Direct HIV testing in blood donations: variation of the yield with detection threshold and pool size.

Abstract

This study was designed to evaluate the potential yield of introducing nucleic acid amplification testing (NAT) in blood donation, according to the detection threshold and the pool size. A mathematical model of early HIV-1 population dynamics in blood has been developed and is used to predict the window period for NAT, according to the detection threshold and the pool size. The corresponding number of undetected, infected blood donations and the residual risk are estimated by using a previously published simulation model for the United States (9.96 million blood donations from regular donors, and an observed rate of HIV antibody-positive blood donations of 3.18/100,000) and for France (2.32 million blood donations, and a rate of antibody-positive donations of 0.9/100,000). The average window period from infection predicted by the mathematical model for NAT ranges from 8.4 to 15.6 days, according to the detection threshold and the pool size. The maximum yield of adding NAT to the current antibody tests is estimated at 14 donations for the United States and 2 for France. The maximum yield of adding NAT to the newly developed combined HIV antibody and p24 antigen tests is 7 donations for the United States and 1 for France. NAT at blood donation could reduce the HIV-1 window period to a minimal value of 8 days without pooling the blood samples, but the yield of NAT would be close to that of combined HIV antibody and p24 antigen tests for high values of the detection threshold and the pool size.