Male homosexual partnerships and anal sex practices have rarely been investigated in Africa, in contrast to their detailed characterization as key risk factors for HIV-1 infection elsewhere . Unprotected penile–vaginal intercourse has been assumed to drive transmission of HIV-1 among adults in Africa, focusing research on high-risk populations such as female sex workers (FSW),  sero-discordant couples, [3,4] and persons in concurrent partnerships . Anal intercourse has rarely been reported as a measured risk factor in African surveillance or research characterizing HIV-1 risk behaviours, despite a continuing search to explain the heterogeneity in severity of epidemics between different populations in Africa [6–8].
Sex between men and participation in sex work remain illegal and socially stigmatizing in many African countries,  and political and cultural barriers make research into anal intercourse practices difficult. In the few studies in which anal intercourse have been included, researchers have been concerned about misunderstanding and measurement error due to stigma . Similarly, the vulnerability to HIV-1 infection of men who have sex with men (MSM) has been largely ignored in Africa, with the first study of HIV-1 amongst MSM from Africa published only in 2005 . Throughout sub-Saharan Africa, there is a distinct lack of services for the prevention, diagnosis, and treatment of diseases transmitted by anal sex, and of interventions targeted toward MSM.
Active populations of MSM have long been known to exist in East Africa . Local terminology for homosexuality exists in Ki-Swahili dictionaries from the nineteenth century  and there are detailed anthropological accounts of transgender identification, same-sex orientation and homosexual behaviour that predate the global emergence of HIV . Since 2005, we have been recruiting volunteers at higher risk for HIV-1 infection into a vaccine preparedness cohort. Unexpectedly, several male volunteers admitted to having sex with men. Our preliminary investigations and those of other researchers suggest that anal intercourse and other sexual acts between men are more common in Kenya than previously thought.
In 2006, we used a capture–recapture methodology to estimate that 739 MSM [95% confidence interval (CI), 690–798] were selling sex in and around Mombasa;  a similar population has been described in Nairobi . Neither of these population-based surveys assessed the prevalence and risk factors for HIV-1 and other sexually transmitted diseases in the Kenyan MSM population. In the current study, we describe socio-behavioural characteristics and risk factors for HIV-1 infection among MSM in the Mombasa area.
Since July 2005, populations in and around Mombasa at higher risk for HIV infection have been targeted for recruitment into an HIV-1 vaccine preparedness cohort. The present study focuses on MSM, a group for which targeted recruitment began in 2005 . Identification and recruitment of potential MSM study participants was conducted by a team of 10–15 trained peer mobilizers, who approached individuals via personal networks and at venues at which MSM meet to establish contact with partners and clients. The criteria for MSM enrollment was a self-report of any anal sex within the last 3 months. The research clinic is situated approximately 20 km north of Mombasa in a busy suburb with regular bus and matatu (minibus) connections. Mombasa is the second largest town in Kenya (population > 900 000), with a busy commercial sea port and an international airport serving large numbers of tourists.
Prospective participants were invited to attend a drop-in centre where they received information about research participation, watched a consent video, and met with a pre-enrolment counsellor. Upon enrolment, a detailed sociodemographic and sexual behavioural history was established by face-to-face interview. Participants were asked if they had previously tested for HIV, and if so the result (if known). Blood was collected for HIV-1 and syphilis screening following the risk assessment.
A standardized medical history and physical examination, including genital (and rectal, if clinically indicated) examination was conducted and recorded by study clinicians. From September 2006, proctoscopy was offered to participants practicing receptive anal sex or experiencing anal symptoms. Syndromic treatment was administered for symptomatic urethritis or proctitis in accordance with World Health Organization treatment guidelines . Participants who were HIV-1-positive at screening were offered enrolment into a parallel HIV-1-positive cohort. Clinical care and referrals were provided irrespective of study participation.
