In Africa, >85% of adult HIV infections are due to heterosexual transmission.1,2 Although HIV-seronegative partners in HIV-discordant (1 partner HIV positive and 1 partner HIV negative) couples are at a 10% annual risk of acquiring HIV infection3 and a large proportion of new HIV infections in Africa occur in stable partnerships, most HIV prevention programs in Africa focus on reducing the number of casual sexual partners, using condoms during casual sex, and increasing fidelity among married partners. However, for persons in an HIV-discordant relationship, encouraging monogamy without safer sexual behavior will not decrease the risk of HIV transmission. HIV discordance among couples is an increasingly important issue in HIV prevention activities.4 Improving our understanding of behavioral and biologic correlates of HIV transmission in couples should help couples cope with discordant results and improve counseling and care strategies for HIV-discordant couples. In addition to behavioral factors, biologic differences in transmissibility may play a role in explaining the spread and patterns of HIV discordance among couples.
Knowledge of HIV status helps people receiving voluntary HIV counseling and testing (VCT) adopt risk reduction strategies appropriate for their personal circumstances and relationships.5 The AIDS Information Center (AIC) is the largest provider of VCT in Uganda. Over 700,000 clients have been served by AIC since 1990. In 1992, only 9% of clients came for VCT with their partners, but by 2000, 28% of clients at AIC tested with their partners. In 2000, ∼9% of all couples and 18% of married couples were HIV discordant. In addition, 74% of all HIV-positive clients presenting at the clinic with their partners had a negative partner.6 HIV discordance is difficult for counselors to explain and presents a challenge to VCT providers to design appropriate behavioral intervention messages. To determine behavioral and biologic correlates associated with HIV concordance (both partners HIV infected) in couples, we assessed the risk factors among HIV-discordant and HIV-concordant couples receiving VCT.
We conducted a cross-sectional study between January and March 1998, offering enrollment to all persons requesting VCT as a couple at the Kampala Branch of AIC. Only couples older than 18 years of age, testing for the first time, and choosing to receive their HIV test results together were eligible for inclusion. Interviews regarding risk behavior were conducted individually with each participant, but couples were counseled together about their HIV status, sexually transmitted diseases (STDs), and risk reduction strategies. We asked questions about highest level of education attained, occupation, history of blood transfusion and domestic violence, whether couples were living together, condom use, and desire to have children. None of the participants in this study was taking antiretroviral therapy.
The Uganda National Council of Science and Technology, the Science and Ethics Committee of the Uganda Virus Research Institute, and the Centers for Disease Control and Prevention Institutional Review Board approved the study protocol.
Plasma samples were tested at the AIC laboratory using the standard AIC sequential testing algorithm (Capillus HIV-1/HIV-2, Trinity Biotech, Bray, Co Wicklow, Ireland; SeroCard HIV, Trinity Biotech, Bray, Co Wicklow, Ireland; and Multispot HIV-1/HIV-2, Sanofi Diagnostics, Marnes la Coquette, France), and the results were confirmed at the Uganda Virus Research Institute using a parallel enzyme immunoassay algorithm: Recombigen HIV-1/HIV-2 (Cambridge Biotech Corporation, Worcester, MA) and Wellcozyme HIV Recombinant (Wellcome Diagnostics, Dartford, UK). Specimens negative by both screening enzyme immunoassays were considered negative; specimens positive by both assays were considered positive. Discordant enzyme immunoassay results were retested by Western blotting (Cambridge Biotech Corporation, Rockville, MD).7 HIV loads for HIV-positive partners were quantified by reverse transcriptase polymerase chain reaction analysis using Amplicor HIV-1 Monitor version 1.5 (Roche Molecular Systems, Branchburg, NJ), and CD4+ T-cell counts were determined using FACSCount (Becton Dickinson, San Jose, CA). Syphilis serologic analysis was performed using the rapid plasma reagin test and the Treponema pallidum hemagglutination assay (Biotech Laboratories, Ipswich, UK). Only individuals positive by both the rapid plasma reagin test and the T. pallidum hemagglutination assay were considered to have active syphilis. Herpes simplex virus type 2 (HSV-2) serostatus was established using Focus HSV-2 (Focus Technologies, Cypress, CA, USA). Urine samples were tested for Neisseria gonorrhoeae and Chlamydia trachomatis using the ligase chain reaction (Abbott Laboratories, Chicago, IL).
