According to estimates by the Chinese Ministry of Health, at the end of 2005, 650,000 individuals in China had been infected with HIV, which is more than 4 times the cumulative number of confirmed cases (n = 141,000).1 1 2 2-5 1
The spread of HIV among men who have sex with men (MSM) in China has received relatively little attention. In 2001, a survey of Beijing MSM found an HIV prevalence of 3.1%.6 7,8 9,10
Research shows that characteristics of “sexual networks” (sets of people with whom individuals have sexual relationships), such as age mixing (partnerships between members of different age groups) and concurrency (partnerships overlapping in time), are significant risk factors for HIV and other sexually transmitted infections (STIs).11-19 14,15 Chlamydia with discordance in race, age, and education.12 Chlamydia .17
“Social networks” (sets of people with whom individuals have social contacts) also can influence HIV/STI transmission because they can facilitate the diffusion of prevention information and because network members can influence safer sex norms.20 21 22
Despite their potential influence on HIV infection, social and sexual networks among MSM in China rarely have been investigated. Moreover, most prior studies have examined the effects of social or sexual networks on HIV risk but not both. Also, little is known about the extent to which networks overlap and whether the overlap affects the risk for infection. We conducted a cross-sectional study to characterize social and sexual networks and to examine their role in the spread of HIV and syphilis among MSM in Shanghai, which, with a population of approximately 17 million, is China's largest city.
METHODS
Participants and Procedures
Study participants were recruited from September 2004 to June 2005 using a snowball sampling methodology. Initial “seed” respondents were recruited through outreach activities at MSM venues, such as bars, dance clubs, and restaurants. Fifty-two seeds met the eligibility criteria: male gender, at least 18 years of age, Shanghai residence, and self-reported sex with men.
Structured 30-minute questionnaires were administered at 1 Shanghai municipal and 5 district Centers for Disease Control and Prevention (CDCs) by 6 male and 2 female health care workers who had been trained as interviewers. Serologic specimens for HIV and syphilis testing were collected after the interviewer provided pretest counseling. After these procedures, participants received HIV and STI prevention literature, 50 Chinese yuan (US $6.50) as compensation, and appointment cards with prenumbered stickers denoting their study identification number. If the results of HIV and syphilis screening were negative, participants received results and posttest counseling on the day of the interview. If either test was positive, they were requested to return 3 days later, when they received posttest counseling and confirmation of results, were offered free treatment for syphilis at a designated clinic, and were referred to an HIV case management service program at the Shanghai municipal CDC.
On completion of the survey, each seed participant was invited to nominate other MSM for the study. He was given “recruitment coupons” to give to potential participants as an invitation to join the study. Coupons were valid for 4 weeks from the participant's interview date, and each was printed with the project name, study identification number, expiration date, and telephone number. For every coupon redeemed, the nominating participant received an additional 10 Chinese yuan (US $1.30). Enrolled participants (“wave 1 recruits”) who were nominated by seed respondents completed the identical study protocol and were invited to nominate additional participants (“wave 2 recruits”) in the same manner. This process continued with no restriction on the number of recruitment waves. Protocols were approved by the Committee for Human Research of the University of California, San Francisco, and by the Institutional Review Board of the Shanghai Municipal CDC.
Measures
The questionnaire collected sociodemographic information (eg, age, self-reported sexual orientation, possession of a Shanghai household registration card [hukou ] allowing residents to access public services) and number of sexual partners over a participant's lifetime and within the past 6 months. Respondents also were asked about venues they frequented to meet other MSM in the past 6 months.
Eight true or false statements were used to assess knowledge about risk of HIV transmission (eg, sexual intercourse without using a condom, sex with someone who looks clean and healthy). Knowledge scores were created by counting the number of correct responses. Perceived personal risk for HIV infection was measured by asking: “What do you think about your own risk for developing AIDS? Would you say your risk is none, low, somewhat high, or very high?”
Social networks were enumerated by asking respondents to name up to 5 adults with whom they had most often spent leisure time in the past 6 months (ie, contact networks), followed by naming up to 5 adults with whom respondents had most often discussed important matters in the past 6 months (ie, conversation networks). When asked about conversation networks, respondents indicated whether any of the members had already been named as members of their contact network. Demographic information was gathered on each network member, including relationship to and sexual history with the respondent as well as knowledge and supportiveness of the respondent's homosexuality. Also assessed was which members of the contact network and conversation network knew each other.
