Abstract
Summary
This study investigated HIV, hepatitis B (HBV) and C (HCV) and syphilis prevalence and associated behavioural correlates to HIV infection among gay, transsexuals and men who have sex with men (GTMSM) in the Dominican Republic using respondent-driven sampling. Eligible men were aged ≥15 years, reported engaging in anal/oral sex with another man within the past six months and living in one of the four cities sampled. GTMSM were recruited in Santo Domingo (n = 510), Barahona (n = 281), La Altagracia (n = 270) and Santiago (n = 327). HIV seroprevalence ranged from 5.1% to 7.6%. HBV ranged from 0.6% to 3.5%, HCV from 0.8% to 6.9% and syphilis from 3.5% to 6.9%. GTM practice numerous risky sexual behaviours including having multiple sex partners and practising inconsistent condom use with male and female, including transactional, sex partners. Although we found lower HIV prevalence compared with previous studies, the high level of extremely risky sexual behaviours practised by GTM in the Dominican Republic will most likely result in increases in HIV and other infections if action is not taken to scale-up effective prevention programmes for this population.
Keywords
INTRODUCTION
The Dominican Republic comprises two-thirds of the Caribbean island of Hispaniola and has a population of roughly 9,500,000. 1 National HIV prevalence estimates among Dominican adults (15–49 years old) fall somewhere between 0.8 and 1.2% and there is evidence suggesting declines in HIV prevalence from the mid-1990s to 2002 among some key sentinel surveillance groups (i.e. persons attending sexually transmitted infection [STI] and antenatal clinics, donating blood, applying for US entry visas and working in commercial sex).2,3 Although HIV in the Dominican Republic has largely been considered to be spread primarily through heterosexual activities, 2 recent analysis of existing data suggests that HIV in this country is most likely characterized by male-to-male sexual behaviours coupled with additional heterosexual activities.4,5
Little has been documented about gay, transsexuals and men who have sex with men (GTMSM) in the Dominican Republic; most likely a result of the high level of denial of, or stigmatization towards, homosexual behaviour in the region. Some data suggest that between 8% and 28% of Dominican men have ever had sex with men. 6 Based on commonly accepted worldwide estimates that 3–10% of adult men (≥15 years old) have had sex with other men, approximately 100,000–330,000 Dominican men are estimated to be having sex with other men. 7 Sex between men is estimated to account for one in eight (12%) reported HIV infections in the region. 2 However, given the high level of stigma in the Dominican Republic, men who become infected with HIV through homosexual contact may not disclose this and therefore be classified as heterosexual.6,8–10 Three studies using different convenience sampling methods found HIV prevalence among MSM to range from 11% to 19% and even higher for those self-identifying as either homosexual (36%) or bisexual (28%).10–12 There are currently no existing data measuring other infections among MSM, including hepatitis and syphilis in the Dominican Republic.
In 2008, The Presidential Council on AIDS (COPRESIDA) conducted a biological and behavioural survey using respondent-driven sampling (RDS) among GTMSM in four cities: Santo Domingo, Santiago, Barahona and La Altagracia. The purpose of the survey was to measure HIV, hepatitis B (HBV), hepatitis C (HCV) and syphilis prevalence and associated HIV risk behaviours among GTMSM in different cities across the Dominican Republic. The four target cities were selected based on a wide spectrum of important economic, geographic and sociocultural differences: Santo Domingo, the governmental and economic centre, is the capital, largest and most important city in the country; Santiago, considered the second most important city in the Dominican Republic, is located in the northwest and is the intellectual, educational and cultural centre as well as a major industrial centre with rum, textile, cigarette and cigar industries; Barahona is located in the south-west, is somewhat rural and an area of sugar production; and La Altagracia includes several coastal areas which are important tourist destinations.
The survey used RDS – an adaptive sampling method that utilizes social network sizes and recruitment patterns to determine probabilities of selection. RDS is a stochastic process in which samples comprised of long recruitment chains can be analytically adjusted to represent the network of the population sampled.13,14 This paper describes key sociodemographic, behavioural risk and disease prevalence and explore correlates of HIV serostatus of GTMSM in the Dominican Republic.
METHODS
Eligibility, sampling and survey sites
RDS was selected as a sampling method for sampling GTMSM based on formative investigation identifying sufficiently large social networks needed to sustain peer-to-peer recruitment. The target sample size for all cities, with the exception of Santo Domingo, was 300. For Santo Domingo, the sample size was increased to 500 due to its importance, size and the participation of organizations actively working with GTMSM. Eligible participants were men who had anal or oral sex with another man (including self-identified transsexuals, gays and homosexuals) in the previous six months, 15 years or older and living in the respective city in which the survey was conducted. Sampling began with initial eligible participants referred to as ‘seeds’ identified through key informants or local organizations working with GTMSM in each city. Seeds were diverse with respect to age, GTMSM self-identification, marital status and sexual risk behaviours.
