Introduction

In the United States, young gay, bisexual, and other men who have sex with men (MSM) have substantially higher HIV incidence rates than older MSM. (Studies vary in their use of age to designate men as young or older. In this article, we consider MSM aged ≤ 29 years to be young and MSM aged ≥ 30 years to be older.) A recent multiracial study found that HIV incidence among MSM aged 18–24 years was 2.5 times that of older MSM [1]. Age-related incidence disparities are particularly concerning for black MSM. Researchers recently reported that HIV incidence among black MSM aged 18–30 years was 4.3 times that of black MSM aged ≥ 31 years [2]. Such high incidence rates could result in most black MSM acquiring HIV infection by age 35 years [3]. Irrespective of age-related disparities among MSM, HIV incidence among young MSM of all races is high. A 21-city study reported a 2.9% incidence density among MSM aged 18–24 years [4], and another study reported a 24-month cumulative incidence rate of 7.3% among very young MSM (aged 16–20 years) [5]. Due to such high HIV incidence, MSM aged 13–29 years accounted for 48.5% of new HIV diagnoses among all MSM in 2016 [6]. Better understanding of the determinants of age-associated HIV incidence disparities might promote improved prevention efforts for all young MSM, especially black young MSM.

Researchers have identified some factors that contribute to these disparities. First, as HIV prevalence increases with age, HIV incidence decreases. This results in the pool of MSM who are HIV negative, but still at risk for HIV infection, being large among young MSM compared with older MSM [7]. Another potential explanation is that young MSM have an increased likelihood for condomless anal intercourse (CAI) [8], but studies have produced mixed results regarding age differences in CAI. Third, because of the elevated HIV prevalence among older MSM [9], young MSM are at high risk for acquiring HIV from older MSM due to some age-disassortative mixing in the sexual networks of MSM [10]. Although these explanations are invaluable, they focus solely on individual-level behaviors and sexual network characteristics. Contemporary epidemiologic studies emphasize the importance of also examining contextual factors that promote HIV infection [11]. Moreover, because MSM sub-groups that are at greatest risk for HIV do not always engage in greater levels of risk behaviors than other sub-groups, it is important to examine structural factors [12, 13].

Therefore, we sought to identify additional factors that might explain age-associated HIV incidence disparities among MSM. Social ecological theory informed our approach. This framework draws attention to behavioral, clinical, psychosocial, and structural factors that promote health disparities [14]. Social ecological theory emphasizes that these types of factors are inextricably connected at different “levels” (e.g., individual and community) within social environments and collectively affect health [14].

Undoubtedly, behaviors such as CAI, multiple sex partners, and receptive anal intercourse (RAI) increase one’s risk for HIV acquisition, especially when these behaviors occur with known HIV-positive sex partners [2, 9, 15]. In the absence of knowing partners’ HIV statuses, having partners who have sex within high-prevalence sexual networks can also increase HIV infection risk. For example, the high background prevalence of HIV among older MSM contributes to high HIV incidence among young MSM [10, 16]. Clinical factors like STIs [17] and HIV infection unawareness [9]—which increases viral load among HIV-positive young MSM and, thereby, facilitates transmission to HIV-negative young MSM—also increase HIV risk. Relevant psychosocial factors include depression, which has been prospectively and independently associated with HIV acquisition [15]; identified as a syndemic factor that increases risk for HIV acquisition over time [18]; and associated with HIV-discordant sex among MSM [19]. Additionally, substance use reduces sexual inhibitions and produces biophysiological changes (e.g., vasodilation) that increase HIV susceptibility [20]. Stimulants (e.g., amphetamines) [15, 20, 21], injection drug use, poppers [21], and polysubstance use [21] are known predictors of HIV acquisition.

