Christopher D.Pilcher, MD
Assistant Professor of Medicine, University of North Carolina at Chapel Hill
Background: Since 2002, NC DHHS’ STAT program has supplemented traditional HIV viral lysate EIA testing with HIV NAAT for all public VCT, using specimen pooling of seronegative specimens. We have detected acute HIV infections in real time, increasing case identification by 4% with high accuracy (Sp>0.999). This has allowed efficient integration of public VCT with highly efficient prevention and surveillance activities. Recently, we have investigated the implications of using newer rapid tests and ELISAs for the potential identification of acute infections at VCT.
Methods: HIV VCT-based diagnostic performance evaluations were conducted in sites with expected high prevalence and incidence of HIV to compare new tests with NAAT/pooling for identification of HIV infection: 1) downtown Atlanta; 2) Lilongwe, Malawi; and 3) Porto Alegre, Brazil. For each study, HIV infection was defined by a positive antibody test or a positive HIV NAAT or antigen with seroconversion. Ultrasensitive NAATs (Roche 1.5 ultra, Gen-Probe TMA) and small pool sizes (1:12-1:50) were used.
Results: No false positive NAAT results were returned in any study.
In Atlanta (n=2202), we assessed performance of Genetic Systems HIV-1/2+O and HIV-1 rLAV 3rdgeneration EIAs. Overall 70 (3.3%) subjects were HIV+. 66 EIA+ and. 4 EIA-NAAT+ [3rd gen EIA sensitivity for detectable HIV, Se=94.3% (95% CI 86.2 to 97.8%)]. In Porto Alegre (n=933) Genscreen-Biorad HIV Ag/Ab 4th generation EIA followed by Abbott HIV MEIA and Biomanguinhos IF assays correctly identified all 183 HIV Ab+ and 5 Ab- infections (4th gen EIA Seall=1.00, Seacute=1.00). In Lilongwe (n=1440), 587 subjects (40.8%) were HIV+. 555 were antibody+ on parallel Unigold and Determine rapid testing (Unigold Se=96.5%, Determine Se=99.0%, 2RTs Se=96.2%). 20 had RT- or RTdiscordant/NAAT+ acute HIV. For evaluation of antibody negative specimens, the performance of p24 Ag for true infection (std. p24 Se=0.75; Sp=0.995; PPV=0.75) was improved using the heat-dissociated (HD) assay (HD p24 Se=0.84; Sp=1.00; PPV=1.00). Interestingly, 7 of 21 subjects with discordant RT results (33.3%) had acute infection in this study.
Discussion: While more sensitive rapid tests and 3rd generation EIAs may reduce the window period relative to viral lysate EIA, our findings suggest that acute HIV detectable only by antigen or NAAT can still be disturbingly prevalent in high risk testing sites. p24 Ag testing (either alone in standard format, with an added heat-dissociation step or as part of a combination 4th generation assay) has acceptable sensitivity for acute HIV. Because HIV prevalence in the US is relatively low, however, the positive predictive value of p24 Ag detection for acute HIV is questionable and merits further study. Numerous formats for acute HIV detection are available, however all require that some provision be made for follow-up testing to confirm HIV status and to link patients to emergency prevention services.

Last Update: April 8, 2005