Whole blood was tested for HIV-1 on site using two rapid test kits (Determine, Abbott Laboratories, Abbott Park, Illinois, USA; Unigold, Trinity Biotech plc, Bray, Ireland) in parallel. Discrepant rapid HIV test results were resolved by a fourth-generation HIV-1 antigen/antibody assay performed at the Kenya AIDS Vaccine Initiative Laboratories in Nairobi . Active syphilis was diagnosed by positive rapid plasma reagin (RPR) titre confirmed by the Treponema pallidum haemagglutination assay (TPHA), at KEMRI, Kilifi. Gram staining of discharge was performed when symptoms or signs of an infection were present. Urethritis and proctitis were confirmed by microscopic observation of ≥ 5 white blood cell (WBC) per oil emersion field and attributed to Neisseria gonorrhoea where Gram-negative, intracellular diplococci were identified .
Data management and analysis
Questionnaire, clinical and laboratory data were entered into a secure database. Individual and aggregate data were subjected to routine accuracy checks by research staff and periodic review by study monitors. Data cleaning, recoding and analysis were conducted using Stata 9.2 . Sociodemographic and behavioural risk factors for prevalent HIV-1 infection were summarized for men who reported sex with men exclusively versus men who reported sex with both men and women. Categorical variables were tested using chi-squared or Fisher's exact tests. Nonnormally distributed continuous and categorical data were compared using the Wilcoxon ranksum test and Spearman rank correlation . Where reported, P values are two-tailed.
Logistic regression was performed to estimate odds ratios (OR) for risk of HIV-1 sero-positivity. All variables associated with HIV-1 status in univariate analysis at a P-value < 0.2 were entered into a multivariate model. Variables were evaluated using backward stepwise estimation, and retained in the final model where an association with HIV-1 status persisted (P-value < 0.2). All regression analyses were restricted to persons not known to have tested HIV-1 positive prior to this study (n = 278), to avoid bias caused by behaviour change after status was learned. Analyses of exclusive sex act preference were restricted to participants reporting anal intercourse in the last 3 months (n = 279).
The study was granted approval by the National Ethical Review Committee under Kenya Medical Research Institute (KEMRI). All participants provided written, informed consent.
Between August 2005 and April 2007, 285 men who reported sex with men were identified; 114 men reported sex with men exclusively (MSME) and 171 men reported sex with both men and women (MSMW). MSM formed approximately one-third of the vaccine-feasibility enrolment population screened; other risk groups screened included 339 women and 210 men at high risk of heterosexually acquired HIV-1 infection.
Table 1 presents the sociodemographic characteristics of these men, comparing MSME with MSMW. MSME were younger than MSMW, but had similar levels of education. Only 18% of MSM in either category were formally employed. MSME reported possession of a greater number of material assets than MSMW. Although the majority of MSM were single, MSMW were most likely to be married, presently or previously. Most MSM identified with an organized religion. Although MSME were significantly more likely to do so, there were no significant differences in the religion identified (Catholic, Protestant or Muslim). Over 90% of these men were circumcized.
Sexual risk behaviour
Almost all MSM (98%) reported participation in anal intercourse in the prior 3 months. Of MSM who reported anal intercourse in the last 3 months, 37% (104/279) reported that all episodes had been unprotected by condoms. Over three-quarters of MSM (82%, 224/279) reported at least one episode of unprotected anal intercourse with any partner in the last 3 months. Reported use of condoms did not vary significantly by the type of anal intercourse most practiced.
Table 2 presents information on sexual risk behaviours, comparing MSME with MSMW. There were no differences in numbers of regular or casual partners in the last month. MSMW were significantly more likely to have paid another person for sex. Twelve (4%) MSM reported having been raped in the last 3 months; this was reported more often by MSME than MSMW. MSME were significantly more likely than MSMW to have practiced receptive anal intercourse and significantly less likely to have practiced insertive anal intercourse. A higher proportion of MSME than MSMW reported practicing both insertive and receptive anal intercourse in the past 3 months (not statistically significant).