A key aspect of social ecological theory is its emphasis on the underlying structural factors that shape the inequitable distribution of disease across populations [14]. Structural factors associated with incident HIV infection among MSM include low socioeconomic status (i.e., education and income) [1, 4] and lack of health insurance [1]. These factors limit access to health care and HIV testing, which prevent ongoing HIV transmission [22]. A structural factor especially pertinent to young MSM is the increasing visibility of homosexuality, which may be inclining MSM to express their sexuality at younger ages. Early sexual debut results in young MSM initiating sex when they are less informed about safer sex decision making than older MSM [23]. If contemporary young MSM have begun to express their sexuality at younger ages than did young MSM in the past, then contemporary young MSM have increasingly become exposed to HIV when they are developmentally vulnerable and are, therefore, at increased risk for HIV infection [11].

Given these factors, we sought to better explain age-associated HIV incidence disparities among MSM. We hypothesized that if young MSM experience more vulnerability than older MSM along these factors, then these factors may help to explain their elevated HIV incidence rates. Our hypotheses were as follows:

Behavioral

  1. 1.

    Young MSM are more likely than older MSM to engage in sexual risk behavior;

  2. 2.

    Young MSM are more likely than older MSM to have HIV-positive sex partners;

Clinical

  1. 3.

    Young MSM are more likely than older MSM to have STIs;

  2. 4.

    HIV-positive young MSM are more likely than HIV-positive older MSM to be viremic;

Psychosocial

  1. 5.

    Young MSM are more likely than older MSM to have depression;

  2. 6.

    Young MSM are more likely than older MSM to use substances;

Structural

  1. 7.

    Young MSM are more likely than older MSM to have low socioeconomic status;

  2. 8.

    Young MSM have less access to health care and HIV testing than older MSM; and

  3. 9.

    Young MSM have earlier ages of sexual expression than older MSM.

Methods

Search Strategy

During January–July 2016, we searched EMBASE, PsycINFO, PubMed, and Sociological Abstracts to acquire U.S. studies published during January 1996–July 2016 (Appendix in Table 2). In each database, we cross-referenced search terms for MSM (i.e., “men who have sex with men,” “MSM,” “gay,” “bisexual,” “gay and bisexual,” “sexual minority,” and “queer”) and HIV (i.e., “human immunodeficiency virus,” “HIV,” “risk behavior,” “infection,” “transmission,” “sexually transmitted disease,” “STD,” “sexually transmitted infection,” and “STI”). After acquiring studies, the 7 members of our research team, all of whom have academic and applied public health training, identified studies for inclusion. Team members recorded descriptive information for each study, including findings relevant for our review, on an abstraction form. The first author then validated all study information after conducting an independent review of the studies.

Inclusion Criteria

We established inclusion criteria to ensure that our review yielded data necessary for our study. All studies had to present data describing age’s association with factors related to at least 1 of our hypotheses. Consistent with the large volume of studies and surveillance reports that considered young MSM to be aged ≤ 29 years—and in consideration of the large number of HIV diagnoses that occur among MSM aged 20–24 and 25–29 years compared with older age groups [6]—we considered young MSM to be those who had not yet reached 30 years of age. We excluded studies containing only MSM aged ≤ 29 years. When multiple research teams analyzed the same data set that provided data for a hypothesis, we used only the most recent or most comprehensive analysis, unless study authors used different measures to assess outcomes. This strategy prevented us from including associations from 1 data set multiple times if associations were replicated across studies.

Analytic Approach

Millett et al.’s approach for explaining black-white HIV disparities among MSM guided our analysis [24]. This approach is useful in exploratory analyses that attempt to explain disparate HIV-related outcomes, and it can inform the subsequent development of meta-analyses that quantify disparities [13]. Additionally, understanding why 2 groups differ in their likelihoods of experiencing a health outcome (e.g., incident HIV infection) necessitates knowing if these groups differ in factors that are already known to contribute to that outcome [25]. Therefore, we considered studies that supported a hypothesis to be those in which young MSM were more likely than older MSM to exhibit characteristics of the hypothesis (e.g., younger age at sexual debut). We report findings from all studies that met inclusion criteria. Supportive studies reported statistically significant findings (ps < .05), 95% confidence intervals that excluded zero, or differences from a census (i.e., case surveillance data).

