Monogamy, together with abstinence, partner reduction, and condom use, is widely advocated as a key behavioral strategy to prevent HIV infection in sub-Saharan Africa. We examined the association between the number of sexual partners and the risk of HIV seropositivity among men and women presenting for HIV voluntary counseling and testing (VCT) in northern Tanzania.
Methodology/ Principal Findings
Clients presenting for HIV VCT at a community-based AIDS service organization in Moshi, Tanzania were surveyed between November 2003 and December 2007. Data on sociodemographic characteristics, reasons for testing, sexual behaviors, and symptoms were collected. Men and women were categorized by number of lifetime sexual partners, and rates of seropositivity were reported by category. Factors associated with HIV seropositivity among monogamous males and females were identified by a multivariate logistic regression model. Of 6,549 clients, 3,607 (55%) were female, and the median age was 30 years (IQR 24–40). 939 (25%) females and 293 (10%) males (p<0.0001) were HIV seropositive. Among 1,244 (34%) monogamous females and 423 (14%) monogamous males, the risk of HIV infection was 19% and 4%, respectively (p<0.0001). The risk increased monotonically with additional partners up to 45% (p<0.001) and 15% (p<0.001) for women and men, respectively with 5 or more partners. In multivariate analysis, HIV seropositivity among monogamous women was most strongly associated with age (p<0.0001), lower education (p<0.004), and reporting a partner with other partners (p = 0.015). Only age was a significant risk factor for monogamous men (p = 0.0004).
Among women presenting for VCT, the number of partners is strongly associated with rates of seropositivity; however, even women reporting lifetime monogamy have a high risk for HIV infection. Partner reduction should be coupled with efforts to place tools in the hands of sexually active women to reduce their risk of contracting HIV.
Citation: Landman KZ, Ostermann J, Crump JA, Mgonja A, Mayhood MK, et al. (2008) Gender Differences in the Risk of HIV Infection among Persons Reporting Abstinence, Monogamy, and Multiple Sexual Partners in Northern Tanzania. PLoS ONE 3(8): e3075. doi:10.1371/journal.pone.0003075
Editor: Landon Myer, University of Cape Town, South Africa
Received: May 4, 2008; Accepted: August 1, 2008; Published: August 27, 2008
Copyright: © 2008 Landman et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was supported in part by Roche. Additional support was obtained from AIDS Clinical Trials Group (U01 AI-39156; Drs. Bartlett, Crump, and Thielman), Mid-career Investigator Award (K24 AI-0744-01; Dr. Bartlett), International Studies of AIDS -associated Co-infections Award (U01 AI 062563-04; Drs. Bartlett, Crump, and Thielman) and the Duke University Center for AIDS Research (P30 AI64516-03; Drs. Bartlett and Crump) from the National Institute of Allergy and Infectious Diseases, from the U.S. Department of State Fulbright Program (Dr. Thielman), from the Fogarty International Center's AIDS International Training and Research Program (D43 TW006732, Drs. Bartlett, Crump, and Thielman), from the Hubert Yeargan Center for Global Health, Duke University Medical Center (Dr. Tribble) and from the University of Rochester Medical Center (Dr. Landman). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Monogamy has been increasingly advocated as one of the cornerstones of HIV prevention in sub-Saharan Africa. As part of the “ABC” approach—urging abstinence, being monogamous, and using condoms—this strategy has been associated with reduced transmission of HIV among sexually active people both in epidemiological modeling studies  and in other research in developing countries , .
Differences between dynamics of female-to-male and male-to-female HIV transmission  and between trends in sexual behavior among men and women in the developing world ,  suggest that sexual behavior may differentially impact transmission risk among men and women. In one Zimbabwean cohort, for instance, the risk of HIV rose with increasing number of sexual partners among women, but not among men .
We examined the association between the number of lifetime sexual partners and the risk of HIV seropositivity among men and women presenting for HIV voluntary counseling and testing (VCT) in northern Tanzania and further identified risk factors for seropositivity among monogamous women and men in this cohort.
Study population and Data Collection
Study subjects were recruited at a free-standing VCT center operated by Kikundi cha Wanawake Kilimanjaro Kupambana na UKIMWI (KIWAKKUKI; Women Against AIDS in Kilimanjaro), a women-led HIV/AIDS advocacy, education, and home care organization based in Moshi, Tanzania that provides services for men and women. Free VCT services are offered to anyone requesting HIV testing at this free-standing walk-in facility, located one block from a busy regional bus station; the majority of clients seen are self-referred. With the exception of a 3 week national VCT campaign period in October 2007, all clients age 18 or older were approached consecutively for study participation and refusal rates were < 5%. At the time of data collection, KIWAKKUKI's VCT center saw an average of 13 clients each weekday , .
