The HIV epidemic in China has shifted from transmission driven by injection-drug-use practices to sexual transmission. Among the 50 000 estimated new HIV infections during 2007 heterosexual or homosexual transmissions were estimated to account for 44.7% and 12.2%, respectively.1 Several surveys conducted in recent years have shown a high prevalence of syphilis infection among men who have sex with men (MSM) in China.2^–4 A special concern has been expressed about potentially increasing transmission of HIV and outbreaks of syphilis among MSM.4 5

The National STD Surveillance System was established in 1987 to include a nationwide STD case-notification system and prevalence surveys in a limited number of sentinel sites.6 In 1995, 42 national HIV sentinel surveillance sites were established in 23 provinces, covering STD clinic attendees, female sex workers, drug users and long-distance truck drivers.7 MSM were not included in the HIV sentinel surveillance programme until 2002. Seroprevalence surveys of HIV and syphilis among MSM have been gradually introduced into sentinel sites, increasing from one in 2002 to four in 2006.7 However, the settings in which MSM are recruited usually vary within surveys2 8 or between surveys across years and sites.9 10 In order to gauge the magnitude of such differences we measured the prevalence of HIV and syphilis infections among MSM from different settings in Shenzhen, Guangdong Province, China.

METHODS

The study population from three surveillance surveys consisted of 1146 MSM living in Shenzhen, including 280 recruited from three saunas, 123 from three bars and 743 from the Shenzhen Institute of Dermatology STD clinic, where special health services are provided to MSM. Of 20 to 30 such settings in Shenzhen, the selected bars and saunas were the main venues for MSM to gather. After informed consent was obtained, participants were interviewed using a short questionnaire. Blood specimens were collected in a separate room on the sites and transferred back to the laboratory at the STD clinic for anonymous testing for syphilis (TRUST, Rongsheng Biotech, Shanghai, China for screening and TPPA, Fujirebio, Tokyo, Japan for confirmation) and HIV (ELISA for screening and western blot for confirmation). Specimens were considered positive for syphilis infection when both TRUST and TPPA were positive; and positive for HIV infection when both ELISA and western blot results were positive. A confidential contact number was used to notify subjects of the results of their anonymous HIV and syphilis tests. All those infected with syphilis or HIV were provided with HIV and STD counselling at the local STD clinic. Statistical analysis was performed using SPSS software (V. 13. 0, SPSS Inc, Chicago, Illinois, USA). p Values were calculated using analysis of variance (ANOVA) for continuous variables and χ² testing for categorical variables.

RESULTS

There were no significant changes in sociodemographic characteristics (eg, mean age, education level) of the study population during the three rounds of surveys, but more MSM reported homosexual behaviours alone and fewer reported bisexual behaviours in subsequent years.

A total of 1146 MSM were included for analysis (table 1). The mean (SD) age was 32.2 (SD 7.9) years, 22.5 (3.9) years and 28.3 (6.5) years for men from saunas, bars and the clinic, respectively (p<0.001). MSM recruited from saunas (39.6%) were more likely to have been married than those from bars (4.1%) or the clinic (18.9%) (p<0.001). A higher proportion of MSM from bars (70.2%) and clinics (68.8%) reported bisexual orientation compared with saunas (44.4%) (p<0.002). More MSM recruited from the clinic (52.8%) admitted to both insertive and receptive anal sex compared with those from saunas (41.2%) or bars (39.5%) (p = 0.002).

The overall prevalence of syphilis and HIV infection also varied by setting. Syphilis positivity was higher among the men recruited from saunas (20.7%) and the clinic (24.3%) compared with men recruited from bars (2.4%) (p<0.001). HIV prevalence was 3.6% at saunas, 4.2% at the clinic and 0.8% at bars (p = 0.09).

DISCUSSION

China has an expanding syphilis epidemic as shown by the increased incidence in national surveillance programmes.11 MSM groups in some studies have the highest prevalence in China.12 During 2000 to 2005 the median syphilis prevalence (interquartile range) for MSM was 14.56% (10.61% to 18.7%).12 A substantial prevalence of syphilis was found among MSM in our study, while HIV prevalence appears to be comparable to that reported in MSM communities in Beijing,2 13 but higher than that in the neighbouring city and provincial capital of Guangzhou8 and many other parts of China.9 10 The relatively high prevalence of HIV and syphilis further emphasises the importance of urgently implementing comprehensive interventions among this high-risk population. However, our study suggests the prevalence varies significantly by setting, showing a higher prevalence among MSM recruited from saunas or clinics than those from bars. For those MSM attending STD clinics the nature of sexual behaviours among this population may be different from those in bars or saunas and some of them are often referred by outreach service providers or intervention programmes to clinics for testing to confirm or exclude infections.

Risk behaviours and sociodemographic characteristics may also be associated with the different settings. For example, unmarried single MSM may be likely to go to bars to find new partners, while those with a family prefer to go to saunas for homosexual activities. The significant difference in terms of demographic characteristics and positivity for HIV and syphilis infections between MSM recruited from bars and those from saunas and clinics warrants consideration of what kind of population we need to include to represent the diversity of the population and/or where we should recruit to reach those at high risk of HIV and syphilis. Surveillance activities among MSM populations may also be hindered because of the stigma associated with homosexuality in traditional Chinese culture, increasing self-reporting bias among this population. Interpretation of epidemiological and behavioural data emerging from MSM in diverse settings, with a potentially high selection bias, is challenging.

In China communities of MSM are relatively hidden and hard to reach because of the stigmatisation of homosexuality. Surveillance activities of MSM therefore face many limitations and challenges. Participants of surveys are usually recruited by convenience sampling from gay venues or over the internet, or increasingly by respondent-driven sampling. The representativeness of these surveys with these sampling strategies is uncertain and results in difficulties in comparing data from location to location or from study to study. Nonetheless, maintaining consistency over time in serial cross-sectional studies is necessary for examining trends in HIV and other STD epidemics among MSM. China is poised to see many one-time surveys of HIV and syphilis among MSM around the country. These surveys could serve as a baseline for evaluating the effectiveness of current and intervention activities among MSM and as a useful tool for further planning surveillance activities. However, as such surveys scale up, we recommend careful consideration of the selection of the appropriate settings to represent the MSM population and to compare trends over time and between cities.