Sexual networks were enumerated separately for male and female sex partners. Respondents were first asked about the number of men with whom they had anal sex during the prior 6 months and then about up to 5 men with whom they had anal sex most recently. Questions were asked about each male partner, beginning with the most recent, including whether he was a steady partner (eg, lover, boyfriend) or a nonsteady partner (eg, one-night stand, casual pick-up); the first and last times (month and year) they had anal sex together; and, in the past 6 months, the number of times they had insertive and receptive anal intercourse and the number of times they used condoms during intercourse. The same procedure was used to ask about female sex partners after asking about the number of women with whom respondents had anal or vaginal sex in the past 6 months.
We calculated the size (number of members), density, and supportiveness of each respondent-reported network. Densities were calculated as the number of network members who knew each other divided by the total number of network dyads, excluding the respondent (range: 0-1). As an example, a conversation network with 4 members contains 6 dyads (ie, 1 and 2, 1 and 3, 1 and 4, 2 and 3, 2 and 4, 3 and 4); if network members from 3 of those 6 dyads know each other, the network density would equal 0.5 (ie, 3 ÷ 6). For network sizes equal to 0 or unity, density was set to 0. Network awareness/supportiveness was operationalized as a 4-category variable: (1) no such network, (2) network members are unaware of a respondent's homosexuality, (3) less supportive (ie, ≤50% of network members being supportive of a respondent's homosexuality), and (4) more supportive (ie, >50% of network members being supportive of a respondent's homosexuality). Using contact networks as an example, this variable allows for several important comparisons: those with versus without a contact network (categories 2-4 vs. category 1); among those with a contact network, having disclosed their homosexuality to at least 1 network member versus not (categories 3-4 vs. category 2); and among those who have disclosed, receiving more versus less network support (category 4 vs. category 3). Finally, we defined network overlap as 1 or more members belonging to more than 1 network, which we operationalized with 3 binary indicators: any overlap between at least 1 social network and the male sex partner network, any overlap between at least 1 social network and the female sex partner network, and any 3-way overlap between both social networks and at least 1 sexual network.
Sex partner concurrency was assessed with a binary indicator. Only the calendar month and year that each sexual partnership began and ended were recorded. Because the exact beginning and ending dates were unknown, we chose a conservative definition that ensured partnership concurrency: (1) 2 partnerships spanned an identical set of 2 or more calendar months or (2) at least 1 partnership spanned 3 or more calendar months, and another partnership began or ended during one of the intervening calendar months.
Serum samples were initially screened for HIV-1 antibodies by the Determine HIV-1/2 rapid HIV antibody test (Dainabot Company Ltd, Matsudo-shi, Chiba-ken, Japan), and samples with a reactive result were confirmed by immunoblot assay (HIV Blot 2.2; GENELABS Diagnostics, Singapore). Syphilis was detected by the toluidine red unheated serum test for nontreponemal antibody (Wantai Company Ltd, Beijing, China) and confirmed by a treponemal antibody test (Serodia Treponema pallidum particle agglutination test; Fujirebio, Tokyo, Japan).
Statistical Analysis
Descriptions of social and sexual networks treated members as the units of analysis. All other analyses considered respondents as the units of analysis. Risk factors for syphilis infection and unprotected intercourse were examined with logistic regression models, which treated respondents as nested within seed cases and applied an adjustment to the variance estimation described by Morel.23 P > 0.20) were removed by backward elimination.
RESULTS
We screened 535 men and identified 515 who met eligibility criteria, 477 (93%) of whom agreed to participate in the study. We enrolled 52 seeds, 159 wave 1 recruits, 126 wave 2 recruits, 84 wave 3 recruits, 39 wave 4 recruits, and 17 wave 5 recruits. The mean, median, minimum, and maximum numbers of participants recruited by the seeds were 15, 3, 1, and 213, respectively.
Respondent Characteristics
The mean age of participants was 28 years (range: 18-56 years; Table 1 ). Seventy-three percent had no college education. Seventy-eight percent were never married. Twenty-two percent had a Shanghai hukou . Most self-identified as homosexual (57%) or bisexual (38%). Participants were knowledgeable about HIV transmission risk (mean correct responses = 6.9, range: 0-8). Most believed that they were at no risk at all (38%) or low risk (52%) for HIV.