After obtaining informed consent, clients presenting for VCT between November 2003 and December 2007 were interviewed by trained, Tanzanian Kiswahili-speaking counselors using a standardized questionnaire developed for this project. The questionnaire gathered data on sociodemographic characteristics, reasons for testing, and sexual behavior, and included the questions, “How many sexual partners have you ever had in your lifetime?” and “Do you have a sexual partner who has other partners?” In addition, trained VCT counselors asked if clients presented for testing because of illness and if they had symptoms commonly associated with HIV infection in this population (recent fever, cough, hemoptysis, night sweats, diarrhea, genital ulcers/discharge, rash, or weight loss) . Client response data were recorded on paper questionnaires by the counselors.
Data from unique records of first-time clients aged 18 and over were entered into a Microsoft Access (Redmond, Washington, USA) sing either EpiInfo 3.3 (CDC, Atlanta, GA, USA) or Teleform 9.0 (Cardiff, Visa, CA) and analyzed using STATA 10 (StataCorp, College Station, TX, USA). Factors associated with HIV seropositivity were identified by multivariate logistic regression models.
Men and women were categorized by the number of lifetime sexual partners reported, and rates of HIV seropositivity were calculated for categories of subjects reporting 0, 1, 2, 3 or 4, and 5 or more partners. Subjects with 3 or 4 reported lifetime partners, and those with 5 or more partners, respectively, were grouped due to lower numbers in these categories. These categories are identical to those used in a similar analysis by other investigators . We compared trends in HIV seropositivity by number of sexual partners, and across various demographic and behavioral characteristics using Chi square and nonparametric trend tests .
To derive an estimate of the annualized risk of HIV infection, we divided the number of positive tests by the total exposure time. A person's exposure time was defined as the time between the age at sexual debut and the age at the test date, accounting for estimates that the HIV epidemic in Tanzania began in 1983 . To assess the degree to which demographic correlates and having a partner with other partners (partner concurrency) affect the risk of seropositivity, we estimated multivariate logistic regression models separately for each gender, and separately for all respondents and for monogamous respondents only. In sensitivity analysis, we conducted stratified analyses for specific population sub-groups. To assess whether selection bias could explain gender differences in seroprevalence, we examined trends in a group comprised of only asymptomatic subjects.
The protocol for this study was approved by the KCMC Research Ethics Committee, the Tanzania National Institutes for Medical Research National Research Ethics Coordinating Committee, and an Institutional Review Board of Duke University Medical Center. All study participants provided written informed consent prior to data collection.
Of 6,549 clients with at least one lifetime partner, 3,607 (55%) were female. The mean (median; range) age of males was 33 (29; IQR 24–40) years and that of females was 32 (30; IQR 24–40) years (p = 0.8384). Nine hundred and thirty-nine (25%) females were HIV seropositive, whereas 293 (10%) males were seropositive (p<0.0001). Four hundred and twenty-three (14%) males and 1,244 (34%) females reported having had only one lifetime sexual partner (p<0.0001). Women having more than one lifetime sexual partner reported fewer total partners (median 3, IQR 2–4) than similar men did (median 4, IQR 3–7, p<0.0001). After excluding clients missing information on sexual debut or on covariates, 6,104 respondents were included in the multivariate analyses.
As expected, rates of HIV were lowest in abstainers and highest among persons reporting 5 or more sexual partners. The absolute risk of HIV seropositivity increased monotonically with the number of lifetime sexual partners in both men and women. Among respondents with at least one sexual partner, the absolute risk (Figure 1) for HIV seropositivity was higher for women than for men; the risk for women relative to men within each subgroup ranged from 3.0 to 5.4. A trend test of increasing rates of seropositivity with increasing numbers of partners was significant for both men (Z = 7.71, p<0.0001) and women (Z = 12.93, p<0.0001, Figure 1).
