Risk Factors for HIV Diagnosis Among Men Who Have Sex with Men: Results of a Case–Control Study in One Sample of Eastern China

Abstract

Substantial increases in human immunodeficiency virus (HIV) have been reported worldwide in recent years, particularly among men who have sex with men (MSM). We conducted a matched case–control study to examine the factors associated with HIV diagnosis among MSM in one sample of eastern China. Between February 2012 and December 2014, we used surveillance records to identify MSM diagnosed with HIV (case participants); we also recruited MSM who did not have HIV (controls) and then matched them (2:1) with control cases in terms of age (±3 years). Multivariate logistic regression models were used to assess the factors associated with HIV diagnosis. According to a multivariate analysis using logistic regression model involving 101 cases and 202 matched controls, a lack of comprehensive knowledge of HIV (adjusted odds ratio [OR] = 0.40; 95% confidence interval [CI] = 0.18, 0.89), a monthly income of ≥4,000 RMB (adjusted OR = 2.99; 95% CI = 1.45, 6.16), having at least two male sexual partners in the past 6 months (adjusted OR = 2.85; 95% CI = 1.28, 6.31), participating in at least four anal sex experiences with a man in the past month (adjusted OR = 3.56; 95% CI = 1.64, 7.73), and having a current syphilis infection (adjusted OR = 3.30; 95% CI = 1.06, 10.25) were associated with an increased risk for HIV diagnosis. MSM with a comprehensive knowledge of HIV were at reduced risk of HIV diagnosis, whereas those with more male sexual partners, more male anal sexual experiences (including receptive or/and insertive anal intercourse, rimming, and fisting), and a current syphilis infection were at increased risk of HIV diagnosis. Focus on protection and safer sex behaviors during male sexual activity (i.e., consistent condom use, pre-exposure prophylaxis, closed sexual networks among MSM) would likely be effective for reducing the HIV transmission rate.

Background

The prevalence of human immunodeficiency virus (HIV) among men who have sex with men (MSM) has increased dramatically worldwide in recent years.^1

–6 MSM constitute a marginalized cohort at high risk for sexually transmissible infections (STIs), including HIV diagnosis,^7,8 due to some of their risky sexual behaviors, such as multiple sexual partners^9,10 and unprotected anal intercourse (UAI).^11
–13 Currently, MSM represent a dominant driving force of the HIV epidemic in China.¹⁴ In 2009, MSM accounted for 33% of new HIV diagnosis in China¹⁵; this is particularly evident in the city of Huzhou city, eastern China, where there is a low AIDS epidemic area in China, but has rapidly increasing HIV prevalence among MSM of all new reported HIV diagnosis from 4.1% in 2009 to 31.7% in 2012.¹¹ In 2014, MSM accounted for 41.7% of all new reported HIV diagnosis in Huzhou.

Previous studies have identified factors associated with HIV diagnosis among MSM in China^6,9,10,16; subjects who had a history of being infected by syphilis, who often went to bars for sexual partners, who were with a monthly income of more than RMB yen 1,000, and who were money boys were more likely to be HIV positive, while receipt of condoms, consistent condom use during last intercourse with a male, consistent condom use during last intercourse with a female, and subjects who had sex with fixed sexual partners or had the protected sex in the past 6 months were associated with a lower risk of HIV diagnosis. However, limited research has been conducted using matched case–control designs, which not only could study the association between multiple factors and certain diseases simultaneously (especially for the cause of exploratory research) but also save manpower, time, and money; it is easy to organize and conduct on-site investigations as well. Therefore, the purpose of this matched case–control study was to identify factors associated with HIV diagnosis in MSM. We assessed associations between related sociocultural factors and the degree of risky behavior among MSM in Huzhou.

