Article Text

Original article
HIV prevalence and risk behaviours among men having sex with men in Nigeria
  1. Mike Merrigan1,
  2. Aderemi Azeez2,
  3. Bamgboye Afolabi3,
  4. Otto Nzapfurundi Chabikuli4,
  5. Obinna Onyekwena4,
  6. George Eluwa4,
  7. Bolatito Aiyenigba5,
  8. Issa Kawu2,
  9. Kayode Ogungbemi6,
  10. Christoph Hamelmann4
  1. 1AED, Gaborone, Botswana
  2. 2Federal Ministry of Health, HIV/AIDS Division, Abuja, Nigeria
  3. 3Department of Medical Statistics, University College Hospital, Ibadan, Nigeria
  4. 4Family Health International, Abuja, Nigeria
  5. 5US Centers for Disease Prevention and Control, Abuja, Nigeria
  6. 6National Agency for the Control of AIDS, Abuja, Nigeria
  1. Correspondence to Dr Mike Merrigan, AED Botswana, PO Box 3780, Gaborone, Botswana; mikemerrigan01{at}


Objective To evaluate HIV and syphilis prevalence among men who have sex with men (MSM) in Nigeria, and assess their HIV-related risk behaviours and exposure to HIV prevention interventions.

Methods Cross-sectional study using respondent-driven sampling conducted in Lagos, Kano and Cross River states, Nigeria, between July and September 2007.

Results A total of 879 MSM participated, 293 from each state. Eight participants (1.1%, CI 0.1% to 2.2%) in Cross River, 27 (9.3%, CI 5.7% to 15.4%) in Kano and 74 (17.4%, CI 12.3% to 23.2%) in Lagos tested positive for HIV. No syphilis was detected. The median age was 22 years. MSM reported an average of 4.2 male anal sex partners in the past 6 months. Between 24.4% (Lagos) and 36.0% (Kano) of MSM reported selling sex to other men. Up to 49.7% of MSM reported sex with a girlfriend and ≤6.5% purchased sex from female sex workers. Consistent condom use in commercial sex with other men over the past 6 months ranged from 28.0% (Cross River) to 34.3% (Kano), in contrast to between 23.9% (Kano) and 45.8% (Lagos) for non-commercial sex. Associations with HIV positivity included age in the three states, having been the receptive partner in anal sex in the past 6 months in Lagos and in Lagos and Kano feeling at risk of HIV.

Conclusion Large differentials in HIV prevalence between states together with high mobility, inconsistent condom use and behavioural links with female sex partners, have the potential for further HIV transmission within MSM networks, and between MSM and the general population.

  • MSM
  • Nigeria
  • HIV
  • risk behaviour
  • homosexual

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


With a population of around 140 million people and HIV prevalence among pregnant women estimated at 4.6%, Nigeria has one of the highest numbers of people living with HIV/AIDS in the world—an estimated 2.95 million individuals.1 According to WHO, Nigeria has a generalised epidemic, and as a result, significant prevention resources have been devoted to general population-based interventions. In accordance with tracking generalised epidemics, Nigeria's second-generation HIV/AIDS surveillance system has historically consisted of two main surveys: ANC (antenatal attendees) Sentinel Surveillance (for HIV prevalence), and the National HIV/AIDS and Reproductive Health Survey (for monitoring knowledge and risk behaviours among the general population).2 To supplement these surveys, behavioural surveillance surveys were carried out among high-risk groups in 2000 and 2005.

In 2007, the Federal Ministry of Health recognised the need for more robust data to determine whether certain population groups contribute disproportionately to HIV transmission. Consequently, Nigeria's first Integrated Biological and Behavioural Surveillance Survey3 (IBBSS) was conducted, and it generated state-level HIV prevalence data from a range of groups, including transport workers, female sex workers, military and police. The survey also broke new ground by including injecting drug users and men who have sex with men (MSM) for the first time ever in a national survey. Little is known about MSM in Nigeria, though high HIV prevalence levels have been detected among MSM in Senegal, Sudan and Kenya.4 Studies from neighbouring countries such as Ghana have found that MSM come from all social classes, religions and ethnicities and have risky sexual behaviour.5 Male to male sexual behaviour is illegal in Nigeria,6 and a recent study reported that Nigerian MSM are socially ostracised by culture, religion and political will, leading same-sex social activities to take place in non-commercial, private venues, where risky sex practices are high.7

The purpose of this survey was to provide insights into the burden of HIV, to understand risk behaviour among MSM, and obtain baseline data to guide the evaluation of future interventions. This paper outlines key findings related to MSM from the 2007 IBBSS study in Nigeria.


