Field trial of four cowside antibiotic-residue screening tests.

Four commercially available screening tests for antibiotic residues in milk were evaluated for their ability to correctly identify the antibiotic status of cows. A field trial, which included 199 cows from 2 herds, was conducted. Sensitivity, specificity, and likelihood ratios for a positive test result were calculated by using the Bacillus stearothermophilus var calidolactis disk assay as the reference test. The relationship of risk factors to the probability of a false-positive result for each screening test was modeled by use of unconditional logistic regression. The risk factors evaluated in these models were loge somatic cell count (scc), intramammary infection, herd, milk appearance, time milk sample frozen before tested, days in lactation, parity, and manufacturer's lot number. The risk factors log(e) scc and intramammary infection were forced into all models. The overall specificities for the 4 tests ranged from 0.78 to 0.95, whereas likelihood ratios for a positive test result ranged from 4.54 to 20.0. When the confounding of cofactors was controlled in the logistic model, there was a positive effect of log(e) scc on the probability of a false-positive result for 3 of the screening tests, that is, for incremental increases in log(e) scc, there was an increasing likelihood for a false-positive result. In some of the tests, parity and intramammary infection also influenced the likelihood of a false-positive result. The goal of cowside testing is to assist in the production of high-quality, antibiotic residue-free milk from dairies.(ABSTRACT TRUNCATED AT 250 WORDS)