Methods

Study population and data collection

This study focused on MSM aged between 16 and 54 years of age in Huzhou city, Zhejiang province, eastern China. Participants were recruited between February 2012 and December 2014. A total of 162 HIV diagnoses among individuals who met the criteria for participation as cases were reported to the Huzhou Center for Disease Control and Prevention; 113 (69.8%) of these men were interviewed. Of the 49 who were not interviewed, 26 could not be contacted, 12 had moved out of the area, 3 were deceased, and 6 declined to participate, and no reason recorded for the nonparticipation of the remaining 2 cases. Of the 113 potential cases, 101 (89.4%) participants who were interviewed were included in the study and 12 were excluded because they reported no male-to-male sexual contact at their initial interview. Of the 432 potential controls who completed the screening survey, 202 (46.8%) were eligible and were included in the study matched (2:1) to cases for age (±3 years). Of the 230 potential controls for ineligibility, 117 (50.9%) were not reporting male–male sexual behavior during the previous 6 months, 86 (37.4%) were not collected blood to test HIV or syphilis, and 27 (11.7%) were the key variable (such as general demographic characteristics, number of male sexual partners or male sexual behaviors, and condom use) deficiency of the questionnaire.

In this matched case–control study, we defined case participants as MSM diagnosed with HIV; all such participants lived or were diagnosed in Huzhou. Controls were recruited as MSM who were not infected with HIV during this period in Huzhou. We identified these individuals using the Chinese HIV/AIDS Reporting System data for Huzhou, which includes information on age, ethnicity, date of diagnosis, route of transmission, and county of residence. We attempted to recruit all potential case participants by telephone or in person.

The protocol was approved by the Ethics Review Committee of the Huzhou Center for Disease Control and Prevention, and all respondents provided written informed consent before their participation. Face-to-face interviews were conducted by five well-trained peer interviewers. The cohort was divided into case (n = 101) and control (n = 202) groups.

Measurements

Potential participants completed an anonymous structured questionnaire pertaining to general demographic characteristics (such as age, marital status, educational level, registered residence, monthly income, and self-identified sexual orientation), risky behaviors (number of male sexual partners in the past 6 months, number of male anal sexual experiences, including receptive or/and insertive anal intercourse, rimming, and fisting, in the previous month, and whether a condom was used during the last male anal sexual behavior and during any other such behaviors during the past 6 months), prevalence of STIs, and utilization of HIV prevention services (i.e., whether they had used any HIV-related prevention services such as public education material, condoms, or advice of health educators in the past 12 months). Eight HIV-related questions were included to assess the knowledge of respondents about HIV; individuals who responded correctly to at least six questions were considered to possess comprehensive knowledge of HIV.¹¹ Blood samples for HIV and syphilis antibodies were collected from all participants. We used an enzyme-linked immunosorbent assay (ELISA, Manufacturer of Infectious Diseases Diagnostics of Beijing Wantai, Beijing, China) for HIV antibodies' screening or syphilis antibodies' screening, used western blot immunoassay (Singapore MP Biomedical Asia Pacific Ltd., Singapore, Singapore) for HIV-1/2 confirmation when the ELISA screen for HIV antibodies was positive, and performed the tolulized red unheated serum test (TRUST, Manufacturer of Infectious Diseases Diagnostics of Beijing Wantai, Beijing, China) for confirming syphilis infection. We provided pre- and posttest HIV/syphilis counseling for all the subjects and made a treatment or referral to hospital for those HIV/syphilis-positive subjects.

Statistical analyses

Bivariate logistic regression analysis was used to assess differences between the case and control groups. Because an educational level of high school or above may affect the extent of knowledge of HIV, we controlled for this factor. Variables with significant bivariate between-group differences (p < .1) were used as candidate variables in a “Forward stepwise regression” multivariate logistic regression model. The independent variables are listed in Tables 1 and 2, and a summary multivariate model was fitted to identify variables independently associated with HIV diagnosis. We estimated adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for HIV diagnosis. All statistical analyses were performed using the SPSS for Windows software package (ver. 19.0; SPSS Inc., Chicago, IL, USA), and a p-value <.05 was taken to indicate statistical significance.

Table 1.

Characteristics of Recruited Men Who Have Sex with Men

	Control group (n = 202)		Case group (n = 101)
Category	n	No. (Col %)	n	No. (Col %)	p
Age (Mean ± SD, years)	27.9 ± 6.8		28.5 ± 7.6		.49^a
Marital status
Unmarried	140	69.3	65	64.4	.68^b
Married	44	21.8	28	27.7
Cohabitation	4	2.0	1	1.0
Divorced/widowed	14	6.9	7	6.9
Registered Residence
Huzhou city	105	52.0	54	53.5	.11^c
Other places	97	48.1	47	46.5
Ethnicity
Han	198	98.0	100	99.0	.67^b
Minority	4	2.0	1	1.0
Local residence time
2 years and below	74	36.6	48	47.5	.07^c
More than 2 years	128	63.4	53	52.5
Education level
Below high school	41	20.3	37	36.6	.002^c
High school or above	161	79.7	64	63.4
Self-identified sexual orientation
Bisexual or uncertain	125	61.9	54	53.5	.16^c
Gay/homosexual	77	38.1	47	46.5

p-Values were determined by the T-test.

p-Values were determined by the Fisher exact test.

p-Values were determined by the chi-square analysis.

