In Sub-Saharan Africa, divorced, separated and widowed individuals who have remarried are more likely than those who are single or married for the first time to be HIV-positive, especially if they are female, according to an analysis of nationally representative data for 13 countries.1 Compared with those in their first marriage, remarried women had higher odds of being HIV-positive in all but one country (odds ratios, 2.0–8.3). Individuals who were formerly but not currently married were also at risk; for example, in the three countries where HIV prevalence was highest, the odds of being infected were elevated at least twofold among formerly married men (2.4–5.3) and women (3.0–3.9).

Although prior studies have found high rates of HIV infection among divorced and widowed individuals in Africa, most of those studies used nonrepresentative samples or did not focus on the remarried population. To address these issues, investigators analyzed data from Demographic and Health Surveys and AIDS Indicator Surveys in which the majority of respondents (70–97% of women and 63–96% of men) had been tested for HIV. Women were eligible for the surveys, which were conducted between 2003 and 2006, if they were aged 15–49; the age range for men varied across countries but was generally 15–54 or 15–59. Data for Ethiopia were available only for women, as male respondents were not tested for HIV. Respondents who had never been married (or in a marriage-like union) were classified as single; those with a history of marriage were categorized as either married for the first time, remarried or formerly married (if they were separated, divorced or widowed but not currently married). To assess the relationship between remarriage and HIV status, the researchers performed descriptive analyses, as well as logistic regression analyses that adjusted for social and demographic characteristics (including the practice of polygyny).

Overall, the prevalence of HIV was relatively low in Senegal, Burkina Faso, Côte d’Ivoire, Guinea and Ethiopia (less than 2% for both sexes); moderate in Ghana, Rwanda, Cameroon, Tanzania and Kenya (3–9% among women and 2–6% among men); and high in Malawi, Zimbabwe and Lesotho (13–26% among women and 10–19% among men). Some 2–9% of men and 4–16% of women had been in a marriage that ended. Among those who were currently married or in union, 6–44% of men and 3–26% of women had previously been married.

In 10 of the 13 countries, the prevalence of HIV was higher among remarried women than among other women. For example, in Côte d’Ivoire, 13% of remarried women were infected, compared with 6% of other women; similar trends were apparent in Kenya (17% vs. 8%) and Lesotho (46% vs. 26%). The pattern was essentially the same for formerly married women, who were more likely than other women to be infected in 10 countries, including Ethiopia (6% vs. 1%), Rwanda (12% vs. 2%) and Zimbabwe (46% vs. 17%).

The relationship between marital history and HIV was less consistent among men. HIV prevalence was greater among remarried men than other men in six countries, including Kenya (10% vs. 4%) and Malawi (20% vs. 7%), and greater among formerly married men than other men in eight countries, including Côte d’Ivoire (11% vs. 2%), Rwanda (8% vs. 2%) and Zimbabwe (44% vs. 13%).

Logistic regression analyses revealed that the odds of HIV infection were generally ele-vated among formerly married and remarried men and women, especially where the prevalence of infection is high. In all three high-prevalence countries (Malawi, Zimbabwe and Lesotho), divorced, separated and widowed individuals who had not remarried had higher odds of HIV infection than those who were single (odds ratios, 2.4–5.3 among men and 3.0–3.9 among women). Findings were similar among remarried men in these countries (2.6–9.0), though remarried women had increased odds of infection only in Zimbabwe (2.3).

In countries where HIV prevalence was low to moderate, the likelihood of being HIV positive was elevated among previously married women in seven countries (odds ratios, 4.1–12.4) and among previously married men in three countries (2.5–9.4). HIV infection was associated with remarriage in five countries among women (2.2–16.2) and in six countries among men (2.2–9.0).

In multivariate analyses that focused on married respondents, the odds of being HIV-positive were higher among remarried women than among those married for the first time in every country but Lesotho (odds ratios, 2.0–8.3). Among men, remarriage was associated with HIV in Rwanda, Malawi and Zimbabwe (2.1–3.0).

Finally, the researchers assessed HIV discordancy among married couples in which at least one partner had been married before. In the vast majority of cases, both partners were HIV-negative. However, in five countries, at least 10% of couples were serodiscordant, and in two countries, both partners were infected in at least 20% of couples. Partners were generally more likely to be serodiscordant than to both be HIV-positive; the primary exceptions were in Zimbabwe and Lesotho.

The investigators acknowledge that the timing and source of formerly married and remarried respondents’ HIV infections cannot be established. Nonetheless, the elevated HIV prevalence among formerly married and remarried women and the high rates of serodiscordance among remarried couples in many countries are of concern, they note. Current HIV prevention programs in Africa, which typically emphasize abstinence and condom use, are most likely “largely irrelevant” to remarried men and women. To help reduce HIV transmission within couples, the researchers recommend that alternative approaches “address gender discrimination and violence” and include testing for both partners.—S. Ramashwar


1. de Walque D and Kline R, The association between remarriage and HIV infection in 13 Sub-Saharan African countries, Studies in Family Planning, 2012, 43(1):1–10.