Research sites

The 2007 IBBSS was a cross-sectional study, surveying MSM in the urban centres of Lagos (southwest), Kano (northwest) and Cross River states (south-south) between July and September 2007. Lagos and Kano were chosen for their large populations (9 million and 9.4 million respectively) and the feasibility of achieving the desired sample size, while Cross River (population 2.9 million)8 was selected to obtain a wider geographical spread.

Sampling design and recruitment

MSM were defined as any man aged ≥18 years who has engaged in anal sex with another man in the 6 months leading up to the survey. Owing to the hidden nature of MSM in Nigeria, the study employed respondent-driven sampling (RDS) to recruit study participants. RDS is a method that combines ‘snowball sampling’ with a mathematical model that weights the sample to compensate for the non-random method of sample recruitment.9 RDS involves a chain-referral method of recruitment whereby members of a target population are recruited by their peers (seeds) already participating in the study. Long referral chains create an equilibrium that balances the tendency of participants to recruit peers who are like them in certain relevant characteristics. The characteristics used in this study to determine whether equilibrium was reached in each state were education, age and HIV status. A total sample size of 293 for each state was planned based on an assumed HIV prevalence of 15%, a design effect of 2.0 and a level of precision of 0.05. Alliance Rights Nigeria, a non-governmental organisation with previous experience in examining HIV/AIDS problems among MSM, worked with the survey teams to identify the initial seeds in urban centres, assisted in screening the study participants for eligibility and contributed personnel who were trained as interviewers for the study.

Ten seeds were initially selected for each site. An attempt was made to diversify the seeds by age and socioeconomic status. Where the recruitment chains terminated before the desired sample size was reached, more seeds were recruited. Each participant was given three serially numbered recruitment coupons to be given to peers personally known to them who were self-identified homosexual or bisexual. When a coupon was rejected, the recruiter filled in a short rejector questionnaire to assess non-response bias and then identified a replacement so as to meet the sample size. Each participant received an incentive of N500 (approximately US$4), representing the estimated cost of transportation to and from the interview venue. The screening process for eligibility was administered by MSM on the survey team and included knowledge of MSM-specific slang and information about MSM peer networks. The coupons were numbered to include the identification number of the original recruiter. A single team of interviewers was used in each state and only one RDS site was established to minimise the risk of duplicating the respondents.

Data collection

A structured, standard and pre-coded questionnaire was used for the behavioural component, which was pilot tested and adapted locally before administration. It captured a variety of demographic, knowledge, attitude and intervention exposure-related variables. The research team consisted of four interviewers, one supervisor and one laboratory technician/counsellor in each state. After central training of trainers, each team underwent 5 days' training on sampling methods, survey tools and ethics. The study used linked anonymous methods. All participants received pre-test and post-test counselling. Respondents testing positive for HIV were referred to a designated antiretroviral therapy site to receive free care and treatment services.

Laboratory testing

The laboratory testing for HIV involved rapid tests (Capillus and Genie II). Syphilis testing involved rapid plasma reagin for screening and Treponema pallidum haemagglutination assay for confirmation, in accordance with the national testing protocol. Samples were sent to the nearest PEPFAR (President's Emergency Plan for AIDS Relief) ‘centre of excellence’ for testing. Samples arriving at testing laboratories were logged into a registry and examined for quality.