Sociodemographic and HIV behavioural risk variables among gay, transsexuals and men who have sex with men (GTMSM) in four cities in the Dominican Republic 2008
RDSAT adjusted estimations, 95% confidence intervals (CI)
*Sample sizes include seeds
Odds ratios (OR) with 95% confidence intervals (CI) for factors associated with HIV among gay, transexuals and men who have sex with men (GTMSM) in four cities in the Dominican Republic, 2008 (n = 1386)
†Weighted by degree and recruitment weights (Heckathorn 2007)
‡Controlling for city differences and number of female sex partners in the past six months
*P < 0.05; **P < 0.01; ***P < 0.001
Biological measures
Blood specimens were processed at the National HIV laboratory for HIV antibody testing (Abbot, Determine® VIH-1/2). Reactive specimens were confirmed using enzyme-linked immunosorbent assay and discordant results were tested with Ortho® HIV-1/HIV-2 (Johnson & Johnson). H BsAg surface antigen antibody was detected with HBsAg BIO-CARD 50T and HCV was detected with Bioblot HCV. Syphilis infection was tested using a rapid plasma regain (RPR) to detect Treponema pallidum antibodies without serological confirmation.
Behavioural measures
The questionnaire included demographic characteristics (e.g. age, education, marital status); sexual orientation; sexual behaviours (e.g. number and types of partners and condom use) with men and women, drug use behaviour, testing history and past STI symptoms. Questions about anal receptive and insertive intercourse were analysed to measure the prevalence and condom use of the more riskier receptive intercourse vis-a-vis the less riskier insertive intercourse. In addition, the question about wanting to know the HIV status of partners was analysed to measure the extent to which GTMSM are discussing HIV with sexual partners. Finally, the question about the number of men known to the participant who are infected with HIV was analysed to provide information about the extent to which HIV is known and exists in the GTMSM community (Table 1).
Analysis
Estimates and 95% confidence intervals (CIs) for each city were calculated using the RDS Analysis Tool 6.0 (RDSAT) (
Logistic regression models were constructed for aggregated cities to evaluate factors associated with HIV infection. The main exposure variables were GTMSM self-identity, types and numbers of sex partners, condom use among sex partners and disease prevalence. Sample weights for the regression models were generated in RDSAT using the social network sizes for HIV infection. Data were then exported to STATA version 9.0 bi- and multivariate analyses. Variables associated with HIV infection at P < 0.2 in the bivariate analysis were included in the initial multivariate model. Those variables that remained associated with HIV infection or GTMSM self-identity at a P < 0.05 significance level or were considered important based on the current literature or research were retained in the model, while controlling for city differences and other confounding variables. Significant adjusted odds ratios (AOR) and 95% CI were calculated and presented in the final model.
RESULTS
Patterns of recruitment
Recruitment began with seven seeds in Santo Domingo, six in Santiago, eight in Barahona and seven in La Altagracia. The total sample comprised 1387 participants from Santo Domingo (n = 510), Barahona (n = 281), La Altagracia (n = 270) and Santiago (n = 327). Recruitment lasted from May to June 2008. Recruitment chains in each city reached at least 11 waves (15 waves in Santo Domingo, 12 in Barahona, 11 in La Altagracia and 13 in Santiago). Mean design effect on HIV and condom use at last sex was 1.75 (less than 2 used for the sample size calculation in these studies), indicating reasonable sample sizes for the variance expected. Almost all GTMSM (mean: 92%) reported having reciprocal relations with the person who recruited them.
Sociodemographic
Median age for each city was between 22 and 24 years (maximum ages were between 53 and 61 years). Most men in Santo Domingo and Santiago had ≥grade 8 education and were never married to a woman whereas those from Barahona and La Altagracia had <grade 8 education and had been married to a woman. Between 12–26.4% were currently married and half or more in Santo Domingo, La Altagracia and Santiago were single. The majority of GTMSM in all cities reported their sexual orientation as bisexual (39–63%). The next most common sexual orientation reported by GTMSM in Santo Domingo, La Altagracia and Santiago was gay and in Barahona was MSM.
