Background:  A new highly reproducible assay, the HIV-1 IgG BED-capture enzyme immunoassay (BED-CEIA), can be used to determine HIV-incidence in cross-sectional specimens with similar performance across subtypes A through E. In Thailand, this assay has only been evaluated in injecting drug users. The generalization of these findings to other risk groups, such as heterosexuals, needs to be evaluated. Our objective was to determine how well BED-CEIA estimated HIV-1 incidence rates compared with observed incidence rates in a cohort of military recruits.

Methods:  In May 1991, 1113 military conscripts entered a 2-year period of service in the Royal Thai Army at Phitsanuloke, Northern Thailand. Screening for HIV infection was done at enrollment and 5, 17, and 23 months later. HIV-positive specimens from these times were tested by BED-CEIA. A normalized OD of 0.8 on the CEIA was used to identify recent infections; incidence rates were estimated using a seroconversion window period of 153 days and the number of seronegatives in the population.

Results:  Of 157 HIV-seropositive specimens tested, 7 were recent. Of these, 2, 4, and 1 were identified from specimens drawn at 5, 17, and 23 months, respectively. The BED-CEIA estimated incidence rates and 95% confidence intervals (95% CI) for these times were 0.55 (–0.2 to –1.3), 1.44 (0.03 to 2.9), and 0.51 (–0.5 to 1.5) per 100 person-years. Observed incidence rates and 95% CI at 5, 17, and 23 months were 0.62 (0.1 to 1.1), 1.36 (0.5 to 2.3), and 1.45 (0.4 to 2.5) per 100 person-years, respectively.

Conclusions:  The BED-CEIA-estimated incidence rates are similar to the observed rates, even when used on samples frozen for > 10 years and from a relatively low incidence, mainly heterosexual risk population. Use of the BED-CEIA for HIV incidence estimates using cross-sectional specimens may be useful for HIV surveillance, intervention, and control. Moreover, it may be a valuable tool in resource-poor settings, because HIV-1 incidence can be estimated from a single specimen collected from a cross-sectional population.

Keywords: HIV incidence; BED-CEIA; STAHRS