Predictive model for Zika virus RNA minipool nucleic acid testing in outbreak scenarios.

Abstract

Zika virus (ZIKV), a mosquito-borne flavivirus, causes asymptomatic infections in blood donors and can be transmitted by transfusion. During the 2016 US outbreak, universal individual-donation nucleic acid testing (ID-NAT) was used to screen the blood supply for ZIKV. Testing pooled samples from multiple donations with minipool (MP)-NAT is less sensitive than ID-NAT, which raised questions about its utility in ZIKV outbreaks. A mathematical model and computer simulation determined the risk of missing ID-NAT-reactive and immunoglobulin (Ig) M-negative donations in a ZIKV outbreak if MP-NAT is used initially instead of ID-NAT. The model calculated the time required for ZIKV RNA to replicate to a concentration detectable by testing donations individually or in pools of 6 (MP6) or 16 (MP16). A computer simulation then randomly selected infection times to determine the probability of detection by the candidate tests. The probability of detecting the first ID-NAT-reactive unit in an outbreak is 92% (2.5th-97.5th percentile, 79%-99%) by MP6 and 85% (2.5th-97.5th percentile, 67%-99%) by MP16. When one donation is detected by MP-NAT, the model predicts that the chance of having missed one or more ID-NAT-reactive donations is 8% to 15%. The probability of missing a unit by MP-NAT is constant over the course of the outbreak (8% by MP6, 15% by MP16). The model predicts that the probability that a candidate MP-NAT will detect the first ID-NAT-reactive unit in a ZIKV outbreak is 85% to 92% and remains constant over time.