Sexual risk
The majority of GTMSM in all cities reported that their same sex partner type during last sexual intercourse was a commercial partner (37.9–57.3%). Few (2.7–6.2%) GTMSM reported having their last sexual intercourse with a stable partner with whom they were living. Most GTMSM reported having insertive anal sex during the last six months (78.6–91.8%) and among those, the majority reported ≥2 partners during that same period (40.7–69.4%). Between 16.6% and 44.2% of GTMSM reported having receptive anal sex during the past six months, and among those, 11.1–25.7% reported ≥2 partners during that same period. With the exception of Barahona (36.1%), most GTMSM reported using condoms at last insertive anal sex with a non-paying partner (68.7–76.6%). Most GTMSM in Santo Domingo (65.6%) and Santiago (55.1%) reported using condoms at last receptive anal sex with a non-paying partner, whereas only 19.4% of GTMSM in Barahona and 27% in La Altagracia reported doing so. GTMSM in Barahona, compared with those in the other three cities, had a significantly lower proportion reporting consistent condom use during anal sex with an occasional male partner (15.1%, P = 0.001 versus 45.9–56% in the other three cities) in the past six months. Consistent condom use during anal sex with a permanent partner in the past six months was low for all cities (<38%), especially in Barahona (17.4%).
High percentages of GTMSM reported ever having transactional sex with a man (between 61.7% and 83.2%) and, with the exception of men in Santo Domingo (54%, P = 0.009), more than 70% of men received payment for sex from a man in the past week. Among those, between 29.2% (Santo Domingo) and 50.6% (Barahona) reported doing ≥2 times. With the exception of Barahona (39.3%, P = 0.001), 70% or more of GTMSM reported consistent condom use with a man during last transactional anal sex.
Most GTMSM in all cities reported having oral, vaginal or anal sex with a non-paid female partner in the past six months. This proportion was significantly lower among GTMSM in Santo Domingo (68%, P = 0.013 versus 83.2–88.5% in the other three cities); among those, between 58.2% and 76.8% reported doing so with ≥2 female partners. Between 35.5% (Barahona) and 62.4% (Santiago) of GTMSM reported consistent condom use with their non-paid female partners during last sex.
Drug use
Past use of non-injection drugs was highest in Barahona (74.6%, P = 0.015) and La Altagracia (55.9%, P = 0.021) and 29.9% in Santo Domingo and 36.5% in Santiago. Among those, ≥60% had used drugs in the past three months. Few GTMSM (0–4.4%) reported injecting drugs in the past one year.
HIV topics, testing history and past STI symptoms
Most GTM in all cities reported that they wanted to know if their anal sex partners are infected with HIV (≥72%). Between 21.4% and 32.7% GTM know two or more men who have HIV. Between 33% and 50.9% of GTMSM reported ever having an HIV test and few reported having symptoms of an STI in the past year (5.5–9.4%).
Prevalence of HIV and other infections
All participants agreed to provide venous blood for testing. There were no significant differences for HIV and other infections among GTMSM in the four cities studied. HIV seroprevalence ranged from 5.1% in Santiago to 7.6% in La Altagracia. Syphilis (RPR tested only) ranged from 3.5% in La Altagracia to 6.9% in Barahona. HBV ranged from 0.6% in Barahona to 3.5% in La Altagracia and HCV ranged from 0.8% in Santo Domingo to 6.9% in Barahona.
Risk factors and co-infection for HIV among all cities
In the bivariate regression analysis, HIV seropositive GTMSM were significantly more likely than seronegative GTMSM to be older, self-identify as gay or transsexual, report receptive anal sex, report ≥2 receptive anal sex partners in the past six months, report using a condom during last receptive anal sex with a non-paying male partner and be infected with syphilis or HBV. HIV seropositive GTMSM were statistically significantly less likely to report insertive anal sex, 1 or ≥2 insertive anal sex partners in the past six months, sex (oral, vaginal or anal) with a female and ≥2 female sex partners in the past six months.
After controlling for city differences and number of female sex partners in the past six months, HIV seropositive GTMSM were statistically significantly more likely to be older, self-identified as gay or transsexual, report ≥2 receptive anal sex partners in the past six months, and test positive for syphilis or HBV.