Transactional sex and intravenous drug use
Most MSM (74%) reported selling sex for money or goods in the previous 3 months and 40% of whom reported buying sex as well. Forty-nine MSM (17%) reported neither selling nor buying sex. MSME and MSMW who reported selling sex had similar numbers of clients in the last week (median one versus one, P = 0.223). MSM selling sex reported higher numbers of casual sexual partners in the last month than MSM not selling sex (median three casual partners versus zero, P < 0.001), and were more likely to report unprotected sex with casual partners in the last week (41 versus 25%, P = 0.013). There was no difference in numbers of regular partners (median one versus one, P = 0.133). Among sex workers, personal asset ownership was inversely correlated with the number of clients in the past week (Spearman's rho –0.17, P = 0.016). Of MSM reporting having sold sex, 44% (86/197) did so on the streets or beach and 34% (66/197) in bars and nightclubs. Where reported, the vast majority of recent MSM clients were Kenyan residents (128/138, 93%) of whom 52% were of African origin, 28% Arab origin and 9% Indian origin. A minority of MSM clients (local or visitor) were of Caucasian origin (9/138; 6%). Only four participants (1.4%) reported the use of intravenous drug use in the last 3 months.
Sexually transmitted diseases
Among 277 participants undergoing genital examination, 62 (22.4%) had urethral discharge at enrolment screening, of which 16.1% were confirmed microscopically (3.2% gonococcal and 12.9% non-gonoccocal urethritis). Prevalence of peri-anal condylomata was significantly higher in men reporting receptive anal intercourse (6.3 versus 0.9% (no receptive anal intercourse), P = 0.032. The same trend was apparent for prevalent peri-anal ulceration, but differences were not significant (3.5 versus 2.0%, P = 0.506, data not shown).
Overall HIV-1 prevalence at enrolment was 24.5% (95% CI, 19.7–30.7%). By contrast, HIV-1 prevalence at cohort enrolment was 31.5% (95% CI, 27–36%) for female sex workers and 12.4% (95% CI, 8–17%) for high-risk heterosexual men. Overall, 25.3% MSM reported previous HIV testing, of whom five MSME and two MSMW disclosed they had tested HIV positive.
Table 2 also presents data on sexually transmitted diseases (STD), comparing MSME with MSMW. There were no significant differences in clinical signs of STD at enrolment between these groups. The HIV-1 prevalence was markedly higher, however, in MSME than in MSMW (43.0 versus 12.3%, P < 0.001). MSME were also more likely than MSMW to have serological evidence of active syphilis (7.0 versus 1.2%, P = 0.009).
Risk factors for HIV-1
Participation in any receptive anal sex in the past 3 months was strongly associated with HIV-1 positivity (unadjusted OR, 4.7; 95% CI, 2.4–9.2). This association persisted when adjusted for age group, religious group, partner preference, anal intercourse without condom (3 months), intravenous drug use (3 months), paying for sex (3 months), peri-anal warts or ulcers, and prior negative testing for HIV-1 (OR, 6.1; 95% CI, 2.4–15.5). In comparison with persons reporting only insertive anal intercourse, the risk of HIV-1 was significantly higher for those reporting only receptive anal intercourse and higher still for those reporting both (Table 3).
Table 3 also shows the analysis of other identified risk factors for prevalent HIV-1 infections. MSME were significantly more likely to be HIV-1 positive in comparison with MSMW (OR, 6.3; 95% CI. 2.3–17, adjusted for other risk factors as above). Clinical evidence of peri-anal condylomata was independently associated with HIV-1 positivity (adjusted OR, 5.1; 95% CI, 1.1–24). An apparent relationship between peri-anal ulceration and prevalent HIV-1 was not statistically significant. Recent intravenous drug use was strongly associated with HIV-1 positivity, but this was rarely reported.
