Study participants and questionnaire
In order to better understand the HIV epidemic among Chinese MSM, a multicenter cross-sectional survey was conducted among Chinese MSM in seven large cities in China (Shenyang, Ji’nan, Zhengzhou, Shanghai, Nanjing, Changsha, and Kunming) from June 2012 to June 2013. The cruising areas and service points for MSM were used as the sampling sites. Site-specific sampling periods were determined based on attendance and hours of operation. Participants were recruited by multiple approaches: advertisements on gay websites, collaborating with local MSM community-based organizations, peer referrals, and venues such as gay bars and bathrooms visited by MSM. To be eligible for this study, participants should be physically male aged 16 years or older, self-reported anal/oral sex experiences within the last year, and able to provide informed consent.
Eligible participants completed an anonymous structured questionnaire concerning sociodemographics, recent sexual behaviors, and other HIV related risk factors. (1) demographics: age, residence, city, education, occupation, marital status, monthly income (USD), predominant sex position in anal intercourse (AI); (2) recent sexual behaviors or HIV risk factors: age of sexual debut with males, main venue of seeking male sexual partners in the past 6 months, group sex in the past 6 months, number of male sexual partners in the past 6 months, commercial sex in the past 6 months, mucosally-traumatic sex in the past 6 months, condom break during AI in the past 6 months, STI-related symptoms in the past year, non-Chinese male sexual partners in the past 6 months.
The positive attitude towards the prevention of HIV infection was assessed by nine relevant questions. If the participant answered all questions correctly, he/she was defined as having an “adequate” positive attitude towards the prevention of HIV infection. The questions were: (1) Is it possible for a person who looks healthy to carry HIV? (2) Is it possible to be infected through transfusion of blood or blood products with HIV? (3) Is it possible to be infected through sharing needles with HIV-infected persons or AIDS patients? (4) Can the proper use of condoms in each sexual activity reduce the risk of HIV transmission? (5) Can having sex with only a single HIV-uninfected sexual partner reduce the risk of HIV transmission? (6) Can an HIV-infected pregnant woman transmit HIV to her child? (7) Is it possible to be infected through eating with HIV-infected persons or AIDS patients? (8) Is it possible to be infected through mosquito bites? (9) If you know or suspect that your partner has AIDS, will you stop having sex with him?
Participants were asked about their use (nonmedical or recreational) of seven commonly used RDs in parties or during sexual contact in the past 6 months: poppers, ecstasy, methamphetamine, amphetamine, codeine, tramadol, and ketamine. Thus, participants were grouped based on the number of drugs used in the past 6 months (i.e., no drug, single drug, two types of drugs, and ≥3 types of drugs). “Polydrug use” was defined as using ≥2 types of drugs simultaneously in the past 6 months.
Laboratory testing
Samples of venous blood were collected from participants to diagnose HIV-1 antibody. HIV-1 antibody was tested using enzyme-linked immunosorbent assay [ELISA] (bioMerieux, Durham, NC, USA), and HIV-seropositive specimens were confirmed by western blotting [WB] (HIV Blot 2.2 WBTM, Genelabs Diagnostics, Singapore). The antibody test for HIV was conducted in respective provincial HIV laboratories of CDC to which the seven study sites were affiliated. These WB-positive samples were tested by the immunoglobulin G (IgG)-capture BED-enzyme immunoassay [BED-CEIA] (Calypte Biomedical Corporation, Rockville, MD, USA) at the key laboratory in Shenyang. Based on the measurement of HIV-1-specific IgG to total IgG after seroconversion, BED-CEIA was able to distinguish between recent and established HIV-1 infections [26, 27]. To test cross-sectional specimens, we followed the algorithm shown in Figure 1. The calibrator (CAL) and control (including high-positive control, low-positive control, and negative control) specimens were tested in triplicate on every plate, and median values were used to calculate the normalized OD (ODn; ODn=specimen OD/calibrator OD). Specimens with initial ODn>1.2 were classified as established infection. Specimens with initial ODn≤1.2 were repeated tested in triplicate to confirm their ODn values by using the median values of the triplicate values. During the retesting, if median ODn values were <0.8, the specimens were considered to be recently infected [28].
Statistical analyses
The chi-square test was used to determine the significance of differences in social demographics of the different subgroups. The Cochran–Armitage trend test was used to analyze the association between different subgroups and social demographics. We calculated the prevalence of nine defined HIV high-risk behaviors in each subgroup. We estimated the ratio of the prevalence of unadjusted HIV high-risk behaviors through the chi-square test, and adjusted this prevalence ratio for social demographics through multivariate logistic regression analysis. An alpha of 0.05 was considered significant. The HIV incidence was estimated using a formula to adjust sensitivity/specificity, and the time window to define recent HIV infection was 168 days. The formula and parameters were recommended by the Chinese Centers for Disease Control and Prevention [29]. Then, we used multivariable logistic regression to determine the adjusted odds ratios (AORs) and respective 95% confidence intervals (CIs) of the different subgroups. We adjusted for social demographics using recent or established HIV infection, as defined by the BED-CEIA, as the outcome. Statistical analyses were carried out using SAS 9.2 (SAS Institute, Cary, NC, USA) and STATA 13.0 (Stata Corporation, College Station, TX, USA).