Analysis of trends and risk factors of HIV incidence in men who have sex with men in Chongqing, China: a retrospective cohort study, 2013–2022

Abstract

Background

China has reported a high rate of HIV transmission through male-to-male sexual contact. This study aims to update the trend in HIV incidence density and identify associated risk factors among men who have sex with men (MSM) in Chongqing, China.

Methods

Data on HIV incidence and sexual behaviors were obtained from the Chongqing CDC for MSM from 2013 to 2022. HIV incidence density was calculated biennially, along with the overall incidence rate. Cox proportional hazards regression was used to identify risk factors associated with HIV infection.

Results

A total of 2,134 HIV-negative MSM were included in the cohort. The overall HIV incidence density was 4.9 (95% CI: 4.3–5.6)/100 person-years (PYs) from 2013 to 2022. The HIV incidence density among MSM showed a declining trend. Multivariable analysis showed that MSM with a high school education or below (HR 1.57, 95% CI: 1.21–2.03), an anal sex role of versatile (HR 1.38, 95% CI: 1.03–1.85), seeking male sexual partners through offline methods (HR 1.73, 95% CI: 1.34–2.23), engaging in unprotected anal intercourse (HR 1.53, 95% CI: 1.19–1.97), and using rush poppers in the past six months (HR 1.60, 95% CI: 1.09–2.33) were identified as significant risk factors associated with HIV infection.

Conclusion

The decline in HIV incidence density among MSM from 2013 to 2022 appeared to reflect the impact of comprehensive HIV prevention strategies implemented in Chongqing. Continued efforts to raise awareness of HIV infection, promote condom use, and reduce high-risk sexual behaviors among MSM remain essential public health priorities.

Clinical trial

Not applicable.

Introduction

HIV/AIDS continues to pose a significant public health challenge both globally and within China. As of December 31, 2023, a total of 1,289,700 individuals living with HIV/AIDS were reported in China, with 457,609 cumulative deaths. In 2023, homosexual transmission accounted for 25.7% of newly reported HIV/AIDS cases [1].

Men who have sex with men (MSM) are defined as males who engage in sexual activity with other men, including homosexual, bisexual, and heterosexual individuals who have same-sex encounters under specific circumstances. According to a United Nations report, MSM face a 24-fold higher risk of HIV infection compared to the general population across all global regions [2]. Since 2007, when sexual transmission became the predominant mode of HIV transmission in China, the proportion of HIV cases attributed to male-to-male sexual contact has increased from 3.4% in 2007 to 25.5% in 2018 and 25.7% in 2023 [1, 3, 4]. Due to their heightened vulnerability to HIV infection, MSM are widely recognized as a key population for targeted prevention efforts [5]. In addition, MSM sexual risk behaviors are influenced not only by individual-level factors but also by structural and socio-environmental conditions [6]. In addition, some MSM may engage in unprotected heterosexual intercourse (UAI), which may be influenced by a range of factors, including sociocultural expectations related to marriage and reproduction in the Chinese context. Consequently, MSM may act as a “bridging population” for HIV transmission to female partners in the general population [7, 8]. The rapid increase in HIV prevalence among MSM is driven by multiple factors, including multiple sexual partners, high-risk sexual behaviors, low HIV testing uptake, coexisting sexually transmitted infections (STIs), stigma surrounding homosexuality, and traditional expectations related to marriage and reproduction. These factors collectively present significant and complex challenges for the prevention and control of HIV/AIDS among MSM in China.

HIV incidence refers to the rate of new infections within a specific population over a defined period and serves as a critical indicator for HIV surveillance efforts [9]. Accurate estimation of HIV incidence is essential for monitoring epidemic trends and evaluating the effectiveness of prevention interventions [10]. However, research on HIV incidence among MSM in China remains limited, with most studies relying on repeated cross-sectional designs. HIV incidence density reflects the average rate of new infections over a given period, typically expressed as the number of infections per 100 person-years (PYs) [11]. The numerator includes newly diagnosed HIV cases within the specified timeframe, while the denominator represents the total person-time at risk contributed by study participants. Incidence density rates among MSM vary across regions. For example, a cohort study conducted in Shanghai and Tianjin reports an incidence density of 2.73/100 PYs among 455 participants [12]. Another cohort study, involving 204 MSM recruited in Kunming, Yunnan Province, reports an incidence density of 2.78/100 PYs from 2018 to 2020 [13].