The present study is the first study describing HIV-1 prevalence and risk factors in a large group of East African MSM. Men who admitted to sex with men exclusively had a high HIV-1 prevalence (43.0%), which was significantly higher than bisexual men (12.3%). This high HIV-1 prevalence among MSM contrasts with the 2005 UNAIDS estimate for adult prevalence (15–49 years) of 6.1% (95% CI, 5.2–7.0%) in Kenya . The MSM in this study reported high levels of receptive anal intercourse and very low condom use. The majority of these MSM were paid for sex in the previous 3 months, and commercial sex was associated with greater sexual risk-taking.
Anal receptive sex, practiced by 75% of the MSME and 40% of the MSMW, and both anal receptive and insertive sex, practiced by 25% of MSME and 32% of MSMW, were strongly associated with HIV-1 infection. Preliminary follow-up of MSM who were HIV-1 negative at enrolment supports these main findings. Within a total MSM cohort follow-up period of 92 person-years of observations (pyo), crude HIV-1 incidence was 8.8 per 100 pyo (95% CI, 2.2–35.2) among men reporting insertive anal intercourse only, 12.9 per 100 pyo (95% CI, 4.2–40.0) among men reporting receptive anal intercourse only, and 20.4 per 100 pyo (95% CI, 7.6–53.8) among men reporting both insertive and receptive anal intercourse [17,23].
Exclusive preference for male partners was associated with prevalent HIV-1 infection. HIV-1 prevalence odds were higher for MSM reporting unprotected versus protected anal intercourse in the last 3 months, but this did not achieve statistical significance. We were not able to identify numbers of partners or unprotected sex with partners as risk factors for HIV-1 infection, suggesting limitations of our data. Plausible explanations for the finding that men paying for sexual services in the past 3 months were less likely to be HIV-1-seropositive include more consistent condom use or higher rates of insertive role-taking when sex is purchased. We were not able to explore this further in the present study, however, due to the way our questions were structured.
There are several limitations to this cross-sectional study of MSM enrolling in a vaccine feasibility cohort. Firstly, although our enrolment criteria included any sexually active MSM, our informant-led community mobilization identified mostly MSM who sell sex. Sex workers usually represent a small but more highly sexually active fraction of a wider MSM population. We strongly suspect our findings are not generalizable to all MSM in Mombasa, particularly those with no links to the sex trade. Further study is underway to identify and engage these men and to better our understanding of the links between sex workers, their clients and the general population. Secondly, questionnaire-based recall of sexual behaviour and risk prevention is known to be subject to numerous respondent biases . Recent risk behaviour is likely to be a relatively poor reflection of cumulative long-term sexual risk behaviour and changes in behaviour over time. Although the higher degree of stigma associated with receptive anal intercourse may have resulted in some misclassification of this behaviour, this would likely result in an underestimate of the association between receptive anal intercourse and HIV-1 infection.
Personal accounts of formative sexual experiences of some of these men are comparable with accounts given by homosexual men elsewhere. The observation that recent partner numbers are, however, inversely related to personal asset ownership among sex workers suggests that for some men, participation in same-sex activity may be due to economic hardship. MSME, locally referred to as queens, are more likely to self-identify as ‘gay’. Sex workers who sell sex to both men and women often practice only insertive anal intercourse with men, and are locally referred to as kings. Kings tend to reject notions of same-sex sexuality. Further socio-behavioral research aimed at clarifying sexual role segregation and versatility (both receptive and insertive anal intercourse) will be key in designing and implementing targeted HIV-1 prevention measures, and understanding the role of MSM behaviour in HIV-1 transmission dynamics .
The extremely high HIV-1 prevalence and low condom use among these MSM calls for urgent public health action. Our study rejects the hypothesis that the local MSM sex trade exists to cater for demand by international visitors, and we also directly challenge another prominent local opinion: that same-sex behaviour among men is done to support intravenous drug habits. Although the high risk of HIV-1 among MSM who are intravenous drug users highlights the need for specific intervention, this risk behaviour was reported by only four MSM (1.4%).