Data Collection and Analysis
All study-specific interview data were collected using standardized questionnaires administered by trained interviewers. Physical examinations were conducted on all participants, and circumcision status for men was verified. Treatment was provided for STDs using standardized syndromic treatment algorithms following Ugandan Ministry of Health guidelines. Analyses were restricted to couples in whom both partners were infected with HIV (concordant) or 1 partner was infected with HIV and 1 partner was HIV negative (discordant). The outcome variable was the couple's HIV status (concordant or discordant). None of the enrolled study participants came in for counseling with >1 partner. Couples with no history of sexual activity within the relationship were excluded from the analysis of potential risk factors associated with being in an HIV-concordant couple. For behavioral and biologic characteristics associated with couple HIV status, we compared proportions of each behavioral characteristic by couple HIV status. Domestic violence was defined as whether partners ever hit, slapped, or pushed each other; the female partner's report was used for this analysis. Living together and circumcision status were treated as couple variables. CD4+ cell count was treated both as a continuous variable to compare the median absolute differences and as a categorical variable (<200, 200-499, and ≥500/μL) when testing for interactions in the multivariate model.
Data management and statistical analysis were conducted using Epi-Info (CDC-USD, Stone Mountain, GA) and STATA (Stata Corporation, College Station, TX). Nonparametric methods were used to test for differences in continuous variables whose distributions were not normal. The median CD4 cell count and median viral load by couple HIV status were tested using the Wilcoxon rank sum test. Univariate and multivariate logistic regression models testing for the probability of being in an HIV-concordant relationship given you are man or a woman and exposed to a given set of covariate(s) were used. Coefficients were tested using the Wald test. We analyzed models with covariate interactions using the likelihood ratio test that compared models with and without interaction terms. Factors that were significant in univariate analysis and those found significant in other studies were included in separate multivariate logistic models for men and for women.
Forty-nine HIV-concordant and 126 HIV-discordant couples were enrolled in this study; 24 HIV-discordant couples were excluded from the analysis because 20 had no history of sexual activity and 4 had contradictory responses to questions on sexual activity. Women were HIV positive in 60 (59%) of 102 HIV-discordant couples and were younger (median age, 29 years; range, 16-51 years) than men (median age, 35 years; range, 21-71; P < 0.001). No age difference was observed between persons with and without HIV infection (P = 0.2). Differences in age between partners were not associated with couple HIV status. A comparison of male and female responses within a couple regarding sexual behavior indicated 90% agreement. Of all men in either HIV-concordant or HIV-discordant couples, 27 (18%) were circumcised. The proportion of circumcised men was higher among HIV-discordant couples (24%) than among HIV-concordant couples (6%) (P = 0.009). Of 24 circumcised men in HIV-discordant couples, 16 (67%) were HIV negative. The proportion of urine samples positive for N. gonorrhoeae and C. trachomatis was <1% for members of both HIV-discordant and HIV-concordant couples.
For men, living together with partner (odds ratio [OR], 12.3; 95% confidence interval [CI], 2.8-53.7) and being uncircumcised (OR, 4.7; 95% CI, 1.3-16.5) were associated with being in an HIV-concordant couple (Table 1). The proportion of men with HSV-2 infection was higher among HIV-concordant couples (86%) than among HIV-discordant couples (61%) (OR, 3.9; 95% CI, 1.5-9.5; P = 0.006). A history of other STDs in the 6 months before enrollment was not associated with couple HIV status. There was no evidence to suggest that education, occupation, history of domestic violence, and history of blood transfusion were associated with couple HIV status.
