Soweto was the most common enrollment site (47%), followed by Gugulethu (39%), and Orange Farm (14%) (Table 1). Most HIV-infected participants (77%) were women. The mean age of HIV-infected participants was 33 years, and 29% had completed secondary school. Most HIV-infected participants (82%) had at least 1 child, and few (4%) reported >1 sex partner in the last month. Two thirds of couples were married or cohabitating; 79% had been together for >1 year (Table 2). On average, men were 4.1 years older than women, regardless of which partner was HIV infected. Few HIV-infected participants (4%) reported recent relationship violence.
At baseline, 13% of HIV-infected participants were tested ≤7 days before baseline, 26% 8–14 days before baseline, 11% 15–30 days before baseline, and 50% >30 days before baseline (Table 1). The median time since HTC was 29 days (interquartile range: 11 days–9.2 months) overall and 9.2 months (interquartile range: 3.8 months–25.6 months) among the previously tested.
At baseline, almost all HIV-infected participants reported ≥1 sex act in the last month with their study partner (new: 94.1%, previous: 96.1%, P = 0.3) (Table 3). Among those newly tested, 53% reported ≥1 unprotected sex act in the last month compared with 25% of those previously tested (OR: 3.3, 95% CI: 2.3 to 4.8). Of those reporting any unprotected sex, the mean numbers of unprotected acts were 8 (newly tested) and 6 (previously tested).
One month after baseline, most HIV-infected participants reported sexual activity with study partners (new: 87.9%, previous: 89.1%, P = 0.7). Nine percent of those newly tested and 13% of those previously tested reported any unprotected sex in the last month (OR: 0.7, 95% CI: 0.4 to 1.3). Of those reporting any unprotected sex, the mean numbers of unprotected acts were 8 (newly tested) and 7 (previously tested).
Twelve months after baseline, most HIV-infected participants continued to report sexual activity with study partners (new: 73.6%, previous: 78.2%, P = 0.3). Six percent of those newly tested and 14% of those previously tested reported any unprotected sex in the last month (OR: 0.4, 95% CI: 0.2 to 0.8). The mean number of unprotected sex acts was 6 in both groups.
In multivariable analysis with adjustment for study site and marital status, the odds of being in a group engaging in unprotected sex were higher among those tested ≤7 days before baseline compared with those tested >30 days before baseline (adjusted OR: 9.3, 95% CI: 3.6 to 24.2). Similarly, the number of unprotected sex acts was higher among those tested ≤7 days before baseline compared with those tested >30 days before baseline (adjusted relative number: 1.7, 95% CI: 1.2 to 2.6). The final adjusted model differed minimally from the unadjusted model (Table 4) or the fully adjusted model (2%, logistic parameter estimate; 8%, NB parameter estimate). Both the adjusted odds and adjusted relative number of unprotected sex acts were higher among those tested 7–14 and 15–30 days before baseline compared with those tested >30 days before baseline (Table 4).
At baseline, the odds of unprotected sex in the last month were substantially higher among those tested ≤7 days before baseline than those tested >30 days before baseline (OR: 7.01, 95% CI: 3.80 to 12.94), but these groups were more comparable by month 1 (OR: 0.53, 95% CI: 0.18 to 1.58) and remained so at month 6 (OR: 0.45, 95% CI: 0.15 to 1.34) and month 12 (OR: 0.40, 95% CI 0.10 to 1.53).
The odds of unprotected sex were lower at month 1 compared with baseline within each group: ≤7 days, OR = 0.03; 8–14 days, OR = 0.09; 15–30 days, OR = 0.19; >30 days, OR = 0.45. In all groups, the odds of unprotected sex remained lower at month 6 compared with baseline: ≤7 days, OR = 0.04; 8–14 days, OR = 0.17; 15–30 days, OR = 0.16; >30 days, OR = 0.69. Similarly, in all groups, the odds of unprotected sex remained lower at month 12 compared with baseline: <7 days, OR = 0.04; 8–14 days, OR = 0.10; 15–30 days, OR = 0.06; >30 days, OR 0.66 (Figure 2a) All ORs were significant at an alpha level of 0.05. Model-building resulted in no adjustment. Full adjustment resulted in a 7% change in the primary comparison of interest (month 1 versus baseline among persons tested ≤7 days before baseline).