Figure 1. Absolute risk of HIV seropositivity among women and men presenting for VCT, by number of lifetime sexual partners in Moshi, Tanzania, 2003–2007, N = 6,531.doi:10.1371/journal.pone.0003075.g001
Among women, subjects between 30 and 39 years old had the highest risk of seropositivity in each category of lifetime sexual partners (Figure 2), while subjects 40 years or older had the greatest rise in risk of seropositivity with increasing numbers of sexual partners. The associations were similar but not as strong among men (data not shown): non-parametric trend tests of associations between the number of partners and seropositivity were significant among men in the youngest and two oldest age groups only (p = 0.016; p = 0.035; and p = 0.001, respectively) and among women showed significant effects in all age groups (p<0.0001 to p = 0.006). Accounting for the beginning of the HIV epidemic in Tanzania in 1983, and the time between subjects' sexual debut and their age at testing (i.e., the exposure time), the annualized risk per 100 person years for monogamous males was 0.51 (95% confidence interval (CI), 0.30–0.86) versus 1.67 (95% CI, 1.47–1.90) for monogamous females, and the risk increased within each category with increasing numbers of sexual partners (Table 1).
Figure 2. HIV seropositivity among women and men presenting for VCT, by number of lifetime partners and age of tester in Moshi, Tanzania, 2003–2007.
Among women, subjects between 30 and 39 years old had the highest risk of seropositivity in each category of lifetime sexual partners (Figure 2, Panel A), while subjects 40 years or older had the greatest rise in risk of seropositivity with increasing numbers of sexual partners. The associations were similar but not as strong among men (Figure 2, Panel B). Non-parametric trend tests of associations between the number of partners and seropositivity were significant among men in the youngest and two oldest age groups only (p = 0.016; p = 0.035; and p = 0.001, respectively) and among women showed significant effects in all age groups (p<0.0001 to p = 0.006).doi:10.1371/journal.pone.0003075.g002
Table 1. Annualized rates of HIV infection among male and female VCT clients in Moshi, Tanzania, 2003–2007.doi:10.1371/journal.pone.0003075.t001
In multivariate analysis, a greater number of sexual partners was associated with seropositivity in women, and to a somewhat lesser extent in men (Table 2). Age was strongly predictive of seropositivity. Females ages 30–39 and males ages 30 and above were most likely to test seropositive. Being widowed and having weekly household expenditures below 7000 Tanzania shillings (TSH; approximately 5.80 United States dollars) were associated with seropositivity in both men and women, while post-secondary education was protective. Having a partner who had other sexual partners or not knowing whether the partner had other partners was associated with seropositivity among monogamous females (odds ratio 1.44, 95% CI, 1.07–1.97) but was not associated with seropositivity in men.
Table 2. Correlates of HIV infection by gender among 6,104 clients presenting for VCT in Moshi, Tanzania, 2003–2007.doi:10.1371/journal.pone.0003075.t002
When assessing HIV seroprevalence in a monogamous population, the partner's risk behaviors are of particular interest. We used parameter estimates from the logistic regression model to calculate the marginal effect of monogamous women's partners having other partners on rates of seropositivity. A monogamous female with a partner who has other partners or who does not know if the partner has other partners is 36% more likely (CI, 7% to 65%, p = 0.016) to be HIV-infected than an otherwise identical female who reports no partners with other partners.
To examine if the gender difference in seropositivity could be explained by high-risk women being more likely to present for testing, we assessed how the HIV risk ratio between women and men varied in more homogenous sub-populations (e.g., age 20–30; not widowed; or no partners with other partners). Despite significant differences in the gender distribution of these sociodemographic risk characteristics, gender differences in rates of seropositivity remained large and statistically significant in all comparisons (p = 0.01 to p<0.001, not shown).
A higher proportion of monogamous women (30%) than men (15%) presented with symptoms or because of illness (p<0.001). However, this difference was not significant among HIV-seropositive clients. Additionally, with restriction of the sample to asymptomatic clients, the risk of HIV infection among monogamous women remained more than 3 times as high as that of men (p<0.001; not shown). In this group, the rate of increase in the risk of HIV infection as numbers of partners increased was similar between men and women (p = 0.9567).
In a large cohort of VCT clients in Moshi, Tanzania, the risk for HIV infection increased monotonically with increasing numbers of clients' sexual partners. The rate of increase was higher among women than among men, and women reporting lifetime monogamy had a significantly higher risk for HIV infection than monogamous men. These findings demonstrate limited protection of monogamy among women and highlight the risk of partner concurrency.
