Evaluation of selective screening of donors for antibody to Trypanosoma cruzi: seroprevalence of donors who answer "no" to risk questions.

ABSTRACT

BACKGROUND: Selective testing of donors for Trypanosoma cruzi infection relies on identification of at-risk donors with screening questions. Using risk modeling and a seroprevalence study, we evaluated the risk of questions failing to identify T. cruzi antibody-positive donors. STUDY DESIGN AND METHODS: The rate of donors with unreported risk was estimated by a telephone survey of 2677 donors who answered "no" to risk questions. The number of T. cruzi antibody-positive donors missed by risk questions was estimated from the product of this rate and the selective testing T. cruzi antibody-positive rate. The 95% confidence interval (CI) was estimated by Monte Carlo simulation. To test the model, 60,132 donors were tested for T. cruzi antibody (26% of donors in selected regions, Phase I). In Winnipeg, Manitoba, the highest-risk region, 26,915 donors were tested (92.5% of donors, Phase II). RESULTS: In the telephone survey, 21 (0.8%) donors reported risk factors that would have identified them for selective testing. Seven were born in Mexico or Central or South America, five had travel risk, and nine had mother or maternal grandmother risk. The 95% CI for predicted number of T. cruzi antibody-positive donors answering "no" to risk questions was 0.71 to 4.38. In Phase I, one Winnipeg donor confirmed positive but had answered risk questions correctly. No other positive donations were identified. CONCLUSION: The estimated risk of T. cruzi-positive donors who answer "no" to risk questions is low and is confirmed by the seroprevalence among these donors.