Background HIV and STI incidence rises sharply during adolescence and young adulthood due to increased sexual activity. At this time youth also undergo major life transitions for example, leaving school, becoming sexually active, leaving home and getting married. It is hypothesised that risk of HIV/STIs increases during or just after such transitions for example, the interval between acquiring a first sexual partner and first marriage is usually the time of most rapid partner turnover. Targeting youth, their sexual partners, and/or families at key transitional periods may be effective in reducing HIV/STI rates. This study explores the association between marital transitions (change in marital status) and HIV seroconversion in a cohort in Tanzania.

Methods In 1998, the MEMA kwa Vijana sexual and reproductive health intervention trial recruited 9645 young people aged 14–17 years in rural Mwanza. We analysed a sub-set of the cohort (3240) who attended follow-up surveys in 2001 and 2007. Reported marital and measured HIV status at both time points was used to describe marital transitions and their association with HIV seroconversion. Logistic regression, adjusting for demographic, behavioural and biological risk factors was used to assess whether marital transition was independently associated with seroconversion.

Results Cases of HIV and the proportion married increased very substantially between 2001 and 2007 (Abstract P1-S5.17 table 1). Seroconversion rates were higher among females remaining unmarried (6.4%, p=0.01) and those transitioning from married or never married to previously married (8.8%, p<0.001). HIV seroconversion was independently associated with recent travel and HSV2 infection among males and number of recent sexual partners, HSV2 infection and lifetime number of pregnancies among females. Marital transition was not an independent risk factor.

Conclusions Women who remained unmarried and/or transitioned from married/unmarried to previously married during the 6 year follow-up were more likely to become infected with HIV. The association between marital transition and HIV seroconversion among females appeared largely due to differences in number of reported sexual partners. Limitations included the lack of information on the exact timing of marriage or breakdown of marriage and of seroconversion, and the limited proportion of cohort members eligible for inclusion. Nevertheless, marital transitions appear important determinants of risk.