Chongqing, the largest directly administered municipality in China, is located in the southwest region of the country. Its economic and cultural appeal has attracted a substantial number of migrant workers from across China. The high mobility of the MSM population may increase the risk of HIV infection by facilitating changes in sexual networks, increasing the likelihood of multiple or anonymous partnerships, and disrupting continuity of access to prevention and testing services [14–16]. In terms of sexual behavior, previous study have shown that MSM in Chongqing are primarily young and middle-aged, with the anal sex role of the top being the most commonly reported practice. High-risk behaviors are prevalent among MSM in the city and contribute to both increased HIV risk and cross-regional transmission [17]. A study analyzing HIV/AIDS epidemiological trends in Chongqing from 1995 to 2015 reports a consistent annual increase in the proportion of infections attributed to homosexual transmission, with a marked rise after 2007. By that time, MSM accounted for 20.55% of all newly diagnosed HIV cases [18]. This indicator reflects the distribution of transmission routes rather than the overall infection level. Furthermore, a five-year cross-sectional study conducted from 2013 to 2017 shows a gradual decline in HIV prevalence among MSM, with reported HIV-positive rates of 23.0%, 21.5%, 20.4%, 19.8%, and 19.2%, respectively [19]. However, the trend in HIV infection among MSM in Chongqing after 2017 remains unclear. Up-to-date data on HIV incidence and its associated risk factors are essential for informing evidence-based prevention strategies and public health interventions [20].

Therefore, the present study aims to update the trend in HIV incidence density among MSM in Chongqing through 2022 and to identify factors associated with HIV infection in this population. These findings are intended to support the development of targeted prevention strategies and to strengthen efforts to control the HIV epidemic among MSM.

Materials and methods

Study design and sampling methods

An open cohort study was conducted among MSM in Chongqing from January 2013 to December 2022. The study was implemented through the Voluntary Counseling and Testing (VCT) clinic at the Chongqing Center for Disease Control and Prevention (CDC), using multiple recruitment strategies, including VCT services, outreach activities, and peer or volunteer referrals. Eligible participants were males aged 16 years or older who had engaged in oral and/or anal sex with other men in the past year, tested HIV-negative at enrollment, and voluntarily consented to participate in a structured questionnaire survey, fingerprint registration (used to identify duplicate testing records belonging to the same individual), and serological testing. Inclusion also required at least two recorded HIV tests during the study period. Data collection was conducted by investigators who had undergone standardized training, strictly following predefined procedures, including face-to-face questionnaire interviews, blood sample collection, and laboratory testing. The questionnaire was administered in an interviewer-led format, with investigators asking questions and participants providing responses. During the questionnaire administration, investigators explained the specific meanings of items as needed. Upon completion of the survey, all original questionnaire data were subjected to double data entry, followed by parallel logical consistency checks to ensure data quality.

Follow-up visits and intervention

Questionnaire data were collected through face-to-face interviews conducted by trained researchers. Following the interviews, professional counselors provided participants with tailored guidance and risk-reduction strategies to prevent HIV and other sexually transmitted infections. Blood samples were then collected from each participant for HIV and syphilis testing. Upon returning to receive their HIV test results, participants were also provided with post-test counseling. After the initial visit, all participants (excluding those who tested HIV-positive and were referred to specialized antiretroviral therapy institutions) were encouraged to return every six months for follow-up visits, during which additional blood samples were collected and repeat HIV testing was conducted. Baseline demographic and behavioral data collected at enrollment were used for the primary analyses, while follow-up HIV testing results were used to ascertain HIV seroconversion during the follow-up period. The study defined a follow-up window period as one month before and after the scheduled follow-up date. The primary study endpoint was HIV seroconversion. New HIV infection was defined as a confirmed positive HIV test result following a previously negative result, both recorded at the Chongqing CDC. Participants who did not seroconvert were treated as right-censored at the date of their last HIV-negative test or at the end of the study period (December 31, 2022), whichever came first. Accordingly, a censoring indicator variable was defined (event = 1 for HIV seroconversion; censor = 0 for no seroconversion).The estimated date of seroconversion is defined as the midpoint between the last negative and the first confirmed positive enzyme-linked immunosorbent assay (ELISA) result, a commonly used approach in cohort studies when the exact infection time is unknown and assumed to occur uniformly within the testing interval.

Measures

The survey questionnaire was based on the core questionnaire used in the comprehensive HIV prevention and control pilot project for MSM populations in China, with modifications tailored to the specific context of Chongqing. The specific content is as follows. The questionnaire covered four domains. The first domain covered sociodemographic information, including age, marital status, household registration, ethnicity, education level, sexual orientation, anal sex role (bottom/versatile/top, as self-identified and self-reported by the participants), methods for seeking male partners (seeking male partners offline defined as meeting partners through physical venues, e.g., bars, bathhouses, or parks), and duration of residence in the area. The second domain assessed participants' knowledge of HIV prevention and control. Participants answered 8 questions related to HIV based on their level of knowledge. According to the standard, participants who correctly answered 6 or more of the 8 questions were classified as knowledgeable about HIV prevention and control [21]. The third domain covered sexual behavior during the 6 months prior to enrollment, including whether they engaged in unprotected anal intercourse with men (defined as never or rarely using condoms during anal sex with all male partners), whether they had a regular male partner, whether they had a casual male partner, whether they engaged in commercial male-to-male sex, whether they had sexual relations with females, whether they used Rush poppers (refer to inhaled alkyl nitrites, a class of recreational drugs that produce a transient vasodilatory effect and a brief sensation of euphoria), and whether they used drugs. The fourth domain covered STIs and HIV-related services, including whether the participant was diagnosed with any STIs in the past year, and whether they received HIV prevention services such as condom promotion and HIV counseling/testing. Biological indicators included HIV antibody testing and syphilis testing. Separately, biological indicators derived from laboratory blood tests included HIV antibody testing and syphilis serological testing (indicating past or current syphilis infection).

