In many African countries, women account for more than half of those living with HIV-1 and young women (<25 years old) are up to 5 times more likely to be HIV-1 infected compared with their male peers.1–3 A combination of biological, behavioral, social, and structural factors may impact young women's susceptibility to HIV-1 infection.4 Engagement in age-disparate relationships, where the male partner is substantially older than the female partner, has been hypothesized as one of the drivers of this higher burden of HIV-1 among young women.5–10 Evidence in support of this hypothesis comes from surveillance data showing consistent sex–age disparities in HIV prevalence,2,3,11,12 where the prevalence of HIV is greater among young women compared with young men, and from epidemiologic studies that report that age-disparate relationships are common among young women.3,13,14 In South Africa, a country that continues to have the greatest HIV-1 burden globally,2 population-based surveys of women aged 15–24 reported that 33%–39% of respondents had a partner >5 years older.3,13 Campaigns to reduce or eliminate age-disparate relationships have been a prominent part of HIV-1 prevention strategies in this region and other high-incidence settings.15–17 Although cross-sectional studies have reported an association between age-disparate relationships and HIV in women,7,14,18 data from prospective studies provide limited and conflicting evidence regarding the contribution of age-disparate relationships to HIV-1 acquisition in South Africa and other African countries.15,18 A recent analysis of data collected between 2003 and 2012 from >2400 women in KwaZulu-Natal, South Africa, observed no association between age-disparate relationships and HIV-1 acquisition in this high-incidence region.15 These data may suggest a shift in local epidemiology of heterosexual HIV-1 transmission but require confirmation. Our objective was to assess the association between age-disparate relationships and HIV-1 acquisition risk among South African women participating in the VOICE trial.
VOICE was a randomized placebo-controlled trial that assessed the safety and effectiveness of oral tenofovir disoproxil fumarate (TDF), oral TDF plus emtricitabine, and vaginal 1% tenofovir gel for HIV-1 prevention in women (Cliniicatrial.gov No. NCT00705679). Detailed methods have been described previously.19 Briefly, between 2009 and 2012, 5029 women from 3 countries (Uganda, South Africa, and Zimbabwe) were enrolled and followed monthly. Eligible women were 18–45 years old, HIV-1-uninfected, not pregnant, sexually active, and willing to use highly effective contraception. All participants provided written informed consent, and all institutional review boards and relevant regulatory authorities approved the trial at each site.
At monthly visits, blood was collected through venipuncture for HIV-1 testing. Structured face-to-face interviews were used to collect data on sexual behaviors, vaginal washing, and contraceptive use. Demographic characteristics of the primary partner, including partner age as reported by the participant, were assessed at enrollment. At quarterly visits, audio computer–assisted self-interviews were conducted to collect additional sexual behavior data, including a change in the primary partner in the last 3 months. Planned follow-up was for a minimum of 12 months and up to 36 months, depending on enrollment date. However, the data safety and monitoring board recommended the TDF arm and the tenofovir arm be discontinued for futility. Participants in the TDF plus emtricitabine and oral placebo arms continued participation until planned study exit (August 2012).
HIV-1 infection status was determined using a standardized algorithm.20 Rapid testing was performed using Determine HIV 1/2 (Abbott Diagnostic Division, Hoofddorp, the Netherlands), OraQuick (Orasure Technologies, Bethlehem, PA), or Uni-Gold Recombigen HIV test (Trinity Biotech, Wicklow, Ireland). Western blot was performed on samples with any positive HIV-1 rapid result (Genetics Systems HIV-1 Western blot kit; BioRad Laboratories, Hercules, CA).
