Risk factors for HIV-1 seroconversion among young men in northern Thailand

Objective.-To identify behavioral and sociodemographic risk factors for incident human immunodeficiency virus-1 (HIV-1) infection among healthy young men in northern Thailand. Design.-Men inducted into military service in northern Thailand in May and November 1991 were followed at 6-month intervals until discharge 2 years later. Trained nonmilitary interviewers identified risk factors for HIV-1 infection through interviews with the men. Setting.-Thirteen military bases in northern Thailand. Participants.-A total of 1932 seronegative men, aged 19 to 23 years (average age, 21 years) at enrollment, conscripted into the Royal Thai Army and Air Force from six upper-northern Thai provinces. Main Outcome Measures.-Human immunodeficiency virus-1 seroincidence as determined through enzyme-linked immunosorbent assay and verified by Western blot, and univariate and multivariate analyses of risk factors related to H1V-1 incidence. Results.-A total of 85 men seroconverted to HIV-1 over the period of observation, giving an incidence rate of 2.43 per 100 person-years. Factors strongly associated with HIV-1 seroconversion were frequency of visits to female commercial sex workers (CSWs), sex with men, and incident sexually transmitted diseases (STDs). High frequency of condom use showed a significant (P<.001) protective effect for HIV-1 incidence among men with a history of recent sex with female CSWs in univariate analysis, but a multivariate model demonstrated no difference in HIV-1 seroconversion rates by consistency of condom use. Multivariate analysis incorporating condom use showed that having sex with men (adjusted relative risk [RR], 2.59; 95% confidence interval [CI], 1.08 to 6.25), having sex with CSWs (adjusted RR ranged from 2.54 [95% CI, 1.81 to 3.58] to 2.74 [95% CI, 1.56 to 4.81]), and incident STDs (adjusted RR, 2.38 [95% CI, 1.31 to 4.32]) to be predictors of incident HIV-1 infection. Substance use was not associated with HIV-1 seroconversion rates in multivariate analysis. Conclusion.-The HIV-1 incidence in this cohort of young men appears to be primarily attributable to having sex with female CSWs. Condom use provided some protection, although nor in multivariate analysis; however, condom use has previously been shown likely to be useful in preventing HIV-1 transmission. Thus, programs to increase effective condom use in brothels are essential. Efforts to extend condom use to non-CSW partners are especially needed. More effective prevention and treatment of STDs may also be necessary to decrease HIV-1 infection in this population.