We included 49 systematic reviews and meta-analyses of high and moderate quality, which included a total of 1,721 studies. Among the included reviews, 35 were global (two or more continents), five only in the United States, six only in China, one in Brazil, one in Africa, and one in Europe. Thirty-two reviews explicitly focused on GBMSM, seven extended the study population beyond GBMSM to include high-risk populations such as injection drug users, sex workers, transgender people, incarcerated people, and LGBT populations more broadly. The remaining ten reviews were broader but included GBMSM in the inclusion criteria for the review. The risk of bias assessments indicated that 35 reviews were of high quality and 14 were of moderate quality. Four reviews were low quality, and nine were critically low quality (of which their data are not presented here).

The overwhelming majority (68%) of the systematic review literature on HIV infection and positive serostatus examined individual-level factors associated with infection. Eight reviews examined interpersonal-related factors, and six examined structural-level factors. Thematic analyses of the results implicated three levels of the socioecological model in the system of HIV risk: individual-, interpersonal-, and structural-level. Within the individual level, we identified (1A) biomedical prevention methods that shape HIV infection; (1B) sexual and sex-seeking behaviors influence HIV infection; (1C) behavioral prevention methods shape HIV infection; and (1D) individual-level characteristics and infections influence HIV infection as central themes. Within the interpersonal level, we identified that (2A) lived experiences and interpersonal relationships influence HIV infection. At the structural level, we found three themes: (3A) country-level income influences HIV infection; (3B) country-level HIV prevalence shapes HIV infection; and (3C) structural stigma shapes HIV infection.

Identified themesIndividual-level of the socioecological model Theme 1A: Biomedical prevention methods shape HIV infection

There were four reviews, one in Europe and three globally that provided high-quality information that the use of pre-exposure prophylaxis (PrEP) reduces HIV infection, emphasizing that greater adherence leads to more protection [22,23,24, 25••]. Wang et al. reviewed 74 studies reporting that post-exposure prophylaxis (PEP) reduces HIV risk, reporting 2.6% seroconversions were observed among 19,456 GBMSM [26••]. However, there was a lack of data from low- and middle-income countries. Two high-quality reviews, one in Europe and the other focused on the United States, Australia, and Europe, indicated that the use of, and in particular, early-initiation of ARV treatment significantly reduces the HIV infection of sexual partners [25••, 26••]. A moderate-quality global review by Jiang et al. indicated that ARV reduces HIV infection risk [27••]. Overall, the review of systematic reviews demonstrates high-quality evidence that biomedical preventative behaviors have a dramatic impact on reducing HIV infection.

Theme 1B: Sexual and sex-seeking behaviors influences HIV infection

Sixteen reviews discussed how sexual and sex-seeking behaviors shape HIV infection. One high-quality global review indicated that transactional sex increases the risk of HIV infection (OR 1.3, 95% CI 1.1–1.6) among GBMSM [28••]. One high-quality review in China indicated that GBMSM who had sex in bathhouses, compared to other subgroups, had the highest prevalence of HIV (OR = 13.4, 95% CI 10.3–17.1, n = 22) [28••]. One high-quality review in China found that GBMSM who had sex with women, as compared to men only, had an elevated prevalence of HIV (6.6% vs. 5.4% 95% CI 1.01–1.6) [29]. A review by Li 2019 found that self-reported rectal douching was associated with elevated odds of living with HIV (OR 2.8, 95% CI 2.32 to 3.39) [30••]. Several high-quality reviews and metanalyses validated the well-documented association between condomless sex and increased HIV infection [26••, 30••, 31, 31, 33••]. Four reviews six (three high and one moderate) analyzed how behavioral health factors, specifically substance use, influence HIV risk [28••, 34••, 35, 36••]. Reviews investigated a variety of drugs. Three reviews found a strong association between substance use and elevated HIV infection [34••, 35, 36••], while one in China did not find a statistically significant association [28••].