The median CD4 cell count at enrollment was 250/mm3 for men in HIV-concordant couples compared with 467/mm3 for HIV-positive men in HIV-discordant couples (Wilcoxon rank sum test, P = 0.0072). The risk of being in an HIV-concordant relationship decreased with increasing CD4 cell count for men (χ2 test for trend, 5.8; P = 0.018). HIV-positive men in an HIV-discordant relationship had a significantly lower median viral load than did men in HIV-concordant couples (60,790 copies/mL [interquartile range, 15,346-209,208 copies/mL] vs. 132,212 copies/mL [interquartile range, 31,954-365,319 copies/mL], respectively; P < 0.001), and the odds of being in an HIV-concordant relationship increased with each log increase in viral load (OR, 2.8; 95% CI, 2.0-4.0).
For women, living together (OR, 12.3; 95% CI, 2.8-53.7), having an uncircumcised male partner (OR, 4.7; 95% CI, 1.3-16.5), desiring more children (OR, 0.2; 95% CI, 0.07-0.8), not using a condom during the last sexual act with partner (OR, 2.4; 95% CI, 1.1-5.1), and self-reporting an STD in the 6 months before enrollment (OR, 2.8; 95% CI, 1.4-5.8) were associated with couple HIV status. There was a 2-fold increase in risk of being in an HIV-concordant couple for each log increase in viral load for women. The risk of being in an HIV-concordant relationship decreased with increasing CD4+ cell count for men (χ2 test for trend, 5.8; P = 0.018).
The association between couple HIV status and living together remained for both men (OR, 11.3; 95% CI, 2.2-57.9) and women (OR, 19.0; 95% CI, 3.9-92.1) in the multivariate analysis (Table 2). Uncircumcised men (OR, 4.5; 95% CI, 1.1-18.8), women with uncircumcised partners (OR, 6.5; 95% CI, 1.6-26.4), and women self-reporting an STD within the 6 months before enrollment (OR, 1.9; 95% CI, 0.9-4.5) were more likely to be in an HIV-concordant couple. HIV load was associated with being in an HIV-concordant couple for men (OR, 3.0; 95% CI, 2.0-4.0) and women (OR, 2.2; 95% CI, 1.5-3.2). There was no effect modification between age, sex, and the effect of viral load on couple HIV status.
Several preventable behavioral and biologic factors were associated with being in an HIV-concordant relationship compared with an HIV-discordant relationship. Living together, having a high HIV load, being uncircumcised for men, and reporting an STD within the 6 months before enrollment for women were associated with being in an HIV-concordant relationship. These results support findings from other studies in which similar risk factors were reported.3,8,9
Living together was a risk factor for being in an HIV-concordant relationship and was most likely a proxy for frequency of sexual contact and duration of relationship. One of the difficulties with counseling HIV-discordant couples is that the increased risk for infection comes from within a stable relationship rather than from outside sexual partners. It is important to provide HIV-discordant couples with information that being in a monogamous stable relationship does not mean that partners are not at risk for HIV transmission. VCT for couples, including married couples, is a critical component of prevention activities.
There is substantial epidemiological evidence that uncircumcised men are at a greater risk of becoming infected with HIV than are circumcised men. Several cross-sectional and cohort studies found that circumcision was protective against HIV infection.10-13 In our study, circumcised men were more likely to be in HIV-discordant relationships. The protection from HIV infection of circumcised men most likely reflects removal of mucosal surfaces and access to cells receptive to HIV. However, there is no conclusive epidemiological evidence that introducing male circumcision among adults as a public health intervention will reduce HIV transmission.14 In addition, other factors associated with circumcision, such as ethnic and religious differences,15,16 may confound the observed association. Currently, 3 randomized, controlled trials are being conducted in Africa to evaluate whether circumcision during adulthood reduces HIV transmission.
STDs may enhance HIV transmission by increasing the concentration of HIV in genital secretions, increasing the number of cells receptive to HIV or the number of receptors per cell. In this study, women with self-reported histories of STDs were approximately twice as likely to be in an HIV-concordant relationship. Other studies17 also showed that the proportion of men and women with histories of N. gonorrhoeae and C. trachomatis infections were higher among HIV-concordant couples, although these infections were rare in our study. In Uganda, >50% of all persons are infected with HSV-2.18 HSV-2 has been associated previously with both acquisition and transmission of HIV and may be a major factor associated with the spread of HIV in Africa.19 Rapid diagnosis and treatment of STDs in HIV-discordant couples, especially acyclovir suppression therapy for HSV-2 infection, are of paramount importance given the high rates of HSV-2 infection among this population.