HIV and syphilis testing

For all MSM participants, 5 mL of venous blood was collected. During collection and storage, an appropriate amount of EDTA-K₂ anticoagulant was added to prevent coagulation. Blood samples were tested for HIV-1/2 antibodies and Treponema pallidum antibodies (syphilis). HIV testing was conducted in strict accordance with the National HIV Testing Technical Guidelines (2015 edition). Initial screening for HIV-1/2 antibodies was performed using a double-antibody sandwich method diagnostic reagent (Bole Life Sciences, Shanghai, China). Confirmatory testing was conducted using the HIV Blot 2.2 Western blot assay (MP Biomedicals Asia Pacific Pte Ltd, Singapore). For syphilis testing, initial screening was conducted using a TP-Ab (Treponema pallidum antibody) ELISA (Enzyme-Linked immunosorbent assay) kit (Shanghai Yaji Biotechnology Co., Ltd., China). For samples with positive results, the TRUST (Toluidine Red Unheated Serum Test; Shanghai Rongsheng Biopharmaceutical Co., Ltd., China) and the RPR (Rapid Plasma Reagin) test (Shanghai Kehua Bioengineering Co., Ltd., China) were performed. A diagnosis of active syphilis was made when both the TRUST and RPR tests yielded positive results.

Ethical approval

The study protocol was approved by the Institutional Review Board (IRB) of the Chongqing CDC (CQCDCLS(2019)002). Before the interviews and blood collection, each participant provided written informed consent. Participants were free to refuse participation or withdraw from the study at any time without consequence.

Data analysis

The database was established using EpiData 3.1 software, with data entry performed via a dual-entry method. Following data entry, cleaning was conducted using Excel worksheets, including logical checks and follow-up for missing data. After reassigning values and generating new variables, the incidence density was calculated based on the number of seroconversions during the study period and the total person-years of follow-up. The formula for calculating the incidence density was as follows: HIV incidence density = Total number of new infections / Total observation time. The total observation time was the sum of the observation time for both new infection cases and those without new infections, expressed as “/100 PYs.” Data analysis was conducted using SPSS 26.0. Continuous variables that followed a skewed distribution were described using the median. Continuous variables with skewed distributions were presented as medians with interquartile ranges (IQR). Categorical variables were presented as proportions or composition ratios. Univariable Cox proportional hazards regression was performed to assess the association between each variable and HIV seroconversion. All variables were treated as candidate covariates and entered into a multivariable Cox proportional hazards regression model, with forward stepwise selection (based on the likelihood ratio test, with entry and removal criteria set at P = 0.05 and P = 0.10, respectively) used to identify independent factors. This approach was chosen to objectively identify the core independent risk factors for HIV incidence while controlling for potential confounding, thereby constructing a parsimonious predictive model. Furthermore, collinearity diagnostics were performed on the final model using the Variance Inflation Factor (VIF), with a VIF < 5 established as the criterion for indicating variable independence and the absence of severe multicollinearity. To prevent the omission of crucial epidemiological confounders by the stepwise algorithm, a sensitivity analysis was conducted. Key demographic variables identified a priori in the literature (such as marital status and household registration) were forced into the final multivariable Cox model using the “Enter” method to fully adjust for potential confounding effects and verify the robustness of the core predictors. A post-hoc power analysis was planned and conducted to evaluate the statistical adequacy of the final sample size for detecting the primary observed effect sizes. Statistical significance was defined as α = 0.05.