Because study authors differentially measured age (e.g., categorically vs. continuously) and assessed its association with outcomes, we summarized findings based on the degree to which age was generally associated with outcomes. However, we noted instances in which young MSM in distinct age categories had different likelihoods of behavioral, clinical, psychosocial, and structural factors. Because multivariable analyses can obscure associations that exist between variables [26], and because different studies use different multivariable models, we report results of studies’ bivariate findings except in a few instances in which studies only reported multivariable findings. To weigh the strength of evidence in support of hypotheses, we accounted for studies’ sample sizes, sample compositions, and designs (i.e., use of prospective analyses or probability-based data).

Results

Our search yielded 3132 studies, and 95 met inclusion criteria. We organized hypotheses with respect to their foci on primarily behavioral, clinical, psychosocial, and structural factors. Table 1 summarizes findings for hypotheses, or hypotheses’ components, supported by published studies.

Table 1 Supported hypotheses that may explain HIV incidence disparities among young and older men who have sex with men (MSM)
Full size table

Behavioral Hypotheses

Hypothesis 1

Young MSM are more likely than older MSM to engage in sexual risk behavior.

Any CAI

A total of 33 studies examined “any” CAI. Eight found that young MSM were more likely than older MSM to have any CAI [27,28,29,30,31,32,33,34] (2 of these examined HIV-discordant CAI [33, 34]), 21 found no age-related association [23, 35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54] (4 of these examined HIV-discordant CAI [51,52,53,54]), and 4 found that young MSM were less likely to have CAI [55,56,57,58] (1 of these examined HIV-discordant CAI [58]).

Within studies, inconsistent findings sometimes occurred. Occasionally, within stratified sub-groups (i.e., HIV-negative men [59] and men with primary sex partners [60]) young MSM were more likely than older MSM to have CAI, but no age-related associations existed for other sub-groups (i.e., HIV-positive men [59] and men without primary partners [60], respectively). Other studies found that young MSM were more likely to have CAI with some, but not all, types of partners (e.g., main vs. non-main) [61, 62].

Condomless Insertive Anal Intercourse (CIAI)

A total of 15 studies examined CIAI. Only 3 found that young MSM were more likely than older MSM to have CIAI [63,64,65] (1 of these examined HIV-discordant CIAI [65]), 6 found no age-related association [23, 29, 41, 47, 54, 66] (1 of these examined HIV-discordant CIAI [54]), and 4 found that young MSM were less likely to have CIAI [55, 67,68,69]. Within 2 studies, age-related associations were mixed and varied by partner type (i.e., casual vs. primary [60] and HIV-positive vs. HIV-negative [70]).

Condomless Receptive Anal Intercourse (CRAI)

A total of 16 studies examined CRAI. Four found that young MSM were more likely than older MSM to have CRAI [29, 47, 63, 71], 6 found no age-related association [23, 41, 64, 66, 72, 73], and 3 found that young MSM were less likely to have CRAI [55, 67, 68]. However, in 3 studies that specifically examined HIV-discordant CRAI—including 1 that used a probability-based sample [65]—young MSM consistently were more likely than older MSM to have HIV-discordant CRAI [2, 54, 65].

Number of Recent Sex Partners

A total of 10 studies examined the number of recent sex partners. Only 3 studies found that young MSM had a greater number of sex partners than older MSM [49, 52, 71], 3 found no age-related association [35, 42, 74], and 4 found that young MSM had fewer partners [29, 64, 70, 75].

RAI

The 3 studies that examined RAI supported the hypothesis [42, 55, 76]. Young MSM were more likely to identify as a “bottom” [42] and less likely to be “top” during their recent same-sex encounters [76]. Although 1 study only included Asian/Pacific Islander MSM [55], these men’s younger partners tended to be receptive during CAI.

Summary of Hypothesis 1

Studies yielded inconsistent findings regarding the association between age and most sexual risk behaviors among MSM. However, young MSM consistently were more likely than older MSM to have HIV-discordant CRAI and RAI.

Table 2 Literature review methodology
Full size table

Hypothesis 2

Young MSM are more likely than older MSM to have HIV-positive sex partners.

Known HIV-Positive Partners

The only study that provided data for this component found that young MSM were less likely than older MSM to have known HIV-positive partners [47].

