Significantly diminished long-term specificity of the BED capture enzyme immunoassay among patients with HIV-1 with very low CD4 counts and those on antiretroviral therapy

OBJECTIVE: To estimate the proportion who test as recent infections by the BED capture enzyme immunoassay (BED) among patients about to commence, and those receiving, antiretroviral therapy. DESIGN: Cryopreserved plasma samples from HIV patients on the national antiretroviral treatment (ART) rollout program at Tygerberg Hospital HIV clinic, South Africa, were tested using the BED assay. PARTICIPANTS: Five hundred five patients qualifying for ART were included in this study. METHOD: All plasma samples from each patient were tested by BED. Basic demographic data, HIV-1 viral load, and CD4 count results were obtained from the laboratory database. MAIN OUTCOME: The proportion presenting as false recently infected is reported. RESULTS: Among patients, with presumed long-term HIV-1 infections, about to commence ART, 11.2% [95% confidence interval (CI): 8.3 to 14.5%] tested recent by BED. The proportion was higher among patients with CD4 counts <50 cells per microliter [odds ratio 2.63, 95% CI: 1.39 to 5.00] and log10 HIV-1 viral load less than 4 [odds ratio 3.03, 95% CI: 1.05 to 9.09]. Proportions testing false recent increased from 11.2% before ART to 17%, 25%, 38%, and 56% at 0.5, 1, 1.5, and 2 years, respectively, after ART initiation. CONCLUSIONS: If the BED method is to be used for the accurate estimation of HIV incidence from cross-sectional surveys, it will be essential, before other statistical adjustment methods, to identify, at least, all cases who are on ART and all those with CD4 counts < 50 cells per microliter. The more general remaining problem is the unequivocal identification of all persons with long-term HIV infections.