Background:  To determine the effect of ART on cross-sectional incidence testing, we compared the Serological Testing Algorithm for Recent HIV Seroconversion (STARHS), an antibody vs the affinity/avidity assay on a mature HIV-1 epidemic receiving ART.

Methods:  From June 29, 2001 to August 16, 2001, 1549 patients entering the Johns Hopkins University Emergency Department were enrolled by interview based, identity unlinked sero-survey. Sera were tested for HIV by enzyme immunoassay (EIA), Western blot, and if positive for HIV RNA by Roche AmplicorTM v1.5. The STARHS and affinity/avidity assays were used to determine recent HIV infection. The Avidity/Avidity method was done with the Genetic Systems EIA, with DEA as the chaotropic agent and an AI cutoff of 35% for recent infection. Survey information was used to determine demographic and risk behavior.

Results:  The HIV prevalence was 11.8%. Of the HIV-infected subjects, 23.8% did not know they were infected. STARHS determined that 11 of 183 individuals were recently infected while affinity/avidity testing confirmed 6 such individuals. In removing the viral load negative individuals, confirmed by affinity/avidity testing to be chronically infected form the incidence calculation, the incidence estimate was lowered from 1.73 to 0.94%/year in this population. All 5 discrepant samples were Western blot positive, viral load undetectable, and of those with interview data (2 of 5) were on ART. 

Conclusions:  The use of ART in a population significantly affects the capacity of STARHS in determining HIV incidence. Finally, these results suggest that inner city emergency departments provide an opportunity to identify previously unrecognized HIV infection and together with incidence testing allow for real-time monitoring of the epidemic. In addition the emergency department could serve as a venue for disseminating information to at risk individuals. 

Keywords: incidence testing; urban; ARV