In longitudinal NB analysis, all newly tested participants (<30 days) were analyzed together because of sparse data. The number of unprotected sex acts was higher among the newly tested than the previously tested at baseline (1.4, 95% CI: 1.1 to 1.8), but the groups were the same by month 1 (relative number: 1.0, 95% CI: 0.6 to 1.8) and remained so at month 6 (relative number 0.9, 95% CI: 05 to 1.7) and month 12 (relative number: 1.1, 95% CI: 0.6 to 2.0). Among the newly tested, the number of unprotected sex acts in the last month was similar at month 1 (0.8, 95% CI: 0.5 to 1.3), month 6 (0.7, 95% CI: 0.4 to 1.1), and month 12 (1.0, 95% CI: 0.6 to 1.7) compared with baseline, but the results were imprecise. Among previously tested persons, the number of unprotected sex acts was the same at month 1 (1.1, 95% CI: 0.9 to 1.5), month 6 (1.0, 95% CI: 0.7 to 1.5), and month 12 (1.2, 95% CI: 0.9 to 1.7) compared with baseline, though also imprecise (Figure 2b). Model-building resulted in no adjustment variables, so the results are not presented. Full adjustment resulted in a 6% change in the primary comparison of interest (month 1 versus baseline among newly tested persons).
These findings strongly suggest that HTC, and particularly couples HTC, lead to the adoption of consistent condom use in these stable HIV-discordant couples. At baseline, HIV-infected participants who had just learned their HIV status were much more likely to report unprotected sex (71%) than HIV-infected participants who had known their HIV status for the full month (26%). One month later, after both groups had received couples HTC, the proportion reporting unprotected sex declined from 71% to 8%. In the presence of monthly counseling for the HIV-infected participant, 3 monthly HTC for the HIV-uninfected participant, and condom access for both, low levels of unprotected sex persisted for 1 year (8%).
The protective nature of couples HTC for HIV-discordant couples is consistent with findings from earlier work in Africa. Couples HTC is associated with high condom use among HIV-infected persons,23,24 particularly persons in HIV-discordant relationships.8–12 Our analysis is the first to show that condom use occurs within the first week after couples HTC.9
Our findings further suggest that a couple’s mutual awareness of HIV discordance is more protective than a person’s individual awareness of HIV-positive status. This finding is supported by the modest decline in unprotected sex observed from baseline to month 1 among HIV-infected persons who had received HTC previously. Although these persons had sought HTC before, some may have sought individual HTC and not disclosed to sex partners, learned their partner’s HIV status, or received counseling with partners until just before baseline when they received couples HTC with study partners. The finding that mutual awareness is more protective than individual awareness is complemented by findings from the full trial: HIV-uninfected participants reported less frequent unprotected sex with study partners, whose HIV status was known, than with outside partners, whose HIV status was often unknown.25
After baseline, all HIV-infected persons received individual counseling monthly, partners received individual or couples HTC quarterly, and condoms were provided. These factors may have contributed to consistent condom use, though we cannot determine how influential these factors were compared to the initial impact of couples HTC.
In spite of behavioral prevention, some HIV-infected persons continued engaging in unprotected sex with study partners, without reducing the number of unprotected acts. Reasons for ongoing risk behavior may include fertility desires, condom dislike, or disinhibition, but cannot be assessed formally in these data. Assessing reasons for nonuse, and acceptability of other prevention strategies, is important in this subpopulation.
Understanding the impact of HTC on HIV prevention is critical given its rapid scale-up. However, HTC is difficult to assess in randomized settings because withholding HTC is unethical, and observational studies are typically subject to confounding. This trial provided an opportunity to address the impact of couples HTC on HIV prevention in an ethical rigorous way. Our results are unlikely to be heavily biased by unmeasured confounding because the main difference between exposure groups was the timing of HTC with respect to study enrollment. This typically differed by only a few months and a priori seems unlikely to be strongly influenced by social or biomedical factors. The similar distribution of observed covariates between exposure groups (Tables 1 and 2) and the need for minimal adjustment support this contention.
In our study, we knew the precise timing of when someone learned their HIV status but could only determine when sexual behavior occurred within a 1-month interval. The discrepancy in the timing of these measures leaves ambiguity regarding the temporal order of HTC and sexual behavior for newly tested persons. For example, someone tested 10 days before baseline spent the first 20 days of the month unaware of their HIV status and the final 10 days aware. If they reported unprotected sex during this 30-day period, it could have occurred before, after, or both before and after HTC. Two features of our study design lend evidence to the strong possibility that unprotected sex occurred predominantly before HTC. First, once newly tested persons had been aware of their HIV status for at least 1 month they reported lower levels of unprotected sex. Second, at baseline, the relationship between the amount of time someone was unaware of their HIV status and the odds of unprotected sex was monotonic. The more time someone spent unaware of their HIV status, the more likely they were to report any unprotected sex at baseline. If this trend were to continue a group unaware for the entire month would be expected to experience an even higher probability of unprotected sex than those who were most recently tested.
