The effects of pooling serum samples from seroconverting individuals or individuals with end stage disease for HIV antibody testing: a comparison of four screen tests and three pool sizes

Objectives: To determine the sensitivity of four HIV screening tests when testing pooled samples from individuals with low antibody levels. Methods: Samples were obtained from 10 individuals with strong positive WB results, 11 seroconverting individuals and 8 individuals with late stage disease. Samples were tested individually and with pool sizes of 5, 10 and 20 using the BioChem test, the SYVA test, the CBC Recombigen test and Organon-Teknika's Vironostika test. Results: Samples from individuals with strong positive WB results tested positive according to all four tests at all pool sizes. One of the samples from the eight individuals with late stage disease tested negative with a pool size of 20 for all four tests and with pool sizes of 5 and greater with CBC Recombigen, SYVA and Organon-Teknika. More than half of the pools of size 20 containing samples from seroconverting individuals tested negative with SYVA, CBC Recombigen and Organon-Teknika. Sensitivity varied between 18% and 82% when samples were pooled in groups of 10 and between between 27% and 100% when samples were pooled in groups of 5. Conclusions: Pooling of samples from individuals with late-stage disease or who are in the process of seroconverting is not recommended when the results of each test are critical. When only the aggregate results are of interest, such as in anonymous seroprevalance studies, the loss in sensitivity due to pooling samples from these individuals will not qualitatively affect prevalence estimates.