Results

Trends of HIV prevalence

Between 2013 and 2022, a total of 2,134 HIV-negative MSM were identified, all of whom had undergone at least two HIV tests (Fig. 1). Out of these, 247 cases of HIV seroconversion were detected, with a total of 5,024.3 person-years (PYs) of observation. The median follow-up time was 1.8 PYs. The overall HIV incidence density was 4.9 (95% CI: 4.3–5.6) per 100 PYs. Across the study period, the number of newly enrolled participants gradually declined, from 642 in 2013–2014 to 86 in 2021–2022, while the number of participants exiting the cohort increased over time. The number of participants under observation peaked in 2017–2018 (n = 1,951) and then decreased in the later periods. The incidence densities for the periods 2013–2014, 2015–2016, 2017–2018, 2019–2020, and 2021–2022 were 11.4 (95% CI: 8.7–14.7), 6.0 (95% CI: 4.7–7.5), 4.5 (95% CI: 3.5–5.7), 3.1 (95% CI: 2.3–4.2), and 2.8 (95% CI: 1.7–4.7) per 100 PYs, respectively (Table 1). Despite these fluctuations in follow-up time and sample size, the incidence density was calculated based on person-time, which accounts for differences in observation time across periods. Since 2013, there has been a declining trend in HIV incidence density among MSM (Fig. 2). A post-hoc power analysis indicated that this final sample size (N = 2,134, with 247 seroconversion events) provided 91.1% power to detect the primary risk factor (e.g., unprotected anal intercourse, HR = 1.53).

Fig. 1.

Retrospective cohort study screening process, from 2013 to 2022

Table 1.

Incidence of new HIV incident infection among MSM in Chongqing

Year	Newly enrolled participants	Number of participants exiting the cohort	Number of participants under observation	Number of seroconversions	PYs of observation	Incidence density(/100PYs)
2013–2014	642	99	909	53	465.9	11.4(8.7–14.7)
2015–2016	545	281	1652	71	1189.0	6.0(4.7–7.5)
2017–2018	510	365	1951	66	1472.7	4.5(3.5–5.7)
2019–2020	351	508	1802	42	1370.0	3.1(2.3–4.2)
2021–2022	86	634	844	15	526.7	2.8(1.7–4.7)
Total	2134	-	-	247	5024.3	4.9(4.3–5.6)

Fig. 2.

Trends in the HIV incidence density among MSM in Chongqing, China, from 2013 to 2022

Demographic characteristics

Among the 2,134 HIV-negative MSM in this open cohort study, the median age was 26 (IQR 22–31) years. A total of 1,874 MSM (87.8%) were not married at the time of cohort entry. Approximately 71.8% of participants had at least a college education. In terms of sexual roles, 627 MSM (29.4%) identified as bottom, 734 MSM (34.4%) as versatile, and 773 MSM (36.2%) as top. Additionally, 585 MSM (27.4%) reported seeking male sexual partners offline (Table 2).

Table 2.

Baseline characteristics of participants from 2013 to 2022 [n (%)]

Characteristics	Group	N(%)	Characteristics	Group	N(%)
Marital Status	Unmarried	1774(83.1)	Sexual orientation	Homosexual	1574(73.8)
	Married/cohabiting	260(12.2)		Bisexual	478(22.4)
	divorced/widowed	100(4.7)		Heterosexual/others	82(3.8)
Household registration	Chongqing	1690(79.2)	Anal sex role	Bottom	627(29.4)
	Other city	444(20.8)		Versatile	734(34.4)
Ethnicity	Han	2093(98.1)		Top	773(36.2)
	Other ethnic groups	41(1.9)	Ways of seeking male sexual partners	Internet/dating apps	1549(72.6)
Education	Unschooled/primary school	26(1.2)		Offline	585(27.4)
	Junior high school	136(6.4)	Local residence duration	Within 2 years	138(6.5)
	Senior high school	439(20.6)		Over 2 years	1996(93.5)
	Junior college/university	1533(71.8)

Sexual behaviors during the 6 months prior to enrollment

The age range for first male-to-male sexual activity was 10 to 74 years, with a median age of 20 (IQR 18–24) years. A total of 936 MSM (43.9%) reported engaging in unprotected anal intercourse (UAI) with a male partner during the 6 months prior to enrollment. Moreover, 78 MSM (3.7%) had more than two regular male partners, while 263 MSM (12.3%) had more than two male casual partners. Fifty-two MSM (2.4%) engaged in commercial sex with a male partner during the 6 months prior to enrollment. Additionally, 262 MSM (12.3%) reported having sex with a female partner. Furthermore, 235 MSM (11.0%) used rush poppers, and 28 MSM (1.3%) used drugs during the 6 months prior to enrollment (Table 3).

Table 3.

Status of sexual behavior, HIV prevention knowledge, services, and STIs history [n (%)]