DISCUSSION
HIV prevalence of GTMSM in each city was above 5% and across the four cities the mean prevalence was 6.1%, indicating potential widespread transmission of HIV in this population. Indeed, a recent study suggests that GTMSM accounted for 33% of new infections in 2010. 5 Mean syphilis prevalence across the four cities was 5.4% and GTMSM who were HIV seropositive were more likely to be infected with syphilis. Co-infection with HIV and other STIs could indicate faster progression to disease as well as increased risk of further HIV spread within this community and to the general population. 17 This survey did not use confirmatory testing for syphilis, possibly resulting in false-positive results. Overall HBV prevalence was low (0.6–3.5%), but our findings only indicate active or chronic infection and does not reflect lifetime prevalence of HBV in this population. HBV is 50–100 times more infectious than HIV 18 and is easily transmitted during sexual activity indicating numerous opportunities for further spread in this population. HIV increases the risk of cirrhosis and end-stage liver disease in HBV co-infection; 19 however, data have not found HBV co-infection to have a substantial impact on the development of AIDS related illness or death.20,21
Although not significantly associated with HIV seropositivity, HCV was 0.8–5.0%. Considering that HCV is transmitted through contact with blood, injection drug use is often considered the most likely method of acquiring this disease. However, this study found low percentages of GTMSM (0–4.4%) reported ever injecting drugs and no information was gathered about ever sharing injection drugs. In general, injection drug use is considered to be rare in the Dominican Republic and data from a similar survey of drug users using the same sampling methodology conducted in the same cities in 2008 found that between 0–4.9% (mean of 2.9% across all four cities) of drug users reported ever injecting drugs, among which few reported sharing needles. 3 Not until recently has male-to-male sexual contact (especially rougher or poorly lubricated, unprotected anal penetration and sex between persons with genital ulcers) been reported as an efficient transmission route for HCV, especially among men who have sex with men.22,23 More research is needed to elucidate factors associated with HCV in this population.
GTMSM in the Dominican Republic practice risky sexual behaviours including having multiple sex partners and practicing inconsistent condom use with occasional, permanent and transactional sex partners. It is not surprising that many GTMSM are not using condoms consistently given that the majority do not know any men who are infected with HIV. However, GTMSM appear to have some concern about possible HIV infection as indicated by the high percentage who are interested in knowing whether their anal sexual partners are infected with the virus.
High percentages of GTMSM reported being bisexual (39–63%) and having unprotected sex with women. Bisexual behaviours and non-disclosure of male-to-male sexual preferences to female partners have been reported in several studies of GTMSM in Latin America24–28 and bisexuality is considered an important component influencing the spread of HIV in the Dominican Republic.4,9,10
The finding of marked geographical variation in sexual behaviours and HIV, hepatitis and syphilis prevalence underscores the value of understanding local contexts in the prevention of HIV infection. Barahona GTMSM were found to have especially high-risk behaviours and disease prevalence when compared with the other three cities, including the highest prevalence of syphilis and HCV and lowest usage of condoms with female paid and non-paid sex partners and male transactional, occasional and regular sex partners. These findings suggest that targeted programmes and interventions focusing on access and correct usage of condoms (and lubricants) are urgently needed in Barahona. In addition, the scale-up of programmes targeting transsexuals and gay men in the Dominican Republic are needed given that these subpopulations were more likely to be HIV seropositive. There is ample evidence that transsexuals are a unique HIV risk group more prone to higher risk sexual behaviours, drug use and distinct physical, mental, social and economic disparities compared with other self-identified MSM groups (29–32). Unfortunately, this survey was only able to capture a small proportion of transsexuals. Including transsexuals as a subpopulation within MSM surveys may hinder the accurate measurement of the specific characteristics, behaviours and prevention needs of transsexuals. The highest percentage of transsexuals participating in the surveys was in the two largest cities, which most likely provide better possibilities for networking and access to social support and health services.
The findings in this study are subject to several limitations. Peer to peer recruitment may have resulted in some subpopulations being over-represented (e.g. men who engage in selling sex) or under-represented (i.e. transsexuals). Indeed, a large number of male sex workers (54% of Santo Domingo participants and 72–84% of participants in other cities) enrolled in the survey. RDS relies on a number of assumptions, some of which are difficult to meet or measure. One assumption is that the social network being sampled comprises one complete network component which results in each of the four survey cities being a separate network component (14). The aggregation of data from all four cities for the regression analysis violates this assumption. Therefore, conclusions drawn from the regression analyses should not be interpreted as being representative of GTMSM in the Dominican Republic and P values may overstate the significance of the model terms. The quality of estimates derived from RDS data has been challenged in several recent papers29–35 and more research is needed to determine how generalizable estimates are to the sampled population. Finally, The multivariate analysis that was conducted for this analysis utilized exported weights for the dependent variable (that is, HIV infection); however, standardized guidelines for multivariate analysis for RDS data are still under development and require validation. Despite these limitations, RDS was successful in gathering relatively large samples of a poorly understood and largely understudied population in the Dominican Republic.
We found lower HIV prevalence among GTMSM compared with studies conducted in the late 1980s and 1990s (HIV prevalence of 19% and 11%, respectively)10,11 and, more recently, in 2004 (HIV prevalence of 11%) 12 which used convenience sampling methods. Additional surveys using more robust sampling methods, such as RDS, as well as surveys using the same sampling method over numerous years are needed to validate these findings and to measure trends over time. A follow-up survey using RDS will occur in late summer 2012 and will be an important step in determining the direction of the HIV epidemic and measuring whether current interventions have been successful in reducing risk among this population.