That MSM have received little attention as a group vulnerable to HIV-1 infection in Kenya, and in Africa in general, remains puzzling in the face of the historic anthropological record documenting the group's existence . MSM in Africa often face profound social and political denial and stigmatization . That our research unit has gained an enduring acceptance from MSM, despite its relative distance from central Mombasa, owes much to the diligence and enthusiasm of peer workers drawn from the MSM community, but we also note the lack of other safe, ‘MSM-friendly’ health services in the area. Our experience also confirms that MSM can be motivated participants within research programmes,  as loss to follow up is markedly lower than among heterosexual men and women.
In summary, this is the first report of anal sex practice by a large group of MSM in relation to HIV-1 infection in East Africa. Unprotected receptive anal sex is common, especially among MSM who sell sex, and is strongly associated with HIV-1 infection. Our preliminary findings are consistent with early studies of HIV-1 risk and MSM sexual behaviours in Western contexts . Commercial sex work is more common in Kenya, however, due to high rates of unemployment and lack of opportunities for further education and training. Local residents who do not identify as ‘gay’ are often clients of MSM sex workers. We believe there is an urgent need to address the sexual health needs of MSM sex workers and their clients. With promising evidence to suggest that adult HIV-1 prevalence in Kenya may be falling , the provision of appropriate interventions targeting high-risk, marginalized groups will be crucial to further advances in HIV-1 prevention.
We wish to commend the commitment of the MSM participants, peer leader recruiters, and Community Advisory Board. We thank our counselling and clinical staff in Mtwapa for their hard work and dedication; Scott Geibel, Population Council Nairobi, for advice on MSM recruitment strategies; Dr A. Kahindi, Provincial Medical Officer, Coast Province, and Dr Esther Getambu, National AIDS Control Council, for support and advice; Helen Thomson and Melanie Onyango, International Aids Vaccine Initiative, Nairobi, for useful input and monitoring; and Mr. Bashir Farah, Kenya AIDS Vaccine Initiative, Nairobi, for confirmatory HIV-1 testing.
Sponsorship: Financial support for this study was provided by the International AIDS Vaccine Initiative (IAVI), New York, USA, and the U.S. Agency for International Development (USAID).
This paper was published with permission from the Director of KEMRI.
1. Brody S, Potterat JJ. Assessing the role of anal intercourse in the epidemiology of AIDS in Africa. Int J STD AIDS 2003; 14:431–436.
2. Baeten JM, Richardson BA, Martin HL Jr, Nyange PM, Lavreys L, Ngugi EN, et al
. Trends in HIV-1 incidence in a cohort of prostitutes in Kenya: implications for HIV-1 vaccine efficacy trials. J Acquir Immune Defic Syndr 2000; 24:458–464.
3. Allen S, Meinzen-Derr J, Kautzman M, Zulu I, Trask S, Fedeli U, et al
. Sexual behavior of HIV discordant couples after HIV counselling and testing. AIDS 2003; 17:733–740.
4. Wawer MJ, Gray RH, Sewankambo NK, Serwadda D, Xianbin L, Laeyendecker O, et al
. Rates of HIV-1 transmission by coital act, by stage of HIV-1 infection, in Rakai, Uganda. J Infect Dis 2005; 191:1403–1409.
5. Halperin DT, Epstein H. Concurrent sexual partnerships help to explain Africa's high HIV prevalence: implications for prevention. Lancet 2004; 364:4–6.
6. Hardy D. Cultural practices contributing to the transmission of human immunodeficiency virus in Africa. Revf Infect Dis 1987; 9:1109–1119.
7. Carael M, Holmes K. Dynamics of HIV epidemics in sub-Saharan Africa: introduction. AIDS 2001; 15(suppl 4):S1–S4.
8. Buve A, Carael M, Hayes R, Auvert B, Ferry B, Robinson NJ, et al
. The multicentre study on factors determining the differential spread of HIV in four African cities: summary and conclusions. AIDS 2001; 15(suppl 4):S127–S131.