Data management and analysis

Data were entered centrally using CS Pro version 3.2 (US Census Bureau). In agreement with the survey protocol, double-data entry was performed with 25% of questionnaires and 100% of biological data. Data analysis involved statistical software packages including RDSAT version 5.6 (available at, accessed 4 July 2010) and STATA version 9.2 (Stata Corporation). We used RDSAT 5.6 to calculate proportion estimates, 95% CIs, homophily and equilibrium proportions. Equilibrium distribution was calculated with a convergence radius of 2% of the sample estimate for essential variables (age, HIV status, education) in all three states. We used STATA 9.2 for bivariate and multivariate analyses, and χ2 or Fisher's exact test for categorical variables. Variables in bivariate logistic regression analyses significant at the level of 0.05 were entered into a multivariate logistic regression model to identify associations with HIV prevalence among MSM, while controlling for potential confounders. The weights for the multivariate analysis were derived by exporting the individualised weights from RDSAT into STATA. The individualised weights were calculated by RDSAT based on each respondent's network sizes and a recent partition analysis performed on the dependent variable in the analysis (HIV status). The weights were then included as probability weights in the bivariate and multivariate analyses. Ethical approval was obtained from the Protection of Human Subjects Committee of Family Health International North Carolina, the US Centers for Disease Prevention and Control and the National Institute for Medical Research, Nigeria.


Table 1 summarises the main results for MSM in the survey. A total of 297, 315 and 293 MSM were contacted in Lagos, Kano and Cross River State, respectively. In Lagos four MSM refused to participate in the study and returned coupons, compared with 22 in Kano, and none in Cross River. No ineligible individuals presented with coupons and five questionnaires (0.6%) were excluded from the analysis owing to incomplete data. The desired sample size of 293 MSM was reached in each state (879 overall); only 43 (4.9%) respondents refused HIV testing. Acceptance of an HIV test was not a condition of enrolment. The response rates were high, at 98.6%, 93% and 100% for Lagos, Kano and Cross River state, respectively.

Table 1

Characteristics of men who have sex with men, HIV prevalence, sexual behaviours and HIV prevention intervention exposure*

In Lagos 11/14 seeds actively recruited other MSM, with varying length of recruitment chains. Lagos state had a maximum of 10 waves of recruitment with an average of four waves for each active seed. In Lagos, equilibrium was reached for HIV status by the third wave and by the fourth wave for age and educational status. In Kano, 6/10 seeds actively recruited, with a maximum of nine waves and an average of six waves for each active seed. Equilibrium was reached in Kano by the fourth wave for age and by the second wave for HIV status and educational status. Cross River had 9/10 seeds actively recruiting, with a maximum of eight waves and an average of four waves for each seed. Equilibrium was reached in Cross River by the fourth wave for HIV status, by the first wave for age and by the third wave for educational status.

The median age was 22 years in all three states. There were wide variations in HIV prevalence, with adjusted estimates ranging from 17.4% (74), to 9.3% (27) and 1.1% (8) in Lagos, Kano and Cross River, respectively. No syphilis was detected among the participants. Experiences with symptoms of sexually transmitted infections in the past 12 months were more common, with 2.1% of MSM in Lagos and 6.2% of MSM in Cross River reporting experiencing genital discharges and ulcers/sores. MSM in Kano were more likely to report experiencing genital discharge (4.2%) than genital ulcers/sores (1.7%).

MSM were mobile. Half the participants had spent more than one continuous month away from their home state in the past year. Only a small number of MSM reported being married (7, 9 and 18 in Cross River State, Lagos and Kano, respectively). Multiple partnerships were common, with a reported average of 4.2 male anal sex partners per respondent in the past 6 months. A sizeable proportion of MSM reported selling sex to other men (24.4% in Lagos, 35.0% in Cross River and 36.0% in Kano). Similar proportions of MSM in Lagos and Cross River reported using condoms at last non-commercial anal sex (unadjusted), compared with those reporting condom use at last commercial sex (unadjusted). This was different in Kano, where a higher proportion of MSM reported using condoms at last commercial sex with another man (53.2%) compared with non-commercial sex (39.4%). In Kano, 23.9% of MSM reported using condoms consistently in non-commercial anal sex over the past 6 months, compared with 30.0% in Cross River and 45.8% in Lagos. There were no significant differences in condom use at last sex or consistent condom use in the past 6 months between those infected and those not infected with HIV. Girlfriends were the most common female sex partner for MSM surveyed. Of the MSM in Lagos, 49.7% reported sex with a girlfriend in the past 6 months, compared with 26.2% of MSM in Kano and 16.9% of MSM in Cross River. Consistent condom use with girlfriends was higher in Lagos (58.8%) than in Kano (23.0%) and Cross River (22.9%). Table 2 describes the types of sexual partners reported by MSM and association with reports of selling sex. For Lagos and Kano states, the odds of MSM selling sex reporting sex with girlfriends and female sex workers was around three times higher than those not selling sex. This can be contrasted with MSM who sold sex in Cross River state, where no significant differences in sex with these types of female partners were detected.