TABLE 1: Characteristics of the 477 Study Participants
TABLE 1: (Continued )Characteristics of the 477 Study Participants
The median numbers of lifetime male and female sex partners were 5 and 1, respectively. Twenty-four percent reported a lifetime history of STIs. The venues where participants most frequently met other MSM during the past 6 months were bars and clubs (70%), house parties (66%), parks or public toilets (64%), bathhouses (61%), Internet sites (59%), and Internet cafés or bars (48%). The percentages of respondents who reported concurrent sexual partnerships within their male, female, and total (male plus female) sexual networks equaled 33%, 3%, and 43%, respectively; 17% reported concurrent sexual partnerships with male and female partners. Fifty-seven percent reported unprotected insertive or receptive anal sex with men in the past 6 months. Sixty-three percent reported unprotected anal or vaginal intercourse with men or women, and 13% reported unprotected anal or vaginal sex with men and women in the past 6 months.
HIV and Syphilis Infection
Of the 475 participants who provided blood samples for the study, 7 (1.47%, 95% confidence interval [CI]: 0.59 to 3.01) tested positive for HIV, 64 (13.47%, 95% CI: 10.53 to 16.88) had markers for syphilis , and 2 (0.42%, 95% CI: 0.05 to 1.51) were coinfected with HIV and syphilis .
Characteristics of Social and Sexual Networks
The 477 respondents named a total of 1244 contact network members and 819 conversation network members. The mean sizes of contact and conversation networks were 2.61 and 1.71, respectively. Mean density was 0.49 for contact networks and 0.25 for conversation networks. Approximately one third of social network members previously had been a sexual partner of the respondent (Table 2 ).
TABLE 2: Characteristics of Members of Study Participants' Social Networks
The total numbers of individuals named for male and female sex partner networks were 1093 and 160, respectively. The mean size and mean density of male sex partner networks were 2.29 and 0.10, respectively. The corresponding figures for female sex partner networks were 0.34 and 0.01. Most male (52%) and female (74%) sex partners had unprotected sex with respondents during the past 6 months (Table 3 ).
TABLE 3: Characteristics of Members of Study Participants' Sexual Networks
Overlap Between Social and Sexual Networks
Fifty-one percent of respondents had at least 1 male sex partner who was also a member of their contact or conversation network. Nine percent reported that at least 1 female sex partner belonged to their contact or conversation network. Twenty-two percent reported that at least 1 sex partner was a member of both social networks.
Correlates of Syphilis Infection and Unprotected Sex
Table 4 reports results of bivariate and multivariate logistic regression analyses of factors associated with syphilis infection and having unprotected anal or vaginal intercourse in the prior 6 months. In the multivariate analysis, having 3 to 5 versus 6 or more lifetime male anal sex partners (P = 0.0361), meeting other MSM at the gym (P = 0.0337) or through the Internet (P = 0.0065), having a steady female sex partner (P = 0.0365), and having overlapping social and sexual networks (P = 0.0052) were negatively associated with syphilis infection. Also, contact network awareness/supportiveness was associated with syphilis infection (omnibus P = 0.0248). Essentially, the risk for infection was lower among respondents who reported having a contact network versus those who did not; all other pairwise comparisons between the 4 categories were nonsignificant. The association between conversation network awareness/supportiveness and syphilis infection was marginally significant (omnibus P = 0.0691), although respondents with at least some network members who were aware of their homosexuality (ie, “≤50% supportive” and “>50% supportive”) had a higher risk of infection than those with no network members aware of their homosexuality (P = 0.0251 and P = 0.0232, respectively). Again, all other pairwise comparisons between the categories were nonsignificant. These findings suggest that among those with a conversation network, disclosure versus nondisclosure to network members was the significant predictor of infection.