Table 2.

Results of Bivariate Analysis of Potential Associations Between Factors and HIV Infection Among Men Who Have Sex with Men

	Control group (n = 202)		Case group (n = 101)
Factors	n	No. (Col %)	n	No. (Col %)	p	Unadjusted OR (95% CI)
Education level
Below high school	41	20.3	37	36.6	<.01	0.44 (0.26, 0.76)
High school or above	161	79.7	64	63.4
Registered residence
Other places	97	48.0	47	46.5	.81	1.06 (0.66, 1.67)
Huzhou city	105	52.0	54	53.5
A monthly income of
<4,000 RMB	154	76.2	54	53.5	<.01	3.01 (1.74, 5.22)
≥4,000 RMB	48	23.8	47	46.5
Sexual orientation
Bisexual or uncertain	125	61.9	54	53.5	.16	1.41 (0.87, 2.28)
Gay/homosexual	77	38.1	47	46.5
Marital status
Married or cohabitation	48	23.8	29	28.7	.3	0.72 (0.39, 1.33)
Unmarried or divorced	154	76.2	72	71.3
Fixed homosexual sex partner
No	124	61.4	64	63.4	.75	0.93 (0.58, 1.49)
Yes	78	38.6	37	36.6
With a comprehensive knowledge of HIV
Does not have a comprehensive knowledge	31	15.3	33	32.7	<.01	0.37 (0.20, 0.66)
Has a comprehensive knowledge	171	84.7	68	67.3
No. of male sex partners in the past 6 months
1 or none	134	66.3	35	34.7	<.01	4.95 (2.65, 9.23)
2 or more	68	33.7	66	65.3
Ever had male anal sexual behavior in the past 6 months
No	30	14.9	7	6.9	.048	2.42 (1.01, 5.79)
Yes	172	85.1	94	93.1
No. of male anal sexual behavior in the past month
3 or below	166	82.2	46	45.5	<.01	5.80 (3.18, 10.56)
4 or more	36	17.8	55	54.5
Using condom during the last male anal sexual behavior
No	96	47.5	64	63.4	<.01	0.50 (0.30, 0.83)
Yes	106	52.5	37	36.6
Always using condom during male anal sexual behavior in the past 6 months
No	137	67.8	85	84.2	<.01	0.40 (0.21, 0.73)
Yes	65	32.2	16	15.8
Ever had heterosexual sex partner
No	137	67.8	63	62.4	.35	1.26 (0.77, 2.06)
Yes	65	32.2	38	37.6
Ever met heterosexual sex partner by Internet
No	50	24.8	30	29.7	.36	0.78 (0.46, 1.33)
Yes	152	75.2	71	70.3
Ever drunk, used stimulants, or aphrodisiac drugs during male sexual behavior
No	161	79.7	77	76.2	.50	1.22 (0.69, 2.13)
Yes	41	20.3	24	23.8
Ever had group sex, abuse sex, one night stand, or sex party during male sexual behavior
No	112	55.4	47	46.5	.15	1.43 (0.88, 2.31)
Yes	90	44.6	54	53.5
Ever been infected with sexually transmitted diseases
No	158	78.2	64	63.4	<.01	2.03 (1.20, 3.42)
Yes	44	21.8	37	36.6
Ever tested for HIV
No	88	43.6	46	45.5	.74	0.92 (0.57, 1.50)
Yes	114	56.4	55	54.5
Ever received HIV prevention services (such as propaganda materials and condoms, consulting services, and training seminars) in the past year
No	95	47.0	39	38.6	.16	1.43 (0.87, 2.34)
Yes	107	53.0	62	61.4
A current syphilis infection
No	185	91.6	81	80.2	<.01	2.69 (1.32, 5.47)
Yes	17	8.4	20	19.8

CI, confidence interval; HIV, human immunodeficiency virus; OR, odds ratio.

