Evaluating the use of packed cell volume as an indicator of trypanosomal infections in cattle in eastern Zambia.

In this study, packed cell volume-values (PCV) are evaluated as indicator of trypanosomiasis infections in cattle. A total of 734 blood samples were collected in 11 different sampling sites in eastern Zambia: 84 calves (< 1 year), 52 young females and 40 young males (between 1 and 3 years), 228 cows, 317 oxen and 13 bulls (> 3 years). All samples were subjected to three diagnostic tests: parasitological examination using the buffy coat method, PCR/RFLP and PCV determination. The results were compared and analysed in a Bayesian model, which allowed the estimation of the infection prevalence and the respective test sensitivities and specificities. The presence of a trypanosomal infection significantly reduced the PCV, independently of the age and sex of the infected animal. The estimated prevalence of trypanosomal infections in the study area was 34% (95% credibility interval: 30-38%). While the specificity of both the parasitological and the PCR/RFLP tests were set to 1, the parasitological diagnosis had a low sensitivity (37%) compared to the PCR/RFLP (96%). When using a cut-off value of 24, the PCV had a high specificity (98%) but a rather low sensitivity (53%) for identifying trypanosomiasis infections. Using 26% as a cut-off increased the sensitivity to 76% without much affecting the specificity (94%). A parallel combination of the parasitological diagnosis and the PCV improved the diagnostic sensitivity (74% and 89% for PCV cut-off values of 24% or 26%, respectively) while specificity remained high (98% and 94% for PCV cut-off values of 24% or 26%, respectively). These results suggest that such a combination could advantageously be used for the diagnosis of cattle trypanosomiasis in the field: it is much more sensitive than parasitological examination alone and it is much cheaper than molecular tests. However, the value of this approach depends largely on the determination of an appropriate cut-off value to consider a sample positive, depending on the required test sensitivities and specificities.