Older Partners

Only 2 studies examined having older partners. Although 1 found that young MSM were more likely than older MSM to have older partners, it used data collected from HIV-positive men during 1993–1994 [49]. The other found no age-related association, but it was limited to 18–35 year-old black and Latino MSM [72].

Summary of Hypothesis 2

Although 1 relatively outdated study found that young MSM were more likely than older MSM to have older partners, studies have not shown that young MSM were more likely to have known HIV-positive partners.

Clinical Hypotheses

Hypothesis 3

Young MSM are more likely than older MSM to have STIs.

Any STI

A total of 7 studies examined “any” STI. Five found that young MSM were more likely to have any STI [2, 34, 49, 77, 78] (1 of these examined “chlamydia or gonorrhea” [78]), 1 found no age-related association [47], and 1 found that young MSM were less likely to have any STI [54]. It is noteworthy that 4 of the 5 supportive studies confirmed infections with diagnostic tests [2, 34, 78] or medical records [49]. The study that found no association used self-reported, outdated data collected during 1993–1994 [47]. The study that found that young MSM were less likely than older MSM to have any STI used self-reported data exclusively from men who viewed sexually explicit online media [54].

Gonorrhea

A total of 4 studies examined gonorrhea. Three found that young MSM were more likely to have gonorrhea infection [77, 79, 80], and 1 found no age-related association [49]. Of the 3 studies that supported this component, 2 used diagnostic tests to confirm infections [79, 80], of which 1 used a very large sample of HIV-positive and HIV-negative MSM (n = 21,927) [80]. The study that found no age-related association used a small sample (n = 336) that only included HIV-positive MSM recruited during 1993–1994 [49].

Chlamydia

A total of 3 studies examined chlamydia. Two found that young MSM were more likely to have chlamydia infection [49, 79], but 1 found no age-related association [80]. The 2 supportive studies have limitations because they included only HIV-positive [49] or incarcerated [79] MSM. The study that found no age-related association used a large sample of HIV-positive and HIV-negative MSM (n = 21,927) [80].

Syphilis

A total of 4 studies examined syphilis. One found that young MSM were more likely than older MSM to test positive for early syphilis (primary, secondary, or early latent) [79]. However, 2 studies found no age-related association for early syphilis [49, 81], and 1 found that young MSM were less likely to be seroreactive for syphilis (including infection that could have been treated previously) [80]. The studies that supported and contradicted the hypothesis both used large, racially/ethnically diverse samples of MSM (ns > 7000).

We identified 2 additional studies that examined age-related trends in primary and secondary syphilis diagnoses during 2004–2008. During this time, primary and secondary syphilis increased among all MSM [82, 83]. However, relative increases in diagnoses among black, Latino, and white MSM aged 25–29 years were greater than increases among older black, Latino, and white MSM, respectively. Moreover, black and Latino MSM aged 13–24 years experienced greater relative increases in syphilis than older black and Latino MSM, respectively. Black MSM in all age groups had greater relative increases than age-matched Latino and white MSM [83].

Herpes Simplex Virus 2

Only 2 studies examined herpes simplex virus 2. Both found that young MSM were less likely than older MSM to test positive for herpes [84, 85].

Summary of Hypothesis 3

The most generalizable data suggested that young MSM were more likely than older MSM to have “any” STI and gonorrhea. Findings for chlamydia and syphilis were inconsistent. However, primary and secondary syphilis diagnoses increased more among young (vs. older) MSM during 2004–2008.

Hypothesis 4

HIV-positive young MSM are more likely than HIV-positive older MSM to be viremic.

Viral Suppression

The 3 studies that examined viral suppression supported this hypothesis. Young HIV-positive MSM were less likely than older HIV-positive MSM to be virally suppressed [86, 87] and use antiretroviral therapy [22, 87]. Robust data appeared in a nationally representative, probability-based study [87] and a study that recruited MSM from 21 cities [22].

HIV Infection Unawareness

A total of 3 studies examined HIV infection unawareness. Two found that young HIV-positive MSM were more likely than older HIV-positive MSM to be unaware of their infection [9, 88], but 1 found no age-related association [89]. Of note, the 2 supportive studies both replicated findings across 2 data collection waves [9, 88], and 1 included 8153 men in 21 cities [9]. The study that found no age-related association only included indigent, substance-using MSM in 1 city [89].