This analysis relied on self-report, which is subject to social desirability. If persons were more likely to overreport condom use after HTC than before, effect measures would be exaggerated. Biomarkers suggest that these differences are unlikely to be explained entirely by this concern. In the full trial population, consistent condom use was strongly associated with reductions in HIV acquisition.26 Additionally, there was strong correlation (84%) between the HIV-infected and HIV-uninfected partners with respect to the number of self-reported unprotected sex acts at baseline.
Caution is needed when generalizing results beyond these stable HIV-discordant couples. Persons enrolling in HIV prevention trials may be more motivated to adopt HIV prevention behaviors than the general population. Additionally, persons who are willing to enroll with partners may differ from persons who are unwilling. Most couples were in long-term marital or cohabiting relationships, and levels of intimate partner violence were low.27 Couples HTC may not be as protective in segments of the population in less stable more violent partnerships.28 Understanding effectiveness of couples HTC in these less stable partnerships is an area for future investigation.
Our findings raise questions about the current HTC paradigm, which is not typically couple oriented. When stable couples learn that they are in HIV-discordant relationships they adopt consistent condom use quickly, but such marked behavior change is not typically reported after individual HTC.29,30 Couples HTC assures simultaneous disclosure, has a substantial impact on sexual behavior,30 and may have an impact on adherence to biomedical prevention.7,31 However, most current HTC efforts are aimed at individuals, not couples, leading to missed HIV prevention opportunities. Strategies, such as home-based testing,32 supportive HIV-disclosure counseling,33 and partner notification,34,35 can be used to inform persons of HIV discordance. Such couple-oriented strategies have recently been recommended in the World Health Organization’s Guidance on Couples HIV Testing and Counseling.13
In summary, our results add to a growing body of evidence demonstrating that couples HTC is effective at rapidly increasing condom use, facilitating ongoing condom use, and likely lowering rates of HIV transmission.10,11,26 Although initial findings were published nearly 2 decades ago, most countries have been slow to implement couple-based strategies. With expanding HTC capacity in Africa,36 decision-makers now need to consider how to reach couples. Such expansion will help a high-risk group make informed sexual health decisions and likely prevent a substantial number of HIV infections.
The authors thank the teams from the 3 South African study sites and at the University of Washington for their dedicated work. The authors also gratefully acknowledge the HIV-discordant couples who participated in this study.
1. Serwadda D, Gray RH, Wawer MJ, et al.. The social dynamics of HIV transmission as reflected through discordant couples in rural Uganda. AIDS. 1995;9:745–750.
2. Dunkle KL, Stephenson R, Karita E, et al.. New heterosexually transmitted HIV infections in married or cohabiting couples in urban Zambia and Rwanda: an analysis of survey and clinical data. Lancet. 2008;371:2183–2191.
3. Gray R, Ssempiija V, Shelton J, et al.. The contribution of HIV-discordant relationships to new HIV infections in Rakai, Uganda. AIDS. 2011;25:863–865.
5. Baeten JM, Donnell D, Ndase P, et al.. Antiretroviral prophylaxis for HIV prevention in heterosexual men and women. N Engl J Med. 2012;367:399–410.
6. Abdool Karim Q, Abdool Karim SS, Frohlich JA, et al.. Effectiveness and safety of tenofovir gel, an antiretroviral microbicide, for the prevention of HIV infection in women. Science. 2010;329:1168–1174.
7. Cohen MS, Chen YQ, McCauley M, et al.. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med. 2011;365:493–505.
8. The Voluntary HIV-1 Counseling and Testing Efficacy Study Group. Efficacy of voluntary HIV-1 counselling and testing in individuals and couples in Kenya, Tanzania, and Trinidad: a randomised trial. Lancet. 2000;356:103–112.
9. Kamenga M, Ryder RW, Jingu M, et al.. Evidence of marked sexual behavior change associated with low HIV-1 seroconversion in 149 married couples with discordant HIV-1 serostatus: experience at an HIV counselling center in Zaire. AIDS. 1991;5:61–67.
10. Allen S, Tice J, Van de Perre P, et al.. Effect of serotesting with counselling on condom use and seroconversion among HIV discordant couples in Africa. BMJ. 1992;304:1605–1609.
11. Allen S, Serufilira A, Bogaerts J, et al.. Confidential HIV testing and condom promotion in Africa. Impact on HIV and gonorrhea rates. JAMA. 1992;268:3338–3343.
12. Allen S, Meinzen-Derr J, Kautzman M, et al.. Sexual behavior of HIV discordant couples after HIV counseling and testing. AIDS. 2003;17:733–740.