TABLE 4: Correlates of Syphilis Infection and Unprotected Anal or Vaginal Sex in the Past 6 Months: Results of Bivariate and Multivariate Logistic Regression Analyses of the 477 Study Participants
TABLE 4: (Continued )Correlates of Syphilis Infection and Unprotected Anal or Vaginal Sex in the Past 6 Months: Results of Bivariate and Multivariate Logistic Regression Analyses of the 477 Study Participants
Marital status (P = 0.0051), HIV knowledge (P = 0.0225), number of lifetime male anal sex partners (omnibus P = 0.0439), type of male (P = 0.0349) and female (P = 0.0309) sex partners, and size of male sex partner networks (P = 0.0003) were independent correlates of having unprotected anal or vaginal intercourse. Respondents who had been married, were more knowledgeable about HIV, had 6 or more versus 3 to 5 lifetime male anal sex partners, had a steady male or female sex partner, and had a greater number of members in their male sex partner network were more likely to have unprotected anal or vaginal intercourse.
DISCUSSION
Although we found a low prevalence of HIV (1.47%) among MSM in Shanghai, the infection rate for syphilis was high (13.47%). Sexual risk behaviors also were commonly reported. More than half of participants (57%) reported unprotected anal intercourse with men in the past 6 months, and one third reported having concurrent male sex partners. These findings are similar to those reported in other studies of MSM in China.6-10,24-26 syphilis , 8% for Chlamydia , 27.7% for nonchlamydial nongonococcal urethritis, 7.8% for herpes simplex virus type 2, and 2.7% for gonorrhea. Almost half reported unprotected anal intercourse with men in the past 3 months.26
Given that many MSM in China are married and have sex with women,27-29 syphilis infection and no association between having sex with men and women and syphilis infection or having unprotected anal or vaginal sex, we cannot discount the possibility that Chinese MSM may spread HIV to their female sex partners. One quarter of respondents reported having female sex partners in the past 6 months. During that period, 17% reported having male and female concurrent sex partners and 13% reported unprotected anal or vaginal sex with both genders.
There were high levels of overlap between social and sexual networks in our sample of MSM. More than half had a male member of their contact or conversation network as a sexual partner, and 22% reported at least 1 of their current male or female sex partners as a member of their contact and conversation networks. Moreover, social and sexual network overlap was negatively associated with syphilis infection. These data suggest that HIV prevention programs should target the social and sexual networks of Chinese MSM.
Our study investigated the influence of 4 different characteristics of sexual networks on syphilis infection among Shanghai MSM: size, density, age mixing, and concurrency. These characteristics, however, seemed to exert little influence on syphilis infection in this risk population. Our study found no relation between syphilis infection and size and density of male and female sex partner networks or, contrary to prior studies,11-15 syphilis infection within the Shanghai MSM community in our study. Sex partner concurrency also was not associated with syphilis infection. One can speculate that the prevalence of syphilis infection within their sexual networks was such that respondents' concurrency did not pose a substantial risk for them with respect to syphilis infection. Given the high rates of infection found in this sample of MSM, however, concurrency possibly could place their sex partners at risk for infection. As demonstrated in other studies,18,19
The size and density of social networks examined in this study had no influence on syphilis infection. Having a contact network lowered the risk for infection, however. It is possible that contact networks tended to inhibit homosexual activity because of respondent concerns about being “outed” to network members or reprisals from some network members, such as family members, neighbors, or employers, who may disapprove of homosexuality. By contrast, among those with a conversation network, those who had disclosed their homosexuality to at least 1 network member were at higher risk than those whose conversation network was completely unaware of their homosexuality. One explanation for this finding may be that some of the conversation network members to whom MSM had disclosed their homosexuality are sex partners or other MSM who have riskier social norms. The findings about the protective effect of contact networks and the adverse effect of conversation networks on syphilis infection suggest that syphilis prevention programs for Chinese MSM might benefit from 2 different approaches targeting social networks: one aimed at forming a contact network and the other aimed at promoting safer sex among conversation network members.
We found no associations between unprotected anal or vaginal sex with men or women in the past 6 months and syphilis infection. These findings may be attributable to how we measured unprotected sex and syphilis infection. Whereas we assessed recent sexual risk behavior, we did not assess recent syphilis infections, such as primary and secondary syphilis , that exhibit acute symptoms. Our syphilis test did not distinguish the duration of the syphilis infection. Thus, our syphilis measure might have included late-latent syphilis of longer than 1 year's or unknown duration.