A number of studies have shown that genital HIV load is lower but proportional to plasma viral load but that, even with undetectable plasma viral load, HIV is still present in secretions.20,21 However, HIV transmission per sexual act is associated with higher plasma viral load. The infectivity of an HIV-positive partner is thought to vary with disease stage because of higher viral load in acute infection and during advanced stages.22,23 In a study in Uganda, each log increase in viral load was associated with an increase in the risk of transmission by a factor of 2.45.1 The same study found no instances of transmissions by HIV-positive subjects with undetectable plasma viral load. We found that HIV-positive individuals in HIV-concordant relationships had a significantly higher median viral load than did HIV-positive individuals in HIV-discordant relationships. Each log increase in viral load was associated with an increase in the risk of being in an HIV-concordant relationship by a factor of 2.2 for women and 3.0 for men. The most important justification for expanding access to antiretrovirals in Africa is to prolong the life of persons with HIV infection; however, a secondary outcome could be a reduction in the risk of HIV transmission because highly active antiretroviral therapy decreases peripheral blood levels of HIV RNA dramatically.21 In our study, none of the participants were taking antiretroviral therapy.
We did not observe significant differences between HIV-concordant and HIV-discordant couples when we looked at education and occupation. Such components of social economic status may be surrogates for perceptions of protective behavior. Other studies showed that women in HIV-discordant relationships were more likely to be employed and to be more educated and to have higher incomes than women in HIV-concordant couples.18 In a study in Kigali, Rwanda,8 all women who seroconverted were either illiterate or had difficulty reading. This may be explained by the fact that the illiterate people found it difficult to understand educational programs on intervention and in learning the skills needed for effective use of condoms. Because of access to information, educated people may be more likely to apply preventive methods, such as condom use. Radio and community-based dramas as well as print education regarding HIV prevention have been widespread throughout Uganda for several years and may have decreased differential access to prevention messages.
There were some limitations to this study. Because it was cross-sectional with no information on dates of seroconversion, a causal relationship between risk factors and outcomes could not be confirmed. Couples who visit VCT services are more likely to perceive themselves at risk and probably more likely to have been exposed to assessed risk factors than the wider population, reducing generalizability. Last, we did not collect information on alcohol use and drug use, factors often linked with sexual activity and failure to follow safer sex practices. Studies on alcohol use and drug use have indicated that they lead to risky behavior, failure to use condoms, and casual and multiple partner contacts.17 In all the HIV-discordant couples in a study in Kigali, Rwanda,8 in which the HIV-negative partner seroconverted, the man was reported to drink alcohol regularly.
VCT centers are important sites for identifying HIV-discordant couples, primarily in populations with mature HIV epidemics, where the prevalence of HIV infection is common among people not considered at high risk of infection.24 However, >70% of VCT clients present as individuals, and most of these do not know the HIV status of their partners (data not shown). Effective communication campaigns and other strategies to promote couples testing are urgently needed. Once identified, HIV-discordant and HIV-concordant couples can take precautionary measures to prevent transmission to their partners or to prevent acquisition. To date, HIV prevention messages in Uganda have reinforced the idea that “zero-grazing,” faithfulness, and marriage provide safety from HIV infection. Our finding that living together results in a >10-fold risk of HIV transmission within HIV-discordant couples underscores the importance of adding a strong message promoting couples testing as part of the “being faithful” message. This transition to a new generation of more complex prevention messages will be necessary if Uganda is to be able to further reduce the incidence of HIV infection in the context of a mature generalized epidemic. To help control the HIV epidemic in Uganda, HIV-discordant couples identified in VCT programs should be provided with STD diagnosis and treatment, basic preventive care, antiretroviral therapy, and specially designed counseling interventions.25
The authors thank Drew A. Linzer, University of California at Los Angeles, and Ana T. Jefferson, University of California at Berkeley, for critiquing the statistical methods in an earlier version of the article.