Summary of Hypothesis 4

HIV-positive young MSM consistently were less likely than HIV-positive older MSM to be virally suppressed and use antiretroviral therapy. The strongest data for HIV infection unawareness suggested that HIV-positive young MSM were more likely to be unaware.

Psychosocial Hypotheses

Hypothesis 5

Young MSM are more likely than older MSM to have depression.

A total of 7 studies provided data for depression. Two found that young MSM were more likely than older MSM to have depression [70, 90], but 5 found no age-related association [19, 36, 71, 91, 92].

The 2 supportive studies both used racially/ethnically diverse samples, and 1 used a large sample from 6 cities (n = 4295) [70]. The 5 studies that found no age-related association included only black MSM [19, 71, 91], predominantly Latino MSM [36], or only HIV-positive black and Latino MSM [92]. Moreover, 3 of these 5 studies had very small samples (ns = 205 [36], 197 [19], and 199 [92]).

Summary of Hypothesis 5

Data from racially/ethnically diverse samples suggested that young MSM as a whole had more depression than older MSM. However, age was not associated with depression in studies limited to black or Latino MSM.

Hypothesis 6

Young MSM are more likely than older MSM to use substances.

Amphetamines

A total of 9 studies examined amphetamines. Four found that young MSM were more likely than older MSM to use amphetamines [47, 70, 93, 94], 2 found no age-related association [42, 95], and 2 found that young MSM were less likely to use amphetamines [49, 96]. One study produced mixed findings: MSM aged 20–29 years were less likely than MSM aged 30–39 years to use amphetamines, but they did not statistically differ from MSM aged ≥ 40 years [62].

Cocaine (Including Crack)

A total of 5 studies examined cocaine. One found that young MSM were more likely than older MSM to use cocaine, but no age-related association existed for crack [70]. Another study found no age-related association for cocaine [96], but 3 found that young MSM were less likely to use cocaine [49, 62, 94] (1 of these examined crack) [94].

Injection Drug Use

A total of 4 studies examined injection drug use. Two found no association between age and injection drug use [49, 97], and 2 found that young MSM were less likely than older MSM to inject drugs [47, 70].

Polysubstance Use

A total of 5 studies examined polysubstance use. Three found that young MSM were more likely than older MSM to engage in polysubstance use [93, 98, 99], 1 found no age-related association [86], and 1 found that young MSM were less likely to engage in polysubstance use [100].

Of note, the 3 studies that supported this component used multiracial samples—1 used a probability-based sample of MSM [99]—and assessed polysubstance use during the past 2–6 months. The 1 study that found no age-related association used an imprecise measure for young age (≤ 39 years) and only assessed polysubstance during the past month [86]. The study that found that young MSM were less likely to engage in polysubstance use was limited to mostly foreign-born Asian/Pacific Islander MSM [100].

Poppers

A total of 3 studies examined poppers. One found no association between age and poppers use [49], and 2 found that young MSM were less likely than older MSM to use poppers [47, 70].

Substance Use During Sex

A total of 3 studies examined substance use during sex. One found that young MSM were more likely than older MSM to use amphetamines during sex, but less likely to use crack during sex [94]. Two found no age-related association for “any” substance use during sex [91, 101].

Summary of Hypothesis 6

The strongest and most generalizable data for polysubstance use suggested that young MSM were more likely than older MSM to engage in this behavior. Findings for other substances were inconsistent.

Structural Hypotheses

Hypothesis 7

Young MSM are more likely than older MSM to have low socioeconomic status.

Education

A total of 5 studies examined education. Two found that young MSM had less education than older MSM [32, 70], 2 found no age-related association [42, 91], and 1 found that young MSM had more education [2].

Although 1 study that supported this component used a large, multiracial sample (n = 4295) [70], the other only included 60 Asian/Pacific Islander MSM [32]. Studies that found no age-related association included only MSM who sought CAI via the Internet [42] or black MSM [91]. The study that contradicted the hypothesis only included black MSM [2].