Characteristics	N (%)
Having UAI during the 6 months prior to enrollment	936(43.9)
Number of regular male sexual partners during the 6 months prior to enrollment
≤ 2	2056(96.3)
> 2	78(3.7)
Number of casual male sexual partners during the 6 months prior to enrollment
≤ 2	1871(87.7)
> 2	263(12.3)
Having commercial sex during the 6 months prior to enrollment	52(2.4)
Having heterosexual sex during the 6 months prior to enrollment	262(12.3)
Use of rush poppers during the 6 months prior to enrollment	235(11.0)
used drugs during the 6 months prior to enrollment	28(1.3)
Knowledge of HIV Prevention and Control
Is AIDS an incurable, serious infectious disease?	1939(90.9)
Are MSM the population most severely affected by HIV in China?	1842(86.3)
Whether someone is infected with HIV based on their appearance?	1937(90.8)
Does having other STIs increase the risk of HIV infection?	1848(86.6)
Does consistent and correct condom use reduce the risk of HIV infection and transmission?	2097(98.3)
Does the use of new drugs increase the risk of HIV infection?	1880(88.1)
Should one proactively seek HIV testing and counseling after engaging in high-risk behavior?	2095(98.2)
Should individuals who intentionally spread HIV bear legal responsibility?	1892(88.7)
Knowledgeable about HIV prevention and control	2000(93.7)
Diagnosed with an STI in the past year	75(3.5)
Self-reported the type of STI was diagnosed in the past year
Gonorrhea	8(10.7)
Syphilis	19(25.3)
Genital warts	27(36.0)
Genital herpes	4(5.3)
Others	17(22.7)
Laboratory-confirmed syphilis seropositivity	80(3.8)
Received condom promotion / HIV counseling and testing services in the past year	696(32.6)
Received peer education services in the past year	707(33.1)
Received HIV testing in the past year	970(45.5)

Abbreviation: UAI: unprotected anal intercourse

HIV-Related knowledge and service, and history of STIs

A total of 93.7% of MSM demonstrated satisfactory knowledge of HIV prevention and control (Table 2). Seventy-five MSM (3.5%) reported being diagnosed with a sexually transmitted disease (STD) in the past year, with 36.0% diagnosed with genital warts, 25.3% with syphilis, followed by gonorrhea (10.7%), genital herpes (5.3%), and other STDs (22.7%). Based on biological testing, eighty MSM (3.8%) were laboratory-confirmed as seropositive for syphilis. Approximately 32.6% of participants had received condom education and/or HIV counseling and testing services in the past year. Seven hundred seven MSM (33.1%) had received peer education services in the past year, and 970 MSM (45.5%) had undergone HIV testing in the past year (Table 3).

Factors associated with the prevalence of HIV

The univariate analysis (Table 4) indicates that MSM aged 16 to 18 years have a significantly higher likelihood of HIV seroconversion (HR 2.08, 95% CI: 1.09–3.98) compared to those aged 40 years or older. Additionally, MSM with a high school education or lower are more likely to contract HIV (HR 1.63, 95% CI: 1.26–2.10) than those with a junior college education or higher. MSM with a versatile anal sex role have a higher likelihood of HIV infection compared to those with a top anal sex role (HR 1.36, 95% CI: 1.01–1.81). Seeking male partners offline (HR 1.67, 95% CI: 1.29–2.17) is a significant factor contributing to the increased likelihood of HIV seroconversion. Moreover, having UAI with a male partner in the past six months (HR 1.57, 95% CI: 1.22–2.01), engaging in commercial sex (HR 2.12, 95% CI: 1.09–4.13), and not having had an HIV test in the past year (HR 1.40, 95% CI: 1.08–1.81) are also associated with a higher risk of HIV infection.

Table 4.

Univariate analysis of HIV infection-related factors among MSM in Chongqing, China (2013–2022)