1. Quinn TC, Wawer MJ, Sewankambo N, et al. Viral load and heterosexual transmission of human immunodeficiency virus type 1. N Engl J Med
2. Piot P, Bartos M, Ghys PD, et al. The global impact of HIV/AIDS. Nature
3. Hugonnet S, Mosha F, Todd J, et al. Incidence of HIV infection in stable sexual partnerships: a retrospective cohort study of 1802 couples in Mwanza Region, Tanzania. J Acquir Immune Defic Syndr
4. Hudson CP. The importance of transmission of HIV-1 in discordant couples in the population dynamics of HIV-1 infection in rural Africa. Int J STD AIDS
5. Campbell CH, Marum ME, Dillon BA, et al. The role of HIV counseling and testing in the developing world. AIDS Educ Prev
. 1997;9(Suppl 8):92-104.
6. Bahizi J, Kalule J, Wangalwa S, et al. HIV discordance among couples receiving voluntary counseling and testing in Uganda [abstract 10BT5-1]. 12th International Conference on AIDS/STD in Africa (ICASA). Ouagadougou; 2001.
7. Nunn AJ, Biryahwaho B, Downing RG, et al. Algorithms for detecting antibodies to HIV-1: results from a rural Ugandan cohort. AIDS
8. 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
9. Gray RH, Wawer MJ, Brookmeyer R, et al. Probability of HIV-1 transmission per coital act in monogamous, heterosexual, HIV-1-discordant couples in Rakai, Uganda. Lancet
10. Quinn TC, Wawer MJ, Sewankambo N, et al. Viral load and heterosexual transmission of human immunodeficiency virus type 1. N Engl J Med
11. Cameron DW, D'Costa LJ, Mitha GM, et al. Female to male transmission of human immunodeficiency virus type 1: risk factors for seroconversion in men. Lancet
12. Bailey RC, Neema S, Othieno R. Sexual behaviors and other HIV risk factors in circumcised and uncircumcised men in Uganda. J Acquir Immune Defic Syndr
13. Moses S, Bailey RC, Ronald AR. Male circumcision: assessment of health benefits and risks. Sex Transm Infect
14. Van Howe RS. Circumcision and HIV infection: review of literature and meta analysis. Int J STD AIDS
15. Osmond DH, Padian NS, Galss S, et al. Risk factors for hepatitis C virus seropositivity in heterosexual couples. JAMA
16. Padian NS, Shiboski SC, Glass SO, et al. Heterosexual transmission of human immunodeficiency virus (HIV) in northern California: results from a ten-year study. Am J Epidemiol
17. Skurnick JH, Kennedy CA, Perez G, et al. Behavioral and demographic risk factors for transmission of human immunodeficiency virus type 1 in heterosexual couples: report from the Heterosexual HIV Transmission Study. Clin Infect Dis
18. Kamali A, Quigley M, Nakiyingi J, et al. Syndromic management of sexually-transmitted infections and behaviour change interventions on transmission of HIV-1 in rural Uganda: a community randomised trial. Lancet
19. Wald A, Corey L. How does herpes simplex virus type 2 influence human immunodeficiency virus infection and pathogenesis? J Infect Dis
20. Vernazza PL, Gilliam BL, Flepp M, et al. Effect of antiviral treatment on the shedding of HIV-1 in semen. AIDS
21. Zhang H, Dornadula G, Beumont, et al. Human immunodeficiency virus type 1 in the semen of men receiving highly active antiretroviral therapy. N Engl J Med
22. Serwadda D, Gray RH, Wawer MJ, et al. The social dynamics of HIV transmission as reflected through discordant couples in rural Uganda. AIDS
23. Deschamps MM, Pape JW, Hafner A, et al. Heterosexual transmission of HIV in Haiti. Ann Intern Med
24. Painter TM. Voluntary counseling and testing for couples: a high leverage intervention for HIV/AIDS prevention in sub-Saharan Africa. Soc Sci Med
25. 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