Income

A total of 6 studies examined income. Three found that young MSM had less income than older MSM [32, 47, 70], 2 found no age-related association [91, 92], and 1 found that young MSM had greater income [2].

It is noteworthy that 2 of the 3 studies that supported this component used large, multiracial samples (ns = 2189 [47] and 4295 [70]), although 1 only included 60 Asian/Pacific Islander MSM [32]. However, the 2 studies reporting no association used single-city samples limited to HIV-negative black MSM [91] or HIV-positive black and Latino MSM [92]. The 1 study that contradicted the hypothesis only included black MSM [2].

Summary of Hypothesis 7

Studies suggested that young MSM as a whole, but not black or Latino MSM, had less income than older MSM. Findings for education were relatively inconsistent.

Hypothesis 8

Young MSM have less access to health care and HIV testing than older MSM.

Health Care

The 4 studies that examined health care access supported this hypothesis [2, 22, 47, 102]. Younger age was associated with lacking medical insurance [47] and, among HIV-positive MSM, not being in medical care [22]. Among black MSM, younger age was associated with lacking a usual place of health care [2], lacking access to necessary care [2], and not visiting a medical provider recently [102].

Lifetime HIV Testing

A total of 8 studies examined lifetime HIV testing. Seven found that young MSM were less likely to ever test for HIV than older MSM [32, 50, 103,104,105,106,107], but 1 found no age-related association [2]. Most of the 7 supportive studies used multiracial samples [103,104,105,106,107], including 1 that used a large sample (n = 7271) from 21 cities [103]. The study that found no age-related association only included black MSM.

Recent HIV Testing

A total of 12 studies examined recent HIV testing. Two found no association between age and recent testing [28, 108], and 10 found that young MSM were more likely than older MSM to recently test for HIV [29, 49, 103, 105, 107, 109,110,111,112,113]. Nine of these 10 studies used multiracial samples—1 used an all-black sample [111]—including 1 that used a large sample (n = 7271) from 21 cities [103]. Studies finding no age-related association included only black and Latino MSM [28] or men in 1 city [108].

Summary of Hypothesis 8

Young MSM consistently had less access to health care than older MSM. Notwithstanding minor inconsistencies, young MSM were less likely to ever test for HIV, but more likely to recently test.

Hypothesis 9

Young MSM have earlier ages of sexual expression than older MSM.

The 4 studies that provided data for this hypothesis supported it [54, 114,115,116]. Two, including 1 that used probability-based data [115], found that young MSM began having anal intercourse at younger ages than did older MSM [115, 116]. A cohort-based study reported that men born in the 1990s initiated anal intercourse at younger ages than men born in the 1970s and 1980s, but not before those born before 1970 [54]. A study examining “gay-related developmental milestones,” including sexual debut and “coming out,” found that gay-related development began at younger ages among young MSM [114].

Summary of Hypothesis 9

Young MSM consistently had earlier ages of sexual expression than older MSM.

Discussion

In this study, we used social ecological theory as a framework for identifying potential determinants of age-associated HIV-incidence disparities among MSM. In doing so, we build upon previous studies’ contributions. Clearly, young MSM are at greater risk for HIV than older MSM because the relatively low HIV prevalence among young MSM results in a high number of young (vs. older) MSM being vulnerable to HIV acquisition [7]. However, as our findings suggest, multiple behavioral, clinical, psychosocial, and structural factors likely exacerbate this phenomenon.

Age-related differences in some sexual risk behaviors existed. Although we did not find that young MSM were more likely than older MSM to have known HIV-positive partners, they were more likely than older MSM to have RAI and HIV-discordant CRAI. HIV-discordant CRAI is the most risky sexual behavior for HIV acquisition [117]. Previous studies suggested that RAI and HIV-discordant CRAI may often occur when young MSM have sex with older partners [7, 10, 16]. Despite the limited data that we acquired for having older partners, studies have shown that age-related interpersonal dynamics can prompt young MSM to be anally receptive [55] and, for black young MSM, less empowered to negotiate condom use during sex with older MSM [118]. In 1 analysis, the association between having CAI and having older partners was strongest for black young MSM compared with other young MSM [119]. These data suggest that the increased likelihood for HIV-discordant CRAI among young MSM and, perhaps HIV-discordant CRAI with older men, promotes age-associated disparities in HIV incidence, especially for black MSM.