Variables	HIV conversions	PYs	Incidence density (/100 PYs)	P	HR(95%CI)
Age group (years)				0.040
< 18	14	134.0	10.4	0.026	2.08(1.09–3.98)
18–40	206	4283.1	4.8	0.843	1.04(0.70–1.56)
> 40	27	607.2	4.4		1.00(Ref)
Marital Status
Unmarried/divorced/widowed	221	4394.3	5.0	0.379	1.20(0.80–1.80)
Married/cohabiting	26	630.0	4.1		1.00(Ref)
Household registration
Chongqing	212	4144.1	5.1	0.101	1.35(0.94–1.93)
Other city	35	880.2	4.0		1.00(Ref)
Ethnicity
Han	245	4944.8	5.0	0.325	2.01(0.50–8.09)
Other ethnic groups	2	79.5	2.5		1.00(Ref)
Education
High school and below	95	1391.6	6.8	< 0.001	1.63(1.26–2.10)
Junior college and above	152	3632.7	4.2		1.00(Ref)
Sexual orientation
Homosexual	176	3685.2	4.8	0.409	0.89(0.68–1.17)
Others	71	1339.1	5.3		1.00(Ref)
Anal sex role				0.039
Bottom	62	1451.9	4.3	0.768	0.95(0.68–1.33)
Versatile	104	1729.9	6.0	0.040	1.36(1.01–1.81)
Top	81	1842.5	4.4		1.00(Ref)
Ways of seeking male sexual partners
Internet/dating apps	147	3530.5	4.2		1.00(Ref)
Offline	100	1493.8	6.7	< 0.001	1.73(1.34–2.23)
Local residence duration
Within 2 years	15	290.9	5.2		1.00(Ref)
Over 2 years	232	4733.4	4.9	0.939	0.98(0.58–1.65)
Overall knowledge status
Knowledgeable	229	248.0	4.8		1.00(Ref)
Unaware	18	4776.3	7.3	0.167	0.71(0.44–1.15)
Age of first homosexual intercourse (years)
≤ 18	80	1391.5	5.7	0.187	1.20(0.92–1.56)
> 18	167	3632.8	4.6		1.00(Ref)
Having UAI during the 6 months prior to enrollment
Yes	135	2195.0	6.2	< 0.001	1.57(1.22–2.01)
No	112	2829.3	4.0		1.00(Ref)
Number of regular male sexual partners during the 6 months prior to enrollment
≤ 2	238	4845.6	4.9		1.00(Ref)
> 2	9	178.7	5.0	0.869	1.06(0.54–2.06)
Number of casual male sexual partners during the 6 months prior to enrollment
≤ 2	206	4342.1	4.7		1.00(Ref)
> 2	41	682.2	6.0	0.115	1.31(0.94–1.83)
Having commercial sex during the 6 months prior to enrollment
Yes	9	83.8	10.7	0.027	2.12(1.09–4.13)
No	238	4940.5	4.8		1.00(Ref)
Having heterosexual sex during the 6 months prior to enrollment
Yes	35	602.7	5.8	0.312	1.20(0.84–1.72)
No	212	4421.6	4.8		1.00(Ref)
Use of rush poppers during the 6 months prior to enrollment
Yes	33	463.7	7.1	0.080	1.39(0.96–2.01)
No	214	4560.6	4.7		1.00(Ref)
Drug use during the 6 months prior to enrollment
Yes	3	31.4	9.6	0.313	1.80(0.58–5.62)
No	244	4992.9	4.9		1.00(Ref)
Diagnosed with an STI in the past year
Yes	10	228.1	4.4	0.891	0.96(0.51–1.80)
No	237	4796.2	4.9		1.00(Ref)
Laboratory–confirmed syphilis seropositivity
Yes	10	177.7	5.6	0.761	1.10(0.59–2.08)
No	237	4846.6	4.9		1.00(Ref)
Received condom promotion / HIV counseling and testing services in the past year
Yes	84	1973.0	4.3		1.00(Ref)
No	163	3051.3	5.3	0.304	1.15(0.88–1.50)
Received peer education services in the past year
Yes	85	1817.5	4.7		1.00(Ref)
No	162	3206.8	5.1	0.815	1.03(0.79–1.34)
Received HIV testing in the past year
Yes	151	2398.6	6.3		1.00(Ref)
No	96	2625.7	3.7	0.010	1.40(1.08–1.81)

UAI: unprotected anal intercourse

All variables from the univariate analysis were used as independent variables (X), and a forward stepwise selection method was employed for the multivariate Cox proportional hazards model. The results of the multivariate analysis (Table 5) showed that MSM with a high school education or below had a higher risk of HIV infection compared to those with a junior college education or higher (HR 1.57, 95% CI: 1.21–2.03). MSM with a versatile anal sex role are more likely to contract HIV compared to those with a top anal sex role (HR 1.38, 95% CI: 1.03–1.85). Seeking male partners offline (HR 1.73, 95% CI: 1.34–2.23), having UAI with a male partner in the past six months (HR 1.53, 95% CI: 1.19–1.97), and using rush poppers in the past six months (HR 1.60, 95% CI: 1.09–2.33) were also identified as significant risk factors for HIV infection in MSM. Collinearity diagnostics revealed that the VIF values for all variables included in the final multivariable Cox model ranged from 1.00 to 1.10, indicating that the risk factors in the model are mutually independent and free from severe multicollinearity. In the sensitivity analysis, after forcibly adjusting for a priori theoretical confounders (marital status and household registration), the primary behavioral risk factors remained robust and statistically significant (all P < 0.05), with no substantial changes in their respective Hazard Ratios.

Table 5.

Multivariate analysis of HIV infection-related factors among MSM in Chongqing, China (2013–2022)

Variables	B	S.E.	Waldχ2	P	HR (95%CI)
High school education or below	0.45	0.13	11.67	0.001	1.57(1.21–2.03)
Anal sex role			6.91	0.032
Top					1.00(Ref)
Bottom	-0.04	0.17	0.05	0.819	0.96(0.69–1.34)
Versatile	0.32	0.15	4.68	0.030	1.38(1.03–1.85)
Finding sexual partners offline	0.59	0.13	19.62	< 0.001	1.81(1.39–2.35)
Having UAI during the 6 months prior to enrollment	0.43	0.13	11.03	0.001	1.53(1.19–1.97)
Use of rush poppers	0.47	0.19	5.89	0.015	1.60(1.09–2.33)

UAI: unprotected anal intercourse

Discussion

This study found that from 2013 to 2022, the HIV incidence density among MSM in Chongqing, China, demonstrated a clear downward trend. However, when compared to the national average incidence density in China [22], the overall HIV incidence density in Chongqing remains relatively high, suggesting that MSM in this region continue to face a substantial risk of HIV infection. This finding underscores the importance of prioritizing MSM in targeted HIV prevention and intervention programs.