Results

Background characteristics

As shown in Table 1, the respondents' mean age was 28.2 ± 7.1 years (range: 16–54 years). There were no significant between-group differences (p > .05) in age, marital status, registered residence, ethnicity, local residence time, or sexual orientation.

Bivariate analysis of associations between factors and HIV diagnosis rate among MSM

The results of the bivariate analysis are summarized in Table 2: an educational level of high school or above (OR = 0.44; 95% CI = 0.26, 0.76), comprehensive knowledge of HIV (OR = 0.37; 95% CI = 0.20, 0.66), condom use during the last male anal sexual behavior (OR = 0.50; 95% CI = 0.30, 0.83), and always using a condom during male anal sexual behavior in the past 6 months (OR = 0.40; 95% CI = 0.21, 0.73) were significantly associated with reduced odds of HIV diagnosis. In contrast, a monthly income of ≥4,000 RMB (OR = 3.01; 95% CI = 1.74, 5.22), at least two male sexual partners in the past 6 months (OR = 4.95; 95% CI = 2.65, 9.23), engagement in any male anal sexual behavior in the past 6 months (OR = 2.42; 95% CI = 1.01, 5.79), at least four anal sex experiences with a man in the past month (OR = 5.80; 95% CI = 3.18, 10.56), ever having been infected with STIs (OR = 2.03; 95% CI = 1.20, 3.42), and a current syphilis infection (OR = 2.69; 95% CI = 1.32, 5.47) were significantly associated with increased odds of HIV diagnosis.

Multivariate summary model

Educational level and the factors listed in Table 2 (p < .1) were eligible for inclusion in the multivariate analysis using the logistic regression model. The “Forward stepwise regression” option was used to assess the significant factors. As shown in Table 3, the multivariate analysis that controlled for educational level showed that those with a comprehensive knowledge of HIV were less likely to be infected with HIV (adjusted OR = 0.40; 95% CI = 0.18, 0.89). Furthermore, those with a monthly income of ≥4,000 RMB (adjusted OR = 2.99; 95% CI = 1.45, 6.16), at least two male sexual partners in the past 6 months (adjusted OR = 2.85; 95% CI = 1.28, 6.31), at least four anal sex experiences with a man in the previous month (adjusted OR = 3.56; 95% CI = 1.64, 7.73), and a current syphilis infection (adjusted OR = 3.30; 95% CI = 1.06, 10.25) were more likely to be infected with HIV.

Table 3.

Results of Multivariate Analysis of Factors Associated with HIV Infection Among Men Who Have Sex with Men

Factors	B	SE	p	Adjusted OR (95% CI)
A monthly income of ≥4,000 RMB	1.094	0.369	.003	2.99 (1.45, 6.16)
With a comprehensive knowledge of HIV	−0.912	0.406	.025	0.40 (0.18, 0.89)
Having at least two male sexual partners in the past 6 months	1.046	0.406	.01	2.85 (1.28, 6.31)
Participating in at least four anal sex experiences with a man in the past month	1.27	0.396	.001	3.56 (1.64, 7.73)
A current syphilis infection	1.194	0.578	.039	3.30 (1.06, 10.25)

Adjusted ORs and CIs were derived from multiple logistic regression analyses that included only the factors listed.

Discussion

Our research implies that those with a comprehensive knowledge of HIV are less likely to be infected with HIV compared with those without such knowledge; therefore, increasing knowledge about HIV by improving public awareness could reduce the risk of HIV diagnosis. Efforts toward increasing knowledge of HIV among MSM should be made in MSM-based HIV/AIDS prevention programs. A previous study indicated that awareness of infection is vital during antiretroviral treatment of AIDS to reduce the risk of transmission among HIV-positive MSM individuals.¹⁷

Our data also show that those who earned ≥4,000 RMB per month, had at least two male sexual partners in the past 6 months, or had at least four male anal sexual experiences in the past month were more likely to be infected with HIV. Spearman correlation analysis also indicated that there was no association between earning ≥4,000 RMB per month and having at least two male sexual partners in the past 6 months or at least four male anal sexual experiences in the past month. These data may represent an entry point for HIV prevention interventions among men with a high monthly income; such interventions should tap existing social dynamics to promote procondom norms among sexually active MSM. Indeed, the promotion of consistent condom use during male sexual behavior is a pivotal approach to reducing the risk of HIV diagnosis among MSM.⁷ In addition, interventions to prevent HIV should emphasize the social acceptability of condoms and the importance of condom use.¹¹

