High risks of HIV transmission for men sex worker — a comparison of profile and risk factors of HIV infection between MSM and MSW in China

Men who have sex with men (MSM), identified as a key population, continue to account for 23% of new HIV infections worldwide [1]. In China, the biggest increase in HIV transmission has occurred among MSM, growing from 2.5% of new reported cases in 2006 to 26% in 2014 [2]. HIV prevalence among MSM in Tianjin, the third biggest Chinese city, has hit 12% [3].

Being deeply influenced by Confucianism, the majority of population in China refers to the behavior on having homosexual sex as a violation of moral. To avoid the stigma and discrimination, MSM often marry to women, although they still hold the “sex activities” with other men [4]. Marriage and risk behaviors were frequently mentioned in qualitative research. A study in China found that majority of MSM get married due to pressures like family line [5]. They conceal their sexual orientation [6], and express the willingness to remain contact with their sex partners [4]. This may build a dangerous “bridge” where HIV propagate from MSM to general population.

By contrast, another vulnerable group are MSWs. They sell or exchange sex for money or goods and encompass a very diverse population across and within countries [7]. In Netherlands, an observation study based on clinics indicated 40% of consultations were positive for sexually transmitted diseases (including HIV) within MSWs, while a meta-analysis conducted in 2014 suggested HIV prevalence among MSWs was 15.6% [8, 9]. Besides, a research conducted in Kenya showed that MSWs had the higher risk of HIV infection than MSM (26.3% vs. 12.2%) [10]. The high HIV prevalence within MSWs suggests that they may face unique barriers to accessing healthcare [7], which involved in stigma related to sexuality, gender identity, HIV status, sex worker status and internalized stigma, not only from general population but also MSM [11]. The indicatives of stigma impede them from accessing HIV-related healthcare. Beyond that, there are other biological and behavioral factors, as some qualitative studies suggests, giving rise to the high prevalence within MSWs like risky sexual behaviors and marriage [12]. However, these evidences have not been well-established in China.

Given that, the purposes of our study were to depict the prevalence of HIV infection among MSM and MSWs and then profile the differences in risk factors. Results from this study can be used to develop empirical public health strategies for reducing HIV transmission between MSWs and MSM.

From 2011 to 2018, 235 MSWs and 4931 MSM were included into our study and 363 bathers were identified as HIV positive. Of MSWs, the prevalence of HIV infection was 17.9% (42/235, 95%CI:13.2% ~ 23.4%). By contrast, the prevalence was 6.5% in MSM (320/4931, 95%CI:5.8% ~ 7.2%).

As depicted by Fig. 1, during the 8 years, the nonlinear spline showed a decreasing trend in prevalence in MSM. However, a “down-up” fluctuation could be observed in MSWs in terms of prevalence of acquired HIV. It is worth mentioning that the prevalence of HIV infection in MSWs was always higher than that in MSM’s. Until 2018, the prevalence of HIV infection has fell below 3% while corresponding number continued to be higher than 10% in MSWs.

Compared with MSM, there were significant differences in some characteristics such as marital status, highest education, districts, sites for seeing partners, HIV awareness, unprotected sex behaviors, having sex with women, drug use, uptake of health service and test last year (P < 0.05), as presented by Table 1.

Engaging in male sex industry was associated with increased risk of HIV infection in multivariable analysis with aOR (3.38, 95%CI:2.14 ~ 5.32). In detail, in multivariable analysis using variable selection, we adjusted those variables, which were accepted as risk factors, including marital status, age, highest education, self-reported chance of HIV acquired, and having sex with women.

Univariate analysis showed that marital status(P < 0.0001), highest education (P < 0.0001), anal sex(P < 0.0001), commercial sex(P < 0.0001), group sex(P < 0.0001) and the frequencies of condom use (P < 0.0001, respectively), uptake of HIV-related service (P = 0.0067) and HIV test last year (P = 0.0041) were significantly associated with HIV infection. Thus, we included those variables into multiple logistic models and then derived aORs (95% CI) of factors in MSM and MSWs, respectively. Tables 2 and 3 showed the different association of some variables on HIV infection in the two groups. As suggested by multiple variables analysis using stepwise selection, there exist some variables with aORs significantly bigger than 1 in MSM while the counterpart in MSWs were not such as age, marital status, the highest education, chance of HIV acquired, having a HIV test last year and so on.