This analysis assessed the association between having an older primary partner at enrollment (age-disparate relationship) and HIV-1 acquisition risk within the first year of follow-up among women enrolled at South African sites (Durban area, Johannesburg area, and Klerksdorp). Because of potential differences in measured and unmeasured factors by country and low HIV incidences in Uganda and Zimbabwe,19 we limited our analysis population a priori to women enrolled in South Africa. Primary partner was defined as, “A man you have sex with on a regular basis or who you consider to be your main partner.” Participants who reported that they did not have a primary partner at enrollment were excluded from this analysis, because questions about partner characteristics were only asked of women who reported a primary partner. Two categorizations of age disparity (as reported by the participant at enrollment) were used: (1) a partner more than 5 years older than the participant; (2) a partner more than 10 years older than the participant. Participants could respond that they did not know their partner's age. Those who indicated “don't know” were retained in the analysis and evaluated as a separate exposure category, because we hypothesized that lack of knowledge may be reflective of a new or less stable partnership and potentially associated with increased HIV-1 risk. Secondary analyses were performed to assess partner age as a continuous variable. The primary outcome was HIV-1 infection within 1 year of VOICE enrollment. Cox proportional hazards models stratified by study site were used to assess the association between age-disparate relationship status at enrollment and time to HIV-1 infection. Participant follow-up was censored at 1 year from enrollment, because we hypothesized that characteristics assessed at baseline would be predictive up to 1 year from assessment. Baseline factors associated with age-disparate relationship status and HIV-1 acquisition were considered for inclusion in the multivariable model.21 Data were available at quarterly visits regarding a change in primary partner in the last 3 months. Because data on partner age were not collected for subsequent partners nor were specimens available to link new infections to specific partners, a sensitivity analysis was conducted censoring women when a change in primary partner was reported (no longer having a primary partner or a new primary partner). The above analyses were repeated stratifying by participant age at enrollment (<25 years versus ≥25 years). All statistical tests were assessed using a 2-sided α of 0.05. Analyses were conducted using Stata version 12.0 (StataCorp, Inc., College Station, TX) and R (R Foundation for Statistical Computing, Vienna, Austria).
Of 4077 HIV-1-uninfected South African participants enrolled in VOICE, 22 had acute HIV-1 infection detected at enrollment and 34 did not return for follow-up. In addition, 148 participants did not report having a primary partner at baseline and 84 were missing data for important baseline factors and were excluded, leaving 3789 participants for analysis. Select characteristics at enrollment are presented in Table 1. The majority of participants were <25 years old and not married or living with their primary partner. Approximately one-quarter [999 (26%)] reported having a primary partner >5 years older and 183 (5%) reported that their primary partner was >10 years older. The proportion of participants who reported that their partner was >5 years older did not differ by participant age category (χ2 P = 0.36); however, a higher proportion of participants ≥25 years old reported having a partner >10 years older (P < 0.001) (Table 1).
There were 243 seroconversions over 3344 person-years (p-yrs) of follow-up [HIV-1 incidence per 100 p-yrs = 7.27; 95% confidence interval (CI): 6.41 to 8.24]. Among participants who reported having a primary partner >5 years older, the HIV-1 incidence was 7.68/100 p-yrs versus 7.12/100 p-yrs among participants whose primary partner was <5 years older [hazard ratio (HR) = 1.10; 95% CI: 0.82 to 1.45]. Results were similar for participants who reported having a primary partner >10 years older (Table 2). A small proportion (2%) of participants did not know their primary partner's age. Among participants who did not know their partner's age, HIV-1 incidence did not differ compared with those who did not report an age-disparate relationship. Effect estimates for all analyses remained unchanged after adjusting for potential confounders. Findings were also similar when partner age difference was considered as a continuous variable. A 1-year increase in partner's age was not associated with risk of HIV-1 acquisition in univariate (HR = 1.00; 95% CI: 0.96 to 1.03) and multivariable models (adjusted HR = 1.01; 95% CI: 0.97 to 1.05). After stratifying by participant age at enrollment, all categorizations of age-disparate relationships (>5 years, >10 years, or continuous variable) continued to not be associated with increased risk of HIV-1 acquisition (Table 2).
There were 790 (21%) participants who reported a partnership change within 1 year of enrollment. Among those reporting a change, median time to partner change was 5.5 months (interquartile range, 2.8–8.3). In sensitivity analyses where follow-up was censored at the visit where the participant reported she no longer had a primary partner or reported a change in primary partner, there were 194 seroconversions over 3005 p-yrs of follow-up (HIV-1 incidence per 100 p-yrs = 6.46; 95% CI: 5.61 to 7.43). Consistent with the primary analysis, age-disparate relationships were not associated with HIV-1 acquisition in the overall study population nor when stratified by participant age at enrollment. Furthermore, the effect estimates were unchanged after adjustment for potential confounders (Table 2).