Epidemiological models have suggested that having concurrent sexual partnerships increases the risk for transmission of viral sexually transmitted infections (STI's), including HIV , . Supporting these models, several studies in the United States have demonstrated concurrency and unawareness of partners' concurrency to be risk factors for transmission of syphilis and viral STI's , . While several reports from sub-Saharan Africa have been unable to demonstrate a link between concurrency and the spread of HIV , , we found significantly higher rates of HIV infection among women reporting certain or possibly polygynous partners. Such women were 36% more likely to be HIV-infected than otherwise identical women reporting that their partner had no other partners.
Our study had two primary limitations. First, the study population was comprised entirely of VCT service users, potentially limiting the generalizability of the findings to the general population. To explore the validity of the observed gender differentials in HIV seroprevalence we considered two alternative explanations. One possibility was a differential distribution of demographic correlates of HIV infection by gender, the other a possible selection bias resulting from gender differences in the probability of presenting for VCT only after becoming symptomatic. Results from multivariate analyses suggest that differences in demographic or selected risk characteristics are not likely to be the primary explanations for the gender differences in rates of seropositivity. The differential risk of HIV seropositivity in women relative to men was consistent across demographic subgroups, and persisted regardless of the presence of risk factors. Similar trends were observed in all subgroups, including only asymptomatic clients, suggesting that self-selection for testing was not likely a primary explanation for the higher burden of HIV seropositivity among the women.
The second limitation is the study's reliance on self-reporting, which may have resulted in inaccurate measurement of risk factors. Specifically, one potential explanation for higher HIV seroprevalence among female subjects was that female subjects underreported their sexual behaviors. This possibility is not easy to rule out. In many parts of sub-Saharan Africa, the belief that women's sexual needs are not as strong as men's and the widespread, socially acceptable practice of polygyny have been associated with underreporting of extramarital sexual activity by women –. One study of rural Tanzanians estimated that single women underreport their numbers of sexual partners by 46% . In men, the evidence is mixed, with studies suggesting underreporting of early sexual debut among young Zambian men  and over reporting of extramarital sexual activity in Tanzania . The magnitude of the difference in HIV seroprevalence between men and women in Figure 1 (men with 5 or more partners have a lower rate of HIV infection than monogamous women) suggests that differential underreporting of sexual exposures by women is not likely a primary explanation. If women consistently misrepresented sexual exposures, the association between exposure and seroprevalence would likely resemble a flatter line or a concave curve; the highly significant and nearly linear trend in Figure 1 for females (p<0.0001) argues for the validity of reported differentials in women's sexual exposures.
The observed high rates of HIV infection in monogamous women, and the strikingly large effect of monogamous women's partners having other partners on infection rate suggest that efforts to promote abstinence, to reduce the number of sexual partners, and to promote mutual monogamy should be coupled with methods which empower women to better control their exposure risk. We recommend increased efforts to educate and empower women with respect to condom use , while continuing other efforts to reduce HIV transmission, including the promotion of male circumcision – (and further evaluation of its impact on HIV transmission risk to females) and development of vaginal microbicides , . In a climate where many women are placed at greater risk because of partners' concurrent relationships, an overly simple formulation of the ABC prevention strategy, “if not A, then B, and if not B, then C,” is misguided. Greater means to control the risk for HIV infection should be given to all women, including those who are monogamous.
We are very grateful to the staff of the KIWAKKUKI AIDS Information Centre for their collaboration, and in particular to the counselors.
Conceived and designed the experiments: KZL JAC DI NMT. Performed the experiments: KZL AM MM AT HC. Analyzed the data: KZL JO EN NMT. Contributed reagents/materials/analysis tools: JO DI JS JAB. Wrote the paper: KZL JO NMT.
- 1. Morris M, Kretzschmar M (1997) Concurrent partnerships and the spread of HIV. AIDS 11: 641–648.
- 2. Shelton JD, Halperin DT, Nantulya V, Potts M, Gayle HD, et al. (2004) Partner reduction is crucial for balanced “ABC” approach to HIV prevention. BMJ 328: 891–893.
- 3. Singh S, Darroch J, Bankole A (2003) A, B and C in Uganda: The Roles of Abstinence, Monogamy and Condom Use in HIV Decline. New York: The Alan Guttmacher Institute.
- 4. O'Farrell N (2001) Enhanced efficiency of female-to-male HIV transmission in core groups in developing countries: the need to target men. Sex Transm Dis 28: 84–91.
- 5. Slaymaker E, Buckner B (2004) Monitoring trends in sexual behaviour in Zambia, 1996–2003. Sex Transm Infect 80: Suppl 2ii85–90.