Consistent with previous reports, infection with STIs, such as syphilis, was associated with risky sexual behaviors.^12,18 This study attempted to distinguish MSM according to whether respondents were presently infected with syphilis. Our data indicate that those who were infected were more likely to also be infected with HIV. It is well known that syphilis infection increases the odds for HIV infection.¹² Improving awareness of HIV and STIs transmission among MSM and curing STIs timely and normatively may help to reduce the risk of epidemics of HIV and other STIs.

HIV prevention services should emphasize the importance of consistently using condoms during male anal sexual encounters to reduce the risk of HIV diagnosis, as this occurs primarily through UAI in MSM. However, perhaps due to this case–control study design and limited sample size, our results indicate that those who used a condom during their last male anal sexual behavior or who always used a condom during male anal sexual experiences during the previous 6 months were not significantly more likely to be infected with HIV. Future research on whether consistent condom use is associated with lower risk of HIV diagnosis should focus on how more advanced response improves the cognitive competence for regular health reasoning.

This study had several limitations. First, because the design of this study is the case–control study, the generalizability of the results of study is limited. Although numerous previous studies of MSM used similar methods, the convenience sampling employed herein was an improper means of random sampling that may have reduced the generalizability of our findings to other MSM populations. Furthermore, although frequently used in sexual behavior studies, self-reported data may be subject to reporting biases, including recall and social desirability biases.¹¹ Respondents may exhibit bias when answering questions pertaining to condom use during the last male anal sexual behavior experiences and to whether they always used a condom during male anal sexual behavior experiences during the past 6 months. Accurate recall of condom use during male anal sexual experiences is notoriously difficult, especially in the context of drinking or drug use, although over three-quarters of both groups denied any alcohol or drugs during sex. To reduce reporting biases, we reassured respondents regarding the strict anonymity and privacy of the interviews. Finally, although we already knew the HIV status of respondents, because biological marker data (i.e., blood samples) were collected, we did not assess HIV subtype; further research is required to distinguish cases according to HIV subtype.

Conclusions

The findings of this study have implications regarding the factors associated with HIV diagnosis among MSM; those with a comprehensive knowledge of HIV were less likely to be infected with the virus, whereas having more male sexual partners, engaging in anal sex with a man, and current syphilis infection increased the odds of HIV diagnosis. Focus on protection and safer sex behaviors during male sexual activity (i.e., consistent condom use, pre-exposure prophylaxis, closed sexual networks among MSM) would likely be effective for reducing the HIV transmission rate.

Footnotes

Acknowledgments

The authors would like to thank everyone for their valuable contributions to this article. The study in this article was funded by the Mega-Project for National Science and Technology Development under the “11th and 12th, Five-Year Plan of the People's Republic of China” (2011ZX09302-003-03, 2013ZX10004-904), Zhejiang Province medical and health research project (2015KYB384), Huzhou Science and Technology Planning projects (2015GY10) & Zhejiang Provincial Program for the Cultivation of High-level Innovative Health talents. They had no role in study design, data collection and analysis, decision to publish, or preparation of the article.

Authors' Contributions

M.J., N.W., J.L., and Z.D. participated in the study design and data collection. Z.H. and Z.Y. performed the statistical analysis. Z.Y., S.Z., and M.J. developed the study, participated in its design and coordination, and helped to draft the article. All authors read and approved the final article.

Author Disclosure Statement

No competing financial interests exist.

References

Park

, Papworth

, Kassegne

, Moukam

, Billong

, Macauley

, Yomb

, Nkoume

, Mondoleba

, Eloundou

, et al.: HIV prevalence and factors associated with HIV infection among men who have sex with men in Cameroon. J Int AIDS Soc, 2013; 16 Suppl 3:18752.

Mimiaga

, Biello

, Sivasubramanian

, Mayer

, Anand

, Safren

: Psychosocial risk factors for HIV sexual risk among Indian men who have sex with men. AIDS Care, 2013; 25:1109–1113.