The Internet has been shown to play a role in recent syphilis outbreaks among MSM in the United States.30,31 syphilis infection. One possible explanation for the potentially protective effect is that Shanghai MSM who use the Internet might be better informed about HIV and other STIs than those who do not, and thus may be more likely to practice safer sex with partners met on-line. In our study, more than half of respondents (54%) reported the Internet as a place where they usually obtain information about AIDS (data not shown). In a study of MSM in Beijing, 44% of respondents reported using the Internet to access HIV prevention information during the prior 2 years.29 9
These findings should be interpreted with caution. First, we may not have obtained a representative sample of MSM in Shanghai; thus, estimates of HIV and syphilis prevalence could be biased. Second, self-reports of sexual behavior are subject to bias, because some respondents might not have accurately recalled their behavior or might have given socially desirable answers to sensitive questions. Third, no causal inferences can be drawn for the predictors of syphilis infection and high-risk behavior because of the cross-sectional nature of the data. Fourth, our data may not be generalizable to MSM residing in China's smaller cities and rural towns, because the greater tolerance for homosexuality in large urban areas, such as Shanghai, may influence behavior. More studies should be conducted with various MSM populations in China to validate the results of this study. Nonetheless, we were able to recruit a large sample of MSM in Shanghai, and we documented for the first time the characteristics of the social and sexual networks of Chinese MSM and the overlap between them.
In recent years, China has implemented new HIV prevention programs, such as methadone maintenance, syringe exchange, better management of blood collection, and free testing and counseling, as part of its efforts to combat the HIV epidemic.2 syphilis infection facilitates sexual transmission of HIV infection.32-34 syphilis infection and unprotected sexual behavior among MSM in Shanghai. We also found that social and sexual networks play a significant role in syphilis infection and unsafe sexual practices in this risk population. These findings suggest that HIV has the potential to spread rapidly among MSM and that network interventions should be considered to reduce the risk for infection. China must expand its commitment to MSM and provide them with resources for HIV prevention programs.
REFERENCES
1. China Ministry of Health, Joint United Nations Program on HIV/AIDS, and World Health Organization.
2005 Update on the HIV/AIDS Epidemic Response in China . Beijing, China: National Center for AIDS/STD Prevention and Control, China Centers for Disease Control and Prevention; 2006.
2. State Council AIDS Working Committee Office and United Nations Theme Group on HIV/AIDS in China.
A Joint Assessment of HIV/AIDS Prevention, Treatment and Care in China . Beijing, China: State Council AIDS Working Committee, National Center for AIDS/STD Prevention and Control, China, UNAIDS China Office; 2004.
3. Zhang L, Chen Z, Cao Y, et al. Molecular characterization of human immunodeficiency virus type 1 and hepatitis C virus in paid blood donors and injection drug users in China.
J Virol . 2004;78:13591-13599.
4. Yu X, Wang X, Mao P, et al. Characterization of HIV type 1 heterosexual transmission in Yunnan, China.
AIDS Res Hum Retroviruses . 2003;19:1051-1055.
5. Saksena NK, Wang B, Steain M, et al. Snapshot of HIV pathogenesis in China.
Cell Res . 2005;15:953-961.
6. Choi K, Liu H, Guo Y, et al. Emerging HIV epidemic among men who have sex with men.
Lancet . 2003;361:2125-2126.
7. Gu Y, Qu P, Su L, et al. Survey of knowledge, attitude, behavior and practice related to STI/HIV among male homosexuality in Shenyang.
Chinese Journal of Public Health . 2004;20:573-574.
8. Wu Y, Bai L, Liu Y, et al. Men who have sex with men and HIV/AIDS prevention in China. In: Lu Y, Essex M, eds.
AIDS in Asia . NewYork: Kluwer Academic/Plenum Publishers; 2004:365-377.
9. Tao X, Cai W, Cai Y, et al. Survey of related high risk behaviors of men in Shenzhen city.
Modern Preventive Medicine . 2004;31:247-248.
10. Zhu TF, Wang CH, Lin P, et al. High risk populations and HIV-1 infection in China.
Cell Res . 2005;15:852-857.
11. Knolle H. Age preference in sexual choice and the basic reproduction number of HIV/AIDS.
Biometrics . 1990;32:243-256.
12. Aral S, Hughes J, Stoner B, et al. Sexual mixing patterns in the spread of gonococcal and chlamydial infections.
Am J Public Health . 1999;89:825-833.