9. McKenna N. On the margins: men who have sex with men and HIV in the developing world. London: Panos Institute; 1996.
10. Ferguson A, Morris C. Assessing the role of anal intercourse in the epidemiology of AIDS in Africa [letter]. Int J STD AIDS 2003; 14:856.
11. Wade AS, Kane CT, Diallo PAN, Diop AK, Gueye K, Mboup S, et al
. HIV infection and sexually transmitted infections among men who have sex with men in Senegal. AIDS 2005; 19:2133–2140.
12. Kibicho W. Tourism and the sex trade: Roles male sex workers play in Malindi, Kenya. Tourism Rev Int 2003; 7:129–141.
13. Krapf J. A dictionary of the Swaheli language. London: Trubner and Co; 1882.
14. Shepherd G. Rank, gender and homosexuality: Mombasa as a key to understanding sexual options. In: Caplan P, editor. The cultural construction of sexuality.
London, Tavistock; 1987. pp. 240–270.
15. Geibel S, van der Elst EM, Kingola N, Luchters S, Davies A, Getambu EM, et al
. ‘Are you on the market?: A capture-recapture enumeration of men who sell sex to men in and around Mombasa, Kenya. AIDS 2007; 21:1349–1544.
16. Onyango-Ouma W, Birungi H, Geibel S. Understanding the HIV/STI risks and prevention needs of men who have sex with men in Nairobi, Kenya. In: Horizons final report
. Washington, DC: Population Council; 2005, pp. 1–44.
17. Sanders EJ, Graham S, van der Elst E, Mwangome M, Mumba T, Mutimba S, et al.
Establishing a high risk HIV-negative cohort in Kilifi, Kenya. AIDS Vaccine 2006 Conference
. Amsterdam, August 2006 [abstract 470.00].
18. World |Health Organization. Guidelines for the management of sexually transmitted infections.
Geneva, World Health Organization; 2003.
19. Binsbergen J van, Siebelink A, Jacobs A, Keur W, Bruynis F, Graaf F van de, et al.
Improved performance of sero-conversion with a 4th generation HIV antigen/antibody assay. J Virol Methods
20. StataCorp. 2007. Statistical Software: Release 9.2. College Station, Texas: Stata Corporation.
21. Armitage P, Berry G, Matthews J. Statistical methods in medical research, 4th edn. United Kingdom: Blackwell Science Ltd; 2002.
23. Sanders EJ, Graham S, Okuku HS, van der Elst EM, Muhaari A, Davies A, et al.
Risk factors for HIV-1 infection and high HIV-1 incidence of men who have sex with men, in and around Mombasa, Kenya. Trop Med Int Health
(suppl. 1):88 [Abstract O26-59].
24. McKirnan DJ, Doetsch J, Vanable P, Buchbinder S, Douglas JM, Judson F. Preparations for AIDS vaccine trials. Developing brief valid screening instruments for HIV-related sexual risk behaviour among gay and bisexual men. AIDS Res Hum Retroviruses 1994; 10(Suppl 2):S285–S288.
25. Goodreau SM, Goicochea LP, Sanchez J. Sexual role of transmission of HIV type 1 among men who have sex with men, in Peru. J Infect Dis 2005; 191(suppl 1):S147–S158.
26. Esu-Williams E, Hawken M. Identifying the sexual health needs of male sex workers in Kenya. 13th International Conference on AIDS and STIs in Africa
, Nairobi, September 2003 [abstract 469593].
27. Darrow WW, Echenberg DF, Jaffe HW, O'Malley PM, Byers RH, Getchell JP, et al
. Risk factors for human immunodeficiency virus infections in homosexual men. Am J Public Health 1987; 77:479–483.
28. Cheluget B, Baltazar G, Orege P, Ibrahim M, Marum L, Stover J. Evidence for population level declines in adult HIV prevalence in Kenya. Sex Transm Infect 2006; 82(Suppl 1):i21–i26.