Evaluating HIV Prevention Programs: Herpes Simplex Virus Type 2 Antibodies as Biomarker for Sexual Risk Behavior in Young Adults in Resource-Poor Countries

Background Measuring effectiveness of HIV prevention interventions is challenged by bias when using self-reported knowledge, attitude or behavior change. HIV incidence is an objective marker to measure effectiveness of HIV prevention interventions, however, because new infection rates are relatively low, prevention studies require large sample sizes. Herpes simplex virus type 2 (HSV-2) is similarly transmitted and more prevalent and could thus serve as a proxy marker for sexual risk behavior and therefore HIV infection. Methods HSV-2 antibodies were assessed in a sub-study of 70,000 students participating in an education intervention in Western Province, Kenya. Feasibility of testing for HSV-2 antibodies was assessed comparing two methods using Fisher's exact test. Three hundred and ninety four students (aged 18 to 22 years) were randomly chosen from the cohort and tested for HIV, Chlamydia trachomatis, Neisseria gonorrhoeae, and Trichomonas vaginalis. Out of these, 139 students were tested for HSV-2 with ELISA and surveyed for sexual risk behavior and 89 students were additionally tested for HSV-2 with a point-of-contact (POC) test. Results Prevalence rates were 0.5%, 1.8%, 0.3% and 2.3% for HIV, Chlamydia trachomatis, Neisseria gonorrhoeae, and Trichomonas vaginalis, respectively. Prevalence of HSV-2 antibodies was 3.4 % as measured by POC test (n=89) and 14.4 % by ELISA (n=139). Specificity of the POC test compared with ELISA was 100%, and the sensitivity only 23.1%. Associations between self-reported sexual behavior and HSV-2 serostatus could not be shown. Conclusions Associations between self-reported sexual risk behavior and HSV-2 serostatus could not be shown, probably due to social bias in interviews since its transmission is clearly linked. HSV-2 antibody testing is feasible in resource-poor settings and shows higher prevalence rates than other sexually transmitted diseases thus representing a potential biomarker for evaluation of HIV prevention interventions.