Table 2

Types of sexual partners reported by men who have sex with men and associations with reports of selling sex

Less than half of MSM surveyed in each state reported feeling at risk of HIV. This was highest in Lagos (41.5%) and lowest in Cross River (15.2%) (Table 1). The proportion of MSM reporting ever having an HIV test was highest in Lagos (40.9%), and lowest in Kano (17.7%), where MSM were also the least likely to have received HIV information from a health worker in the past year (8.1%). Radio and television were the primary sources of HIV information for MSM in the previous 12 months, with over 60% of MSM in each state receiving information through these channels. In Lagos 69.9% of MSM had obtained cheap/free condoms in the past 12 months, compared with 17.4% in Kano and 76.4% in Cross River.

Table 3 outlines results from a multivariate logistic regression used to investigate associations with HIV infection. While each state had age, sex with a receptive partner, sex with an insertive partner, educational level, feeling at risk of HIV and condom use at last anal sex included in the multivariate analysis model, a bivariate analysis yielded different significant independent variables (p<0.05) for each state, which were used in the logistic regression. For Lagos state, the odds of those aged ≥25 years being HIV positive was nearly six times higher than for MSM aged <25 years (OR=5.6 95% CI 2.2 to 14.1, p<0.001). Higher proportions of MSM who were the receptive partner in anal sex were HIV positive than the non-receptive partner in anal sex (OR=4.2, 95% CI 1.5 to 12.3, p=0.008). HIV prevalence was also higher among MSM who felt at risk of HIV compared with those who did not feel at risk (OR=4.4, 95% CI 1.6 to 12.1, p=0.004). For Cross River, the odds of MSM aged ≥25 years being HIV positive was over five times higher than younger MSM (OR=5.8, 95% CI 0.8 to 44.2, p=0.088). In Kano, there was also an association between age and HIV prevalence (OR=6.5, 95% CI 1.6 to 27.1, p=0.01), albeit weaker than with the associations observed in Lagos and Cross River. In contrast to Lagos state, MSM feeling at risk of HIV in Kano had lower HIV prevalence than those not feeling at risk of HIV (OR=0.2, 95% CI 0.1 to 0.8, p=0.02).

Table 3

Multivariate analysis of risk factors for HIV among men who have sex with men*


Results from the 2007 IBBSS suggest that MSM are highly affected by HIV in parts of Nigeria. The HIV prevalence among MSM surveyed is in contrast to a national HIV prevalence of 4.6% among pregnant women obtained through the national ANC survey in 2008, where wide variations at state level were also apparent (5.1% in Lagos, 2.2% in Kano and 8% in Cross River).1 The HIV prevalence among MSM in Lagos was second only to that in brothel-based sex workers (23.5%), while in Kano the HIV prevalence among brothel- and non-brothel-based sex workers was significantly higher (49.1% and 44.1%, respectively), and MSM were only slightly less affected by HIV than injecting drug users (10%). In Cross River, HIV prevalence among MSM was the lowest among all groups surveyed. The syphilis findings are not unusual when compared with very low levels of syphilis detected in other groups, including 0.5% prevalence among non-brothel-based sex workers and armed forces personnel.

