Accuracy of an indirect fluorescent-antibody test and of a complement-fixation test for the diagnosis of Babesia caballi in field samples from horses.

We evaluated the indirect fluorescent-antibody (IFA) test and complement-fixation (CF) test for diagnosis of equine piroplasmosis in the absence of a gold standard. Using Evan's blue, we estimated the specificity of the IFA test on a parasite-free, field horse population to be 98% (95% confidence interval=97, 99). We observed an excellent test agreement (kappa=0.83) between two collaborating laboratories when the IFA test was performed on identical samples from an endemic area. Using Bayesian analysis with informative prior probability distributions, we estimated the sensitivity of the IFA test to be 92% (95% probability interval, PI=81, 98), and specificity to be 95% (95% PI=88, 99). The CF test sensitivity and specificity estimates were 28% (95% PI=15, 47) and 99% (95% PI=96, 100), respectively. We found the IFA to be superior to the CF test, and the inclusion of Evan's blue in test protocol improved the performance of the IFA test. We conclude that the IFA test for Babesia caballi is a sensitive and specific test for the diagnosis of equine piroplasmosis.