Scott

, Vittinghoff

, Irvin

, Sachdev

, Liu

, Gurwith

, Buchbinder

: Age, race/ethnicity, and behavioral risk factors associated with per contact risk of HIV infection among men who have sex with men in the United States. J Acquir Immune Defic Syndr, 2014; 65:115–121.

Mayer

, Wang

, Koblin

, Mannheimer

, Magnus

, del Rio

, Buchbinder

, Wilton

, Cummings

, Watson

, et al.: Concomitant socioeconomic, behavioral, and biological factors associated with the disproportionate HIV infection burden among Black men who have sex with men in 6 U.S. cities. PLoS One, 2014; 9:e87298.

Knussen

, Flowers

, McDaid

: Factors associated with recency of HIV testing among men residing in Scotland who have sex with men. AIDS Care, 2014; 26:297–303.

Liu

, Qu

, Ezeakile

, Zhang

, Liang

: Factors associated with HIV infection among men who have sex with men in Henan Province, China: A cross-sectional study. BMC Public Health, 2013; 13:356.

Ramanathan

, Chakrapani

, Ramakrishnan

, Goswami

, Yadav

, Subramanian

, George

, Paranjape

: Consistent condom use with regular, paying, and casual male partners and associated factors among men who have sex with men in Tamil Nadu, India: Findings from an assessment of a large-scale HIV prevention program. BMC Public Health, 2013; 13:827.

Okall

, Otieno

, Nyikuri

, Mills

, Hardnett

, Odero

, Turner

, Gust

: Men who have sex with men in Kisumu, Kenya: Support group membership and knowledge of HIV-risk factors. Cult Health Sex, 2013; 15:968–980.

Bai

, Huan

, Tang

, Chen

, Yan

, Liu

, Yang

, Peng

, Zhao

, Yu

, et al.: A survey of HIV infection and related high-risk factors among men who have sex with men in Suzhou, Jiangsu, China. J Biomed Res, 2011; 25:17–24.

10.

Feng

, Wu

, Detels

, Qin

, Liu

, Wang

, Zhang

: HIV/STD prevalence among men who have sex with men in Chengdu, China and associated risk factors for HIV infection. J Acquir Immune Defic Syndr, 2010; 53 Suppl 1:S74–S80.

11.

Jin

, Yang

, Dong

, Han

: Correlates of consistent condom use among men who have sex with men recruited through the Internet in Huzhou city: A cross-sectional survey. BMC Public Health, 2013; 13:1101.

12.

Das

, Li

, Zhong

, Ouyang

, Mahapatra

, Tang

, Fu

, Zhao

, Detels

: Factors associated with HIV and syphilis co-infection among men who have sex with men in seven Chinese cities. Int J STD AIDS, 2015; 26:145–155.

13.

Pham

, Nguyen

, Hoang

, Cao

, Khuu

, Phan

, Mai

, Tran

, Wilson

, Zhang

: Prevalence of HIV/STIs and associated factors among men who have sex with men in An Giang, Vietnam. Sex Transm Dis, 2012; 39:799–806.

14.

Wang

, Wang

, Norris

, Li

, Guo

, Ding

, Wang

: HIV prevalence and influencing factors analysis of sentinel surveillance among men who have sex with men in China, 2003–2011. Chin Med J, 2012; 125:1857–1861.

15.

Wang

, Wang

, Wu

, Guo

, Sun

, Poundstone

, Wang

, National Expert Group on HIV/AIDS Estimation: Estimating the number of people living with HIV/AIDS in China: 2003–2009. Int J Epidemiol, 2010; 39 Suppl 2:ii21–ii28.

16.

Zhang

, Zhang

, Yu

, Wang

, Liu

, Wang

, Li

, Shang

, Li

: Prevalence of HIV infection and associated risk factors among men who have sex with men (MSM) in Harbin, P. R. China. PLoS One, 2013; 8:e58440.

17.

Wejnert

, Le

, Rose

, Oster

, Smith

, Zhu

, Gabriela Paz-Bailey for the NHBS Study Group: HIV infection and awareness among men who have sex with men-20 cities, United States, 2008 and 2011. PLoS One, 2013; 8:e76878.

18.

She

, Zhang

, Wang

, Xu

, Zhang

, Fan

, Sun

, Zhang

: Associated factors for HIV and syphilis infection among men who have sex with men only and men who have sex with both men and women in cities of China. Int J STD AIDS, 2013; 24:293–300.