The inability to reach older MSM is a limitation of the survey. Attempts were made to vary the characteristics of the seeds according to age and occupation, but with a median age of just 22 years, the sample of MSM in the 2007 IBBSS appears consistent with findings of a previous study in Nigeria suggesting that MSM recruited by RDS are significantly younger than men recruited by other methods.10 Given the association between HIV prevalence and age in the multivariate analysis for two of the three states, HIV prevalence in a more representative sample may be higher than that detected in this study. The nominal incentive given to MSM participants in the 2007 IBBSS was based on estimated transportation costs in Nigeria. While recognising the potential limitation or bias financial incentives and recruitment patterns have on the results, RDS remains the best available method for sampling MSM in Nigeria, and other hidden populations.11 12

Although data on the nature and coverage of HIV prevention services targeting MSM in Nigeria are not available from published sources, it is likely that interventions have been ongoing in Lagos with more intensity and a longer duration than other states surveyed. Considering the age of respondents, their high mobility, frequent partner exchange and low condom use in anal sex, it is possible that HIV prevalence among MSM in Kano and Cross River will increase towards the levels seen in Lagos. Education levels of MSM were similar to those of the military and police in Lagos and Cross River (with over 80% completing a secondary education). Considerably lower proportions of MSM in Kano had completed a secondary education, though there was no significant relationship between educational attainment and HIV prevalence. The relatively high levels of exposure to HIV information through mass media do not appear to translate into a heightened sense of personal risk for MSM or more consistent condom use, a finding consistent with a study among MSM in Uganda.13 Exposure to interpersonal communication with health workers or peer educators, referrals to counselling and testing and sexually transmitted infections services, were reportedly very low, signifying the need for more targeted and appropriate prevention-related communication and services for MSM in Nigeria.

The results share characteristics with studies from other African countries in pointing to an emerging concentrated epidemic among Nigerian MSM, and highlighting risk behaviours which may fuel HIV transmission between MSM, MSM and other most-at-risk populations (eg, female sex workers)14 and the general population. The study has made an important contribution to increasing knowledge of the dynamics surrounding male-to-male sexual risk behaviour in Nigeria. The use of RDS for study recruitment combined with a careful screening process facilitated access to this hard-to-reach population who are at risk of discrimination, stigmatisation and incarceration if openly identified. Further research is needed on MSM in Nigeria, with the aim of better targeting effective and acceptable prevention interventions. Given the stigma attached to male-to-male sexual behaviour, the release of these results needs to be linked to policy and programming initiatives which create an enabling environment for promoting access to services and safer behaviours,15 recognising that striving for a lower HIV prevalence among MSM will have important benefits for Nigeria as a whole.

Key messages

  • Men who have sex with men (MSM) in Nigeria are highly affected by HIV/AIDS and condom use in anal sex is low in commercial and non-commercial partnerships.

  • HIV is more established among MSM in Lagos; high mobility among MSM underlines the potential for a worsening epidemic among MSM in other states.

  • Significant proportions of Nigerian MSM have unprotected sex with female partners, indicating a potential bridge for HIV transmission between MSM and the general population.


The study was conducted by the HIV/AIDS division, Federal Ministry of Health, Nigeria, with technical assistance from the Global HIV/AIDS Initiative Nigeria (GHAIN) project, Family Health International, and the US Centers for Disease Control and Prevention, Nigeria. The opinions expressed in this article are those of authors, and not the institutions with which they are affiliated or which provided funding. The authors thank the survey Technical Committee, staff from the Strategic Information branch of the HIV/AIDS division, Federal Ministry of Health and State Ministry of Health officials from participating states, site supervisors, interviewers, counsellors, laboratory technicians and other field workers. Sara Hersey, Mukhtar Mohammed, Kyle Bond, Ifeyanyi Ora, Inoussa Kabore, Kale Feyisetan, Samson Bamidele and Serge Xueref provided valuable input into the design and implementation of the study. Special thanks are given to Alliance Rights Nigeria and the respondents.


Supplementary materials


  • Funding United States Agency for International Development and US Center for Disease Control and Prevention.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the Protection of Human Subjects Committee (PHSC) of Family Health International North Carolina, the US Centers for Disease Prevention and Control and the National Institute for Medical Research (NIMR), Nigeria.

  • Provenance and peer review Not commissioned; externally peer reviewed.