13. Gregson S, Nyamukapa CA, Garnett GP, et al. Sexual mixing patterns and sex-differentials in teenage exposure to HIV infection in rural Zimbabwe.
Lancet . 2002;359:1896-1903.
14. Morris M, Zavisca J, Dean L. Social and sexual networks: their role in the spread of HIV/AIDS among young gay men.
AIDS Educ Prev . 1995;7 (Suppl 5):24-35.
15. Service S, Blower S. HIV transmission in sexual networks: an empirical analysis.
Proc R Soc Lond B Biol Sci . 1995;260:237-244.
16. Rosenberg MD, Gurvey JE, Adler N, et al. Concurrent sex partners and risk for sexually transmitted diseases among adolescents.
Sex Transm Dis . 1999;26:208-212.
17. Fenton KA, Korovessis C, Johnson AM, et al. Sexual behaviour in Britain: reported sexually transmitted infections and prevalent genital Chlamydia trachomatis infection.
Lancet . 2001;358:1851-1854.
18. Potterat JJ, Zimmerman-Rogers H, Muth SQ, et al. Chlamydia transmission: concurrency, reproduction number and the epidemic trajectory.
Am J Epidemiol . 1999;150:1331-1339.
19. Koumans EH, Farley TA, Gibson JJ, et al. Characteristics of persons with
syphilis in areas of persisting
syphilis in the United States.
Sex Transm Dis . 2001;28:497-503.
20. Wasserheit JN, Aral SO. The dynamic topology of sexually transmitted disease epidemics: implications for prevention strategies.
J Infect Dis . 1996;174(Suppl 2):S201-S213.
21. Youm Y, Laumann E. Social network effects on the transmission of sexually transmitted diseases.
Sex Transm Dis . 2002;29:689-697.
22. Yoshikawa H, Wilson PA, Chae DH, et al. Do family and friendship networks protect against the influence of discrimination on mental health and HIV risk among Asian and Pacific Islander gay men?
AIDS Educ Prev . 2004;16:84-100.
23. Morel G. Logistic regression under complex survey designs.
Surv Methodol . 1989;15:203-223.
24. Liu H, Yang H, Li X, et al. Men who have sex with men and human immunodeficiency virus/sexually transmitted disease control in China.
Sex Transm Dis . 2006;33:68-76.
25. Lin X, Liu Y, Huang G, et al. A survey on STD/HIV infection among men who have sex with men and their sexual behavior.
Modern Preventive Medicine . 2004;31:618-619.
26. Yang Z, Fang X, Cai W, et al. Survey on
syphilis and HIV infection and sexual behavior among men having sex with male.
Chinese Journal of Public Health . 2003;19:1292-1293.
27. Jiang J, Cao N, Zhang J, et al. High prevalence of sexually transmitted diseases among men who have sex with men in Jiangsu Province, China.
Sex Transm Dis . 2006;33:118-123.
28. Zhang BC, Chu QS. MSM and HIV/AIDS in China.
Cell Res . 2005;15:858-864.
29. Choi K, Gibson DR, Han L, et al. High levels of unprotected sex with men and women among men who have sex with men: a potential bridge of HIV transmission in Beijing, China.
AIDS Educ Prev . 2004;16:19-30.
30. Klausner JD, Wolf W, Fischer-Ponce L, et al. Tracing a
syphilis outbreak through cyberspace.
JAMA . 2000;284:447-449.
31. Wong W, Chaw JK, Kent CK, et al. Risk factors for early
syphilis among gay and bisexual men seen in an STD clinic: San Francisco, 2002-2003.
Sex Transm Dis . 2005;32:458-463.
32. Fleming DT, Wasserheit JN. From epidemiological synergy to public health policy and practice: the contribution of other sexually transmitted diseases to sexual transmission of HIV infection.
Sex Transm Infect . 1999;75:3-17.
33. Paz-Bailey G, Meyers A, Blank S, et al. A case-control study of
syphilis among men who have sex with men in New York City: association with HIV infection.
Sex Transm Dis . 2004;31:581-587.
34. Taylor MM, Hawkins K, Gonzalez A, et al. Use of the serologic testing algorithm for recent HIV seroconversion (STARHS) to identify recently acquired HIV infections in men with early
syphilis in Los Angeles County.
J Acquir Immune Defic Syndr . 2005;38:505-508.