Other characteristics of sexual networks, including age-concordant sex, are also relevant. Young MSM consistently had earlier ages of sexual expression than older MSM. This phenomenon has increased the amount of time during which contemporary cohorts of young MSM are exposed to HIV, and it has resulted in young MSM initiating sex when they are developmentally vulnerable. Young MSM were also more likely than older MSM to have “any” STI and gonorrhea, which is consistent with age-related findings from population-based data for U.S. men and women [120]. Most notably, primary and secondary syphilis disproportionately increased among young MSM during 2004–2008. This suggests that young MSM are increasingly having sex within networks where syphilis and HIV are becoming more prevalent [83]. Therefore, the likelihood of young MSM acquiring HIV from other young MSM, and not older MSM alone, has likely increased over time.

Psychosocial challenges likely enhance HIV risk among young MSM. They had increased likelihood for polysubstance use and, in multiracial samples, depression. Together, these factors can reduce sexual inhibitions and sydemically interact with other factors that promote HIV acquisition [18]. Although we did not find that young MSM were more likely than older MSM to use substances during sex, our findings do suggest that young MSM may be at increased risk for experiencing reduced cognition and impaired decision making prior to and during sexual encounters. These factors, in turn, increase their vulnerability to HIV infection [77].

The aforementioned factors emerge within broader contexts of socioeconomic vulnerability. Compared with older MSM, young MSM had lower income and less access to health care. These factors are determinants of HIV-related disparities [12]. Therefore, it was not surprising that young MSM were less likely to ever test for HIV and, if HIV-positive, use antiretroviral therapy, be virally suppressed, and have awareness of their infection. Socioeconomic and other structural barriers make it difficult for HIV-positive young MSM to be engaged along the HIV care continuum. These barriers also make young MSM vulnerable to transmitting HIV to their partners, including other young MSM. Although studies did not find greater socioeconomic vulnerability among black young (vs. older) MSM, socioeconomic vulnerability likely has dire health-related consequences for black young MSM. Black young MSM fare worse than non-black young MSM with regard to poverty, insurance coverage, engagement in HIV care, and viral suppression [13].

This review has limitations. We could not draw strong conclusions for some hypotheses due to having limited data. This was notable for our hypothesis regarding age-related differences in having older and known HIV-positive partners. Second, some supported hypotheses had few studies that consistently provided evidence to support them. An example includes HIV infection unawareness, for which our inclusion criteria yielded relatively few studies. Third, although black young MSM are more vulnerable to HIV than other young MSM, the lack of studies examining age-associated differences among black MSM prevented us from highlighting more factors that are potentially relevant for black young MSM. Finally, given the exploratory nature of our study, we did not use methodologies (e.g., meta-analyses) that could quantify differences in factors that contribute to HIV acquisition among young and older MSM. Therefore, our review provides no effect-size data for the relationship between age and the factors that we examined. Future studies should examine age in relation to HIV risk factors to provide better understanding of age-associated HIV incidence disparities among MSM. Meta-analyses, prospective designs, and greater attention to black young MSM would provide critical insights.

Our review identified behavioral, clinical, psychosocial, and structural factors that could promote high HIV incidence among young MSM. Interventions that address these factors might prevent CRAI, delay sexual debut among young MSM, and decrease risks from sex with older MSM. Increasing uptake of HIV pre-exposure prophylaxis among HIV-negative young MSM, as well as early linkage to care and initiation of antiretroviral therapy for HIV-positive young MSM, could reduce HIV acquisition and transmission among young MSM. Screening and treatment for other STIs, mental health problems, and substance use would also be beneficial. Myriad, interrelated factors shape HIV risk and contexts in which this risk emerges among young MSM [11]. Multifaceted prevention strategies that address these factors could help to reduce age-associated HIV incidence disparities among MSM.