Consistent with previous research, high rates of new HIV infections among MSM have been widely reported. For example, the CDC’s HIV Surveillance Report in the United States indicated that 65% of the 1.23 million individuals living with HIV were infected through male-to-male sexual contact [23]. In the Philippines, a study similarly based on convenience sampling and snowball recruitment at VCT clinics reported the incidence density among MSM from 2014 to 2018 was 2.7/100 PYs [24]. Within China, a cohort study in Chengdu, which recruited participants through non-governmental organizations (NGOs) using convenience and snowball sampling, reported the incidence density among MSM in Chengdu from 2012 to 2018 was 5.95 (95% CI: 5.37–6.56)/100 PYs [10], which is higher than that observed in the present study for Chongqing. In contrast, the incidence density in Tianjin from 2013 to 2022 was reported as 2.3 (95% CI: 2.0–2.7)/100 PYs [11]. Notably, the Tianjin study utilized the same VCT clinic-based convenience and snowball sampling approaches as our study, yet reported a lower rate than Chongqing. These discrepancies may be partly attributable to differences in study design, sampling strategies, and recruitment approaches across studies, which can influence the composition of MSM populations and their underlying risk profiles. Nevertheless, both the Tianjin and Chengdu studies reported similar downward trends in incidence, aligning with the findings of this study.

The continuous decline in HIV incidence density observed in this study may be associated with the implementation of HIV prevention strategies in Chongqing. In recent years, the Chongqing CDC has launched a series of comprehensive interventions targeting MSM, addressing educational, behavioral, biomedical, and structural factors. These initiatives include strengthening screening and intervention services for key populations, expanding access to post-exposure prophylaxis (PEP), promoting condom use, and improving HIV testing coverage across healthcare settings. Aligned with the goals of the “13th Five-Year Plan for HIV/AIDS Prevention and Control in Chongqing” [25], the city has expanded HIV testing services and promoted VCT to facilitate early diagnosis [26]. Individuals diagnosed with HIV are linked to antiretroviral therapy (ART) to support early treatment initiation [27, 28]. In addition, pre-exposure prophylaxis (PrEP) has been introduced as part of biomedical prevention strategies among MSM [29]. PrEP has been shown to effectively reduce the risk of HIV acquisition [30–32]. Studies suggest that even in settings where PrEP coverage is suboptimal or associated with out-of-pocket costs, it may be acceptable among MSM; however, barriers such as high cost, limited accessibility, and stigma in clinical settings remain important challenges to its wider implementation in China [33]. Furthermore, community-based interventions led by NGOs, including outreach in offline venues (e.g., bars, clubs, and bathhouses) and online health promotion campaigns, may also have contributed to behavioral changes among MSM [34]. However, given the observational nature of this study and the lack of direct measures of individual exposure to these interventions, causal inferences cannot be established. The observed decline in HIV incidence should therefore be interpreted with caution, and future studies are needed to evaluate the effectiveness of specific prevention programs.

Findings from the Cox proportional hazards model highlight several key risk factors for HIV infection among MSM. Lower educational attainment and seeking male sexual partners through offline venues were significantly associated with higher HIV risk. This association may be related to broader socioeconomic and behavioral factors, such as differential access to prevention resources and patterns of partner-seeking or venue attendance [35]. Notably, HIV-related knowledge was not significantly associated with seroconversion in this study, suggesting that factors beyond knowledge alone may play a more important role. Engaging in UAI remains a major risk factor for HIV transmission, consistent with findings from other studies [10, 36, 37]. The rectal mucosa is highly susceptible to tearing or bleeding during intercourse, especially in the absence of condom use, facilitating the transmission of HIV [38]. Thus, UAI is widely recognized as a primary driver of HIV infection among MSM [39], emphasizing the continued need to promote and normalize condom use within this population [40, 41]. Additionally, MSM who identified their anal sex role as versatile were found to have a higher risk of HIV infection compared to the top. This is in line with prior research [42, 43], which has noted that versatile MSM may engage more frequently in UAI or have a greater number of casual partners [44–46]. The dual-role behavior places them at risk for both acquiring and transmitting HIV [47]. In contrast, those identifying as the top do not require the sphincter-relaxing effects of nitrite inhalants during intercourse, likely contributing to their lower prevalence of use [48]. Additionally, previous studies have established nitrite inhalant use during anal intercourse as a risk factor for HIV infection among MSM [49]. These findings suggest the necessity of designing more targeted prevention strategies informed by sexual role–related risk patterns, in addition to general population-level interventions, while avoiding stigmatizing specific subgroups. The use of rush poppers also emerged as a significant risk factor. Poppers have become the most commonly used recreational drug among Chinese MSM [50], and users often tend to be younger, engaging in high-risk behaviors [49, 51]. Therefore, intensifying health education, raising awareness of the risks associated with drug use, and reinforcing condom use promotion remain crucial for reducing HIV incidence in this population. These findings have important implications for targeted HIV prevention. Interventions should prioritize MSM with higher-risk behavioral profiles, such as those engaging in versatile sexual roles or seeking partners through offline venues. Tailored strategies may include venue-based outreach in bars, bathhouses, and other physical settings, as well as targeted health education and condom distribution programs. In addition, given the association between rush popper use and HIV risk, integrating substance use education and harm reduction strategies into existing HIV prevention programs may further reduce risk behaviors. Importantly, such interventions should be designed to avoid stigmatization and ensure accessibility across diverse MSM subgroups.

