91%, resp.), but method is less specific for Staph. aureus (67% against 92% Str. agalactiae). Figs. 1 and 2 show coordinate graphs of the results obtained while Figs. 3 and 4 document the distribution of frequency of the particular values in data sets. The values do not exhibit normal distribution (P < 0.01). Spearman's coefficient of rank correlation of bulk milk and individual examinations amounted to 0.823 and 0.900 for Staph. aureus and Str. agalactiae, respectively. Four mathematical models were tested in the course of quantitative analysis, describing the relation between bulk milk examination and individual examinations: (1) linear regression, (2) linear regression with fixed starting point, (3) logarithmic regression and (4) irrational function. A model based on the equation of irrational function (4) was found to be best: y = a + bx + c root x + k +/- a(1) + i(t)c(1) root x + k(1). Tab. II shows the parameters of the equation for examined microorganisms. Correlation coefficients for the above equation are r = 0.733 and r = 0.842 for Staph. aureus and Str. agalactiae, respectively. Prediction curves (Figs. 5 to 8) and confidence regions of prediction curves were also determined for the best model, and a prediction table was constructed (Tab. III). It was confirmed that the milking machines were not a significant source of direct contamination of bulk milk samples with the examined pathoges (Tab. V)

FULL PAPER on request

« back