Theme 1C: Behavioral prevention methods can shape HIV infection

Eleven reviews examined how individual-level behaviors shape HIV infection. Four reviews (three high quality and one moderate) indicated that serosorting is less effective at HIV prevention, as compared to condoms. Still, serosorting is consistently associated with a lower risk of HIV when compared to condomless anal intercourse with no serosorting [25••, 32, 37, 38]. Eight reviews (seven high quality and one moderate quality) identified that insertive sexual positions are more protective against HIV than receptive only or receptive and insertive sexual roles [25••, 26••, 39,40,41,42,43,44]. Lastly, eight reviews (one moderate and seven high quality) consistently showed that condom use during anal intercourse prevents HIV transmission [25••, 26••, 28••, 30••, 31, 33••, 37, 45].

Theme 1D: Individual-level characteristics and infections influence HIV infection

Fourteen systematic reviews explored associations between HIV infection and individual-level characteristics such as socioeconomic status, education level, age, circumcision status, and syndemic infections. One moderate quality review of 47 studies from 17 countries across Latin America and the Caribbean revealed that low socioeconomic status was associated with elevated HIV prevalence [30••]. Two high-quality reviews indicated lower education levels were associated with higher HIV prevalence [28••, 33••]. Associations between age and HIV infection were mixed. Two high-quality reviews indicated that younger age was associated with more significant HIV infection [32, 34••]. In contrast, one moderate quality review showed older aged MSM had higher rates of HIV [30••]. The association between circumcision status and HIV risk has been studied extensively with mixed results. Six systematic reviews examined circumcisions and HIV, of which four reviews of three high quality and one moderate quality found no significant relationship with reduced HIV infection [25••, 40,41,42]. Two other high-quality reviews found some protection against HIV with male circumcision but mostly in low-middle-income countries and among insertive partners [44, 45]. The two reviews suggest that mixed results may be due to variations in study design, geography, or the country’s economic status. For example, 29 of 33 studies in Zhang et al.’s 2019 review found no significant association between circumcision and HIV infection. However, subgroup analysis revealed a protective association was found among cross-sectional studies (OR, 0.92; 95% CI, 0.87–0.98) compared to the cohort studies, which found nonsignificant associations (OR, 1.01; 95% CI, 0.86–1.19) [45]. Global data from 62 observational studies involving a total of 119, 248 MSM revealed that circumcision was protective against HIV infection among MSM in LMICs (0·58, 0·41–0·83; k = 23; I2 = 77%) but not among MSM in high-income countries (0·99, 0·90–1·09; k = 20; I2 = 40%) [44]. Lastly, HIV infection was associated with other sexually transmitted infections. Seven reviews, all high quality, indicated that the presence of other STIs, covering syphilis, gonorrhea, chlamydia, HPV, HBV, and HSV-2, was associated with higher rates of HIV infection [31, 33••, 45, 46••]. The thematic analysis of the individual-level factors reifies the complex nature of HIV risk. The body of research on individual-level factors portrays the complexity of factors that all can serve to increase risk of HIV infection.

Interpersonal-level of the socioecological model Theme 2A: Lived experiences and interpersonal relationships influence HIV infection

Eight systematic reviews examined how an individual’s lived experience and interpersonal relationships might shape their risk of HIV. A 2018 moderate-quality review among incarcerated GBMSM from 24 middle- and high-income countries revealed that incarcerated GBMSM are at five times higher risk of HIV infection than incarcerated men who do not have sex with men [47]. However, the risk among incarcerated GBMSM varied by geography such that HIV prevalence among incarcerated GBMSM is 10 times higher in Latin America and 20 times higher in Western Europe [47]. Another high-quality review indicated that the continuity of HIV services, such as testing and testing pre-and-post release could reduce HIV cases among Black GBMSM who experience incarceration and reenter their communities post-release [48].

Three systematic reviews (two moderate and one high quality) examined the relationship between HIV infection and experiences of abuse (childhood and intimate partner violence), all indicating that GBMSM who experienced abuse had elevated HIV infections [30••, 48, 49]. A moderate 2012 systematic review of 12 papers revealed higher HIV infection among men with a history of childhood sexual abuse as compared with men with no history [OR = 1.54; 95% CI (1.22–1.95)] [50••]. Two reviews (Coelho and Buller) found positive associations between violent relationships and elevated HIV infection [30••, 48]. Buller et al.’s global review of data from 8,835 MSM found that exposure to intimate partner violence was associated with a 1.5 times greater chance of a positive HIV status [AOR = 1.5, 95% CI (1.3, 1.7)] [49].