This study has several limitations. First, the use of a convenience sampling strategy, combined with recruitment through multiple channels (VCT clinics, outreach, and peer referral), may have introduced selection bias, as these approaches are more likely to capture individuals with higher health awareness and distinct health-seeking behaviors, thereby limiting the representativeness and generalizability of the findings. This is further reflected in the relatively high proportion of participants with college education or above. Therefore, the findings may not be fully generalizable to the broader MSM population in Chongqing, particularly those who are less engaged with HIV testing and prevention services. Second, although the overall sample size was substantial, the analysis was restricted to individuals with at least two HIV testing records at the CDC during the study period, which may have excluded other high-risk individuals and skewed the sample toward more health-conscious participants. In addition, MSM not engaged with CDC-based testing services were not represented and may have different behavioral and epidemiological characteristics. Furthermore, the declining cohort size in later follow-up years may have reduced the stability of incidence estimates over time, particularly in the later study period. Therefore, caution is warranted when interpreting incidence fluctuations observed in the later years of the cohort. Third, STI diagnoses and behavioral data were based on self-reported questionnaires involving sensitive topics, and are therefore subject to underreporting, recall bias, and social desirability bias, potentially affecting the accuracy of the findings. Finally, the study did not account for key social determinants, such as perceived stigma, social support, disclosure of sexual orientation, experiences of discrimination, and community connectedness, which may influence risk behaviors and engagement with health services. These limitations should be considered when interpreting the results. In addition, given the extended study period (2013–2022), temporal variations in socioeconomic factors, prevention programs, and social contexts may have influenced HIV incidence and associated risk factors. However, conducting stratified analyses by time period may reduce statistical power and was beyond the primary scope of this study. Future studies are needed to explore these temporal dynamics.

In conclusion, this study found that HIV incidence density among MSM in Chongqing showed a consistent decline from 2013 to 2022, although the overall level remains substantial. Several behavioral factors, including seeking male sexual partners through offline venues, engaging in unprotected anal intercourse, versatile sexual roles, and the use of rush poppers, were significantly associated with HIV seroconversion. These findings highlight the need for targeted and feasible interventions, including strengthening venue-based outreach, promoting consistent condom use, expanding access to HIV testing and PrEP, and integrating substance use risk reduction into prevention programs. Future efforts should focus on implementing context-specific and non-stigmatizing strategies to further reduce HIV transmission among MSM.

Abbreviations

HIV

Human immunodeficiency virus

AIDS

Acquired immunodeficiency syndrome

UAI

Unprotected heterosexual intercourse

MSM

Men who have sex with men

STIs

Sexually transmitted infections

VCT

Voluntary counseling and testing

CDC

Center for disease control and prevention

ELISA

Enzyme linked immunosorbent assay

PEP

Post-exposure prophylaxis

ART

Antiretroviral therapy

PrEP

Pre-exposure prophylaxis

Author contributions

TT and LJ drafted and revised the manuscript. LJ, LL and WZ analysed the data and interpreted the data. YH and CB conceptualised the study and review the manuscript. CZ, RL and GW conceptualised and designed the study, edited, revised and critically reviewed the manuscript. All authors have read and approved the manuscript, and ensure that this is the case.

Funding

This study was supported by Chongqing Medical Scientific Research Project(2026MSXM113) and Chongqing Natural Science Foundation Project (CSTB2025NSCQ-GPX0291).

Data availability

The data reported in this work is available upon request from the corresponding author.

Declarations

Ethics approval and consent to participate

The study protocol is approved by the Institutional Review Board (IRB) of the Chongqing CDC (CQCDCLS(2019)002). Before the interviews and blood collection, each participant provides written informed consent. The human data covered in this study adhere to the Helsinki Declaration.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.