- 6. Gregson S, Garnett GP, Nyamukapa CA, Hallett TB, Lewis JJC, et al. (2006) HIV Decline Associated with Behavior Change in Eastern Zimbabwe. Science 311: 664–666.
- 7. Thielman NM, Chu HY, Ostermann J, Itemba DK, Mgonja A, et al. (2006) Cost-effectiveness of free HIV voluntary counseling and testing through a community-based AIDS service organization in Northern Tanzania. Am J Public Health 96: 114–119.
- 8. Chu HY, Crump JA, Ostermann J, Oenga RB, Itemba DK, et al. (2005) Sociodemographic and clinical characteristics of clients presenting for HIV voluntary counselling and testing in Moshi, Tanzania. Int J STD AIDS 16: 691–696.
- 9. Cuzick J (1985) A Wilcoxon-type test for trend. Stat Med 4: 87–90.
- 10. Ministry of Health (2005) National AIDS Control Programme HIV/AIDS/STI Surveillance Report: January - December 2004. Report Number 19.
- 11. Doherty IA, Shiboski S, Ellen JM, Adimora AA, Padian NS (2006) Sexual bridging socially and over time: a simulation model exploring the relative effects of mixing and concurrency on viral sexually transmitted infection transmission. Sex Transm Dis 33: 368–373.
- 12. Drumright LN, Gorbach PM, Holmes KK (2004) Do people really know their sex partners? Concurrency, knowledge of partner behavior, and sexually transmitted infections within partnerships. Sex Transm Dis 31: 437–442.
- 13. Koumans EH, Farley TA, Gibson JJ, Langley C, Ross MW, et al. (2001) Characteristics of persons with syphilis in areas of persisting syphilis in the United States: sustained transmission associated with concurrent partnerships. Sex Transm Dis 28: 497–503.
- 14. Ferry B, Carael M, Buve A, Auvert B, Laourou M, et al. (2001) Comparison of key parameters of sexual behaviour in four African urban populations with different levels of HIV infection. Aids 15: Suppl 4S41–50.
- 15. Lagarde E, Auvert B, Carael M, Laourou M, Ferry B, et al. (2001) Concurrent sexual partnerships and HIV prevalence in five urban communities of sub-Saharan Africa. Aids 15: 877–884.
- 16. Mitsunaga TM, Powell AM, Heard NJ, Larsen UM (2005) Extramarital sex among Nigerian men: polygyny and other risk factors. J Acquir Immune Defic Syndr 39: 478–488.
- 17. Gersovitz M, Jacoby J, Dedy F, Tapé AG (1998) The balance of self-reported heterosexual activity in KAP surveys and the AIDS epidemic in Africa. Journal of the American Statistical Association; Applications and Case Studies 93: 875–883.
- 18. Nnko S, Boerma JT, Urassa M, Mwaluko G, Zaba B (2004) Secretive females or swaggering males? An assessment of the quality of sexual partnership reporting in rural Tanzania. Soc Sci Med 59: 299–310.
- 19. Ahmed S, Lutalo T, Wawer M, Serwadda D, Sewankambo NK, et al. (2001) HIV incidence and sexually transmitted disease prevalence associated with condom use: a population study in Rakai, Uganda. AIDS 15: 2171–2179.
- 20. Gray RH, Kigozi G, Serwadda D, Makumbi F, Watya S, et al. (2007) Male circumcision for HIV prevention in men in Rakai, Uganda: a randomised trial. Lancet 369: 657–666.
- 21. Bailey RC, Moses S, Parker CB, Agot K, Maclean I, et al. (2007) Male circumcision for HIV prevention in young men in Kisumu, Kenya: a randomised controlled trial. Lancet 369: 643–656.
- 22. Auvert B, Taljaard D, Lagarde E, Sobngwi-Tambekou J, Sitta R, et al. (2005) Randomized, controlled intervention trial of male circumcision for reduction of HIV infection risk: the ANRS 1265 Trial. PLoS Med 2: e298.
- 23. Smith RJ, Bodine EN, Wilson DP, Blower SM (2005) Evaluating the potential impact of vaginal microbicides to reduce the risk of acquiring HIV in female sex workers. AIDS 19: 413–421.
- 24. Mayer KH, Karim SA, Kelly C, Maslankowski L, Rees H, et al. (2003) Safety and tolerability of vaginal PRO 2000 gel in sexually active HIV-uninfected and abstinent HIV-infected women. AIDS 17: 321–329.