In this secondary analysis of prospective data from South African women participating in an HIV-1 prevention trial, the incidence of HIV-1 infection did not differ among women who reported being in an age-disparate relationship at enrollment versus those who did not. Findings were consistent after adjustment for potential confounding factors, stratification by participant age at enrollment and in sensitivity analyses that censored follow-up when a woman reported that she no longer had a partner or reported a change in partner.
Our results are consistent with research from the Africa Centre in South Africa, which reported no association between age-disparate relationships and HIV-1 risk among women living in rural KwaZulu-Natal.15 South African women participating in VOICE were similar to women participating in the Africa Centre study with regard to potential HIV risk factors including age, not being married, and having a partner >5 years or >10 years older.15 The overall HIV-1 incidence was nearly identical and very high (VOICE = 7.27/100 p-yrs versus Africa Centre = 7.75/100 p-yrs), demonstrating the urgent need for effective HIV-1 prevention interventions for women in this region. There are also several notable differences in the populations that should be considered when comparing the results of these studies. The Africa Centre was a longitudinal cohort study that included women aged 15–29 and censored follow-up when participants reached age 30, whereas VOICE was a clinical trial that enrolled women aged 18–45, although the majority of VOICE participants were <25 years old. In addition, follow-up in the Africa Centre study was up to 9 years with surveys administered annually, allowing for the assessment of new partnerships over time. In VOICE, partner age was only assessed at enrollment; therefore, our analysis is limited to the primary partnership reported at enrollment into the trial. Finally, Africa Centre study participants were all enrolled from rural KwaZulu-Natal. VOICE enrolled the majority of its South African participants from the KwaZulu-Natal region (areas within ∼50 km of Durban) and also included women from other areas of South Africa. Despite these differences in population and study design, findings from these 2 studies are remarkably similar and further suggest that age-disparate relationships may not be a driver of the current HIV-1 epidemic among young women in South Africa.
Although the prevalence of HIV is higher among older men, it is possible that same-age sexual relationships among young men and women could be driving high endemic levels of HIV and incidence among young women. In South Africa, young men are more likely to be newly infected with HIV and are less likely to be on antiretroviral therapy compared with older HIV-infected men,22,23 thus increasing the potential infectiousness of HIV-infected young men. In addition, mathematical modeling of the population-level impact of behavior change on HIV incidence in generalized epidemics suggests that, in the absence of other behavioral changes, a decrease in age-disparate relationships does not substantially alter the spread of HIV within a sexual network.24 Regardless of partner age, young women in South Africa, particularly in the KwaZulu-Natal region, continue to have one of the highest HIV incidences globally and should be considered a priority population for scale-up of effective HIV prevention interventions, such as pre-exposure prophylaxis.10 In addition, continued epidemiologic surveillance and research into the dynamic and culturally defined risk factors for HIV among young men and women in this region remains critical.
The findings from this study should be interpreted in the context of several strengths and limitations. Strengths included assessment of data collected in the course of a clinical trial with high retention and inclusion of women from several geographic regions within South Africa. The large sample size and number of HIV-1 infections allowed for ample statistical power as demonstrated by the narrow CIs surrounding the effect estimates. Quarterly evaluation of relationship status combined with monthly HIV-1 testing reduced the potential for misclassification due to partnership changes in our sensitivity analysis. However, because data on partner characteristics were only collected at enrollment, we were unable to include information on subsequent primary partners in our analysis. Data were not collected on the age of nonprimary sexual partners; therefore, our findings do not extend to age-disparate relationships with nonprimary partners. Finally, this study included women enrolled in a biomedical HIV-1 prevention trial, thus the findings may not generalizable to women in the general population.