Compared with MSM, MSWs tend to be younger, have lower education and most of them were not locals from Tianjin. These characteristics reflected some noteworthy profiles including the low-income level of the MSW population, sex services for goods, and high mobility, which have become obstacles to their accessibility and intervention [13].

In our study, within MSM, the prevalence of HIV infection was 6.5% while 17.9% among MSWs, which was similar to the research in Shanghai [14] and Shenzhen [15]. Undoubtedly, MSWs have become a non-negligible group for HIV infection in China.

From 2011 to 2018, the comparison between the downward trend of HIV infection among MSM, which has parallel results with Cui [16] and the fluctuation among MSW population indicated, to a certain extent, that the strategy for MSM is not applicable to MSWs. The reasons for this phenomenon might be as follows: firstly, due to the health promotion and education from the station set up in the bathhouse, MSM has increased their awareness of protection and reduced dangerous sexual behaviors; secondly, MSWs tend to be younger, have lower education in comparison with MSM [17] and they usually sell sex for money and goods, so changes in cognition may not be necessarily reflected in actions; thirdly, during sex with others, male sex workers are usually passive, as they need to meet the needs from customers.

The test rate among MSWs was significantly higher than MSM. In terms of sexual behavior, the frequencies of using condoms among MSWs was not lower than MSM in high-risk sexual behaviors (homosexual anal intercourse, group sex and commercial sex). Nevertheless, the prevalence of HIV infection for MSWs was still more than twice that of MSM. There may be some reasons for this. Firstly, there might not be use of condom in the entire process of sexual behavior. Due to the professional requirements of MSWs, they would try their best to meet their customers’ sexual desires. Therefore, though MSWs have a better grasp of HIV knowledge than MSM, unprotected sexual behavior still occurs; secondly, the frequency of drug use in MSWs was significantly higher than MSM. According to related research [18], drugs are positively correlated with HIV infection. As a psychological booster, it could obscure the consciousness of MSWs and clients, causing them to have no awareness of condom use during sex with their customers. This undoubtedly greatly increases the risk of HIV infection in MSWs. Finally, the information we collected about condom use was self-reported by MSM and MSWs, which may be overestimated due to social-expectation bias.

The differences in characteristics may lead to differences in association between risk factors and HIV infections. As suggested by Table 2 and Table 3, the associations were different in MSWs and MSM. The UWD group has higher risk than MC, which is in line with a previous study in China. While that association was observed in MSM, it was not in MSWs. This may be due to the fact that most of MSWs were youngers, and we cannot get enough examples aged > 30 just like in MSM. What’s more, due to characteristic of MSWs, even though some got married, they often don’t stop dangerous sex behaviors. Similarly, no matter what the highest education was, they must continue to meet the needs from clients to exchange money and goods. Therefore, although they were likely to have more frequencies and more uptake of HIV-related service than MSM, similar results could not be found.

Our study had several strengths. Firstly, to our knowledge, this bath-based study was the one of the largest observational studies collecting more than 5000 MSM in China. Moreover, we focus on MSWs, the understudied population in MSM in China. The present study disclosed a remarkably higher risk of HIV infection among MSWs, indicating urgent measures should be considered on the vulnerable population in grey areas. Secondly, diversity of MSM was recruited in the present study. The age of patrons coming into the bathhouse ranges from 18 to 80 years old, the hukou of participants covers most regions of China and the variety of ways of recruit (bars, clubhouses, peer referral and internet) are adopted. Feasibility, authenticity, and diversity can be guaranteed. Finally, our study explored the discrepancy in demographic and sex behavioral features between MSM and MSWs, which could provide valuable information for implementation of targeted HIV/AIDS health and education interventions among MSWs. Our study also had some limitations. A snowball sampling was used for accessing MSM or MSWs, which can cause some selection bias. But in the context of the Tianjin Bathhouse, we could obtain all kinds of information from patrons around China. The data of this study was obtained through interview questionnaires and was inevitably subject to information bias due to particularities of MSWs. However, our investigators themselves are also MSM, along with that survey was conducted on a one-to-one basis using words and expressions from their cultural circles. Thus, the bias has been minimized. Although the questionnaire contains information involved in demographics, behaviors and services, there were still some issues not mentioned such as their posture during sex, which needs to be studied further.

In conclusion, there were some differences in characteristics between MSWs and MSM and the amounts of effects of risk factors on HIV infection were also different. These disagreements should be considered in design of strategies targeting to MSWs and MSM, respectively.