One moderate-quality review examined how GBMSM’s experiences of interpersonal homophobia and homophobic abuse shaped HIV diagnoses. The 2021 rapid review and meta-analysis of 121 cross-sectional studies found an elevated risk of diagnosed HIV among GBMSM in the US who experienced homophobia (OR = 1.34, 95% CI = 1.10–1.64, I2 = 86.3%) [50••]. Stratified analysis showed different levels of HIV infection by race and ethnic group, with Latinos experiencing the greatest.

Lastly, a review examined motivational interviewing as an intervention to support the prevention of HIV among MSM. However, the results concluded that evidence of motivational interviewing’s effectiveness was lacking [51]. The various interpersonal-level factors that shape HIV infection serve to further complicate conceptualizations of HIV “risk.”

Structural-level of the socioecological model Theme 3A: Country-level income influences HIV infection

Two review papers (moderate and high quality) discussed how country-level socioeconomic factors influence HIV prevalence. The moderate quality review by Baral et al. indicated a lower odds ratio for HIV infection among GBMSM in low-income countries (OR = 7.8, 95% CI 7.2, 8.4), as compared to GBMSM in middle-income countries (OR = 23.5, 95% CI 22.8, 24.0) [52••]. A review by Blondeel et al. indicated that MSM in low-middle income countries had a 19.3 increased chance of HIV infection, as compared to the general population [61i].

Theme 3B: Country-level prevalence influences HIV infection

In the same two reviews, GBMSM in very low prevalence countries had the highest OR of HIV infection [OR = 58.4, 95% CI (56.3, 60.6)], as compared to low prevalence [OR = 14.4, 95% CI (13.8, 14.9)], and high prevalence countries [OR = 9.6, 95% CI (8.9, 10.2)] [52••]. Another review showed geographic variation in HIV burden, with the highest prevalence of HIV among GBMSM found in Caribbean countries and sub-Saharan Africa, and lowest in countries within the Middle East and North Africa [52••].

Theme 3C: Structural stigma shapes HIV infection

Two high-quality reviews examined the role of structural stigma, including how homophobic and criminalizing policies elevate HIV infection [53••, 54••]. Oldenbrug et al.’s review indicated an 11% lower HIV prevalence in countries with protective laws for MSM, compared to countries lacking legal protections [53••]. Countries with protective language for men who sell sex had a 7% lower prevalence of HIV, as compared to those without such protections [54••]. A systematic review and meta-analysis of 75 papers found that GBMSM were more likely to be aware of their HIV status in countries with less repressive legislation (22% vs. 6.7%) and less severe penalties for same-sex relations [55]. These findings indicate that structural factors could influence HIV infection and complicate our understanding of HIV risk further.

Dynamic and emergent nature of HIV risk

In totality, the amalgamation of our thematic analyses and findings from our systematic reviews of reviews suggests that the risk of HIV infection operates in an emergent, dynamic, and complex nature across multiple levels of the socioecological model that interact with one another to elevate GBMSMS’s risk of HIV further. Structural factors such as stigmatizing policies, macroeconomic factors, and population-level epidemiology can shape environments in ways that produce HIV infection—moreover, interpersonal factors such as experiences of abuse and homophobia further shape HIV infection among GBMSM. Lastly, numerous individual-level factors such as sexual behaviors, use of biomedical prevention methods, and sexual relationships, amongst others, further perpetuate the risk of HIV infection among GBMSM. This systematic review of reviews and meta-analyses and the developed visualization indicated that HIV risk among GBMSM worldwide is socially patterned by numerous interrelated factors, such that the ecosystem itself is the driver of the disproportionate risk of HIV that burdens GBMSM (Fig. 2).

Fig. 2

Complex systems visualization of HIV risk among GBMSM, globally. Legend: Yellow: individual-level. Blue: interpersonal-level. Green: community-level. Purple: institutional-level. Red: structural-level