In summary, among South African women enrolled in a large HIV-1 prevention trial, report of being in an age-disparate relationship at enrollment was common but not associated with HIV-1 acquisition within 1 year. There continues to be an urgent need for interventions to reduce the HIV-1 incidence among young women in this region.
The authors sincerely thank the women who participated in this study. They gratefully acknowledge the MTN-003 study team and study sites for their work on data and sample collection and the Statistical Center for their work on data management.
1. UNAIDS. UNAIDS World AIDS Day Report 2011. Geneva, Switzerland: UNAIDS; 2011.
2. UNAIDS. UNAIDS Report on the Global AIDS Epidemic. Geneva, Switzerland: UNAIDS; 2013.
3. Pettifor AE, Rees HV, Kleinschmidt I, et al.. Young people's sexual health in South Africa
: HIV prevalence and sexual behaviors from a nationally representative household survey. AIDS. 2005;19:1525–1534.
5. Leclerc-Madlala S. Age-disparate and intergenerational sex in southern Africa: the dynamics of hypervulnerability. AIDS. 2008;22(suppl 4):S17–S25.
7. Gregson S, Nyamukapa CA, Garnett GP, et al.. Sexual mixing patterns and sex-differentials in teenage exposure to HIV infection in rural Zimbabwe. Lancet. 2002;359:1896–1903.
8. Kaiser R, Bunnell R, Hightower A, et al.. Factors associated with HIV infection in married or cohabitating couples in Kenya: results from a nationally representative study. PLoS One. 2011;6:e17842.
10. Dellar RC, Dlamini S, Karim QA. Adolescent girls and young women: key populations for HIV epidemic control. J Int AIDS Soc. 2015;18(2 suppl 1):19408.
11. Shisana O, Rehle T, Simbayi LC, et al.. South African National HIV Prevalence, Incidence and Behaviour Survey, 2012. Cape Town, South Africa
: HSRC Press; 2014.
12. Katz I, Low-Beer D. Why has HIV stabilized in South Africa
, yet not declined further? Age and sexual behavior
patterns among youth. Sex Transm Dis. 2008;35:837–842.
13. Maughan-Brown B, Kenyon C, Lurie MN. Partner age differences and Concurrency in South Africa
: Implications for HIV-infection risk among young women. AIDS Behav. 2014;18:2469–76.
14. Luke N. Age and economic asymmetries in the sexual relationships of adolescent girls in sub-Saharan Africa. Stud Fam Plann. 2003;34:67–86.
15. Harling G, Newell ML, Tanser F, et al.. Do age-disparate relationships
Drive HIV incidence in young women? Evidence from a population cohort in rural KwaZulu-Natal, South Africa
. J Acquir Immune Defic Syndr. 2014;66:443–451.
18. Kelly RJ, Gray RH, Sewankambo NK, et al.. Age differences in sexual partners and risk of HIV-1 infection in rural Uganda. J Acquir Immune Defic Syndr. 2003;32:446–451.
19. Marrazzo JM, Ramjee G, Richardson BA, et al.. Tenofovir-based Preexposure prophylaxis for HIV infection among african women. N Engl J Med. 2015;372:509–518.
21. Nair G, Balkus JE, Marrazzo JM, et al.. Baseline Predictors of HIV-1 Acquisition Among Women Participating in MTN-003 (VOICE). Boston, MA: Conference on Opportunistic Infections and Retroviruses; 2014.
22. Tromp N, Michels C, Mikkelsen E, et al.. Equity in utilization of antiretroviral therapy for HIV-infected people in South Africa
: a systematic review. Int J Equity Health. 2014;13:60.
23. Muula AS, Ngulube TJ, Siziya S, et al.. Gender distribution of adult patients on highly active antiretroviral therapy (HAART) in Southern Africa: a systematic review. BMC Public Health. 2007;7:63.
24. Hallett TB, Gregson S, Lewis JJ, et al.. Behaviour change in generalised HIV epidemics: impact of reducing cross-generational sex and delaying age at sexual debut. Sex Transm Infect. 2007;83(suppl 1):i50–i54.
Keywords:Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
HIV-1 in women; age-disparate relationships; male partner; South Africa; sexual behavior