14. Celum C, Wald A, Lingappa JR, et al.. Acyclovir and transmission of HIV-1 from persons infected with HIV-1 and HSV-2. N Engl J Med. 2010;362:427–439.
15. Lingappa JR, Baeten JM, Wald A, et al.. Daily aciclovir for HIV-1 disease progression in people dually infected with HIV-1 and herpes simplex virus type 2: a randomised placebo-controlled trial. Lancet. 2010;375:824–833.
16. Lingappa JR, Lambdin B, Bukusi EA, et al.. Regional differences in prevalence of HIV-1 discordance in Africa and enrollment of HIV-1 discordant couples into an HIV-1 prevention trial. PLoS One. 2008;3:e1411.
17. Lingappa JR, Kahle E, Mugo N, et al.. Characteristics of HIV-1 discordant couples enrolled in a trial of HSV-2 suppression to reduce HIV-1 transmission: the partners study. PLoS One. 2009;4:e5272.
18. Long JS, Freese J. Regression Models for Categorical Dependent Variables Using Stata. College Station, TX: StataCorp LP; 2006.
19. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika. 1986;73:13–22.
20. Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology. 1999;10:37–48.
21. Mickey RM, Greenland S. The impact of confounder selection criteria on effect estimation. Am J Epidemiol. 1989;129:125–137.
22. Rothman KJ, Greenland S, Lash TL. Modern Epidemiology. 3rd edn. Philadelphia, PA: Wolters Kluwer Health/Lippincott Williams & Wilkins; 2008.
23. Turner AN, Miller WC, Padian NS, et al.. Unprotected sex following HIV testing among women in Uganda and Zimbabwe: short- and long-term comparisons with pre-test behaviour. Int J Epidemiol. 2009;38:997–1007.
24. Weinhardt L, Care M, Johnson B, et al.. Effects of HIV counseling and testing and sexual risk behavior: a meta-analytic review of published research, 1985-1997. Am J Public Health. 1999;89:1397–1405.
25. Ndase P, Celum C, Thomas K, et al.. Outside sexual partnerships and risk of HIV acquisition for HIV uninfected partners in African HIV serodiscordant partnerships. J Acquir Immune Defic Syndr. 2012;59:65–71.
26. Hughes JP, Baeten JM, Lingappa JR, et al.. Determinants of per-coital-act HIV-1 infectivity among African HIV-1-serodiscordant couples. J Infect Dis. 2012;205:358–365.
27. Were E, Curran K, Delany-Moretlwe S, et al.. A prospective study of frequency and correlates of intimate partner violence among African heterosexual HIV serodiscordant couples. AIDS. 2011;25:2009–2018.
28. Jewkes RK, Dunkle K, Nduna M, et al.. Intimate partner violence, relationship power inequity, and incidence of HIV infection in young women in South Africa: a cohort study. Lancet. 2010;376:41–48.
29. Becker S, Mlay R, Schwandt HM, et al.. Comparing couples' and individual voluntary counseling and testing for HIV at antenatal clinics in Tanzania: a randomized trial. AIDS Behav. 2010;14:558–566.
30. Kennedy CE, Medley AM, Sweat MD, et al.. Behavioural interventions for HIV positive prevention in developing countries: a systematic review and meta-analysis. Bull World Health Organ. 2010;88:615–623.
31. Ware NC, Wyatt MA, Haberer JE, et al.. What's love got to do with it? Explaining adherence to oral antiretroviral pre-exposure prophylaxis (PrEP) for HIV serodiscordant couples. J Acquir Immune Defic Syndr. 2012;59:463–468.
32. Tumwesigye E, Wana G, Kasasa S, et al.. High uptake of home-based, district-wide, HIV counseling and testing in Uganda. AIDS Patient Care STDS. 2010;24:735–741.
33. Kairania RM, Gray RH, Kiwanuka N, et al.. Disclosure of HIV results among discordant couples in Rakai, Uganda: a facilitated couple counselling approach. AIDS Care. 2010;22:1041–1051.
34. Brown L. HIV partner notification is effective and feasible in sub-Saharan Africa: opportunities for HIV treatment and prevention. J Acquir Immune Defic Syndr. 2011;56:437.
35. Mohlala BK, Boily MC, Gregson S. The forgotten half of the equation: randomised controlled trial of a male invitation to attend couple VCT in Khayelitsha, South Africa. AIDS. 2011;25:1535–1541.
Keywords:© 2013 Lippincott Williams & Wilkins, Inc.
HIV; condom; unprotected sex; HIV counseling and testing; discordant couple; South Africa