In the United States, 24% of HIV infections have been attributed to high-risk heterosexual contact.1 Studies have shown that the segment of the heterosexual population most at risk for HIV infection is persons of lower socioeconomic status or those living in poverty-stricken areas, making this group a priority for HIV surveillance and prevention activities.1 One explanation for the disparity of HIV burden being higher among persons living in poverty areas is the possibility that people living in low socioeconomic status areas may have different sexual network transmission dynamics than those in other areas of the United States.
Epidemiologic definition for the high-risk heterosexual contact transmission category by Centers for Disease Control and Prevention specifically refers to an individual's reported exposure through heterosexual sex with someone who is known to be HIV positive, a person with a history of injecting drugs, or a man who has sex with men. It is unclear, therefore, if HIV infections among heterosexuals, and heterosexual women in particular, are a consequence of female heterosexual contact with heterosexual men or with men who have sex with men (MSM) or men who are persons who inject drugs (PWID). Furthermore, direct evidence of sustained chains of heterosexual transmission is lacking, at least in the United States. Phylogenetic studies have shown transmission links between heterosexual women and MSM that are more numerous (44% of connections to women) than from heterosexual men to women (18%) and from heterosexual male PWID to heterosexual women (13%).2 With the greatest concentration of HIV in the United States located among key populations, such as MSM and PWID,1 interventions to mitigate HIV among these higher HIV burden groups may have a highly effective secondary effect of reducing HIV acquisition among heterosexual women. That is, focusing HIV interventions among a relatively small group of MSM and PWID (IDU) as opposed to a large number of women, who may arguably not know that they are a risk and thus unlikely to engage in interventions, may be the most effective means to reduce infections among women.
To provide additional support to the conclusions drawn from the phylogenetic studies mentioned above, we examined this question through a different methodological lens. To this end, we investigated the possibility of transmission risks between MSM and female heterosexuals by examining the levels of multiple transmission risk behaviors among MSM and comparing these levels with the background epidemiologic profile of HIV in selected US cities. We hypothesized that should transmission of HIV from MSM, a population with high HIV rates, to women be occurring, we would observe this relationship in a correlation between the percentage of men who have sex with men and women (MSMW) in each city and the percentage of those cities' female HIV cases that were attributed to heterosexual transmission. Alternately, if there was no overlap, there would be no correlation between the prevalence of MSMW behavior and the level of female heterosexual HIV cases. Finally, we sought to characterize cities in terms of race/ethnicity and poverty as these factors have been implicated in high rates of HIV infection among heterosexual women to determine if these characteristics overlapped with our theorized relationship between MSMW and HIV among heterosexual women.
We accessed data from National HIV Behavioral Surveillance (NHBS) to describe transmission risk behaviors of MSM. NHBS is a Centers for Disease Control and Prevention–led collaboration that samples, interviews, and offers HIV testing to MSM using time location sampling every 3 years, which, in 2011, included 20 US cities. The methods used to sample MSM have been documented elsewhere.3 Data gathered in 2011 from 12 NHBS sites are included in the current analysis from NHBS among MSM conducted in 2011. These 12 sites include the following: Baltimore, Detroit, Denver, Houston, Los Angeles, Miami, New Orleans, New York, Philadelphia, San Francisco, Seattle, and the District of Columbia. Men eligible for the study and included in the analysis were 18 years or older, lived in one of the study cities, and reported ever having had sex with a man. For epidemiologic profiles of people living with HIV in each city at the end of 2010, we elicited aggregate data from mandatory HIV case reporting for each city, which were available either on the internet or by request from individual health departments. “City” is defined as the NHBS data collection jurisdiction for each individual-funded entity and can include an entire metropolitan statistical area (MSA), a city, or a combination of counties. Moreover, HIV case reporting monitors morbidity and mortality of individuals with HIV over time. Jurisdictions are able to report the number of living HIV cases (ie, infected with HIV and still living during the period of the report) over a given time in the jurisdiction and by transmission category. We chose to use living HIV cases at the end of 2010 to match the cross-sectional nature of the NHBS data. Finally, we accessed US census data to describe cities in terms of poverty and race/ethnicity composition.4
For MSM, we categorized men into transmission risk groups based on their self-reported behaviors over the 12 months preceding the NHBS survey. Men sampled as part of the MSM sample in NHBS were categorized as having (1) only male partners, (2) male and female partners, (3) female partners only (but identified as gay/bisexual), (4) IDU as only risk, (5) male and female partners and IDU, (6) IDU and male partners, (7) female partners and IDU, and (8) none of the above behaviors in the past 12 months. Men who had any female partners were categorized as men who were MSMW.
HIV case reports are categorized following HIV surveillance guidelines into behavioral risk groups during the process of case reporting. The categories are female heterosexual contact, male heterosexual contact, female IDU, male IDU, MSM, MSM IDU, and no identifiable risk (NIR). We focused only on male and female heterosexual and NIR cases and MSM cases in the current analysis.
NHBS data were analyzed in SAS 9.3 to produce counts and frequencies of risk behaviors among MSM. These data were then tabulated along with living HIV case count data as of December 2010 from each city. Correlations between the percentages of male and female heterosexual cases and the percentage of MSM in NHBS who are MSMW and corresponding P values were calculated in SAS 9.3. Mean poverty and proportion of each city that was African American/Black were calculated using US Census data. Comparisons of these were used to characterize clusters of cities.
Transmission Risk Behaviors—NHBS
The 12 NHBS sites collected data from 6137 (mean per city = 511) men who identified as gay/bisexual or had sex with another man in the past 12 months in 2011. The majority of MSM (80% or higher) in most sites reported only MSM behavior in the past 12 months. Notably, MSM in Baltimore had the lowest percentage reporting only male–male sexual behavior in the last 12 months at 68.6%. There was a wide range of percentages across sites of MSM reporting only having female partners in the past 12 months. No MSM (0%) in Denver reported having only female partners in the past 12 months compared with 5% in Baltimore, with most other sites having a percentage of approximately 1%. The percentage of men across the sites that reported male and female partners in the past 12 months also varied widely. The lowest was reported in San Francisco at 5.3% and the highest in Baltimore at 21.1%. Only 4 sites had a percentage of reporting male and female partners above 10% but below 20% while this percentage was approximately 7.5% in 5 sites. In contrast, all sites had low percentages (<2.5%) reporting MSM and IDU behavior in the past 12 months with the exceptions of San Francisco (3.5%), Seattle (5.3%), and Denver (3.5%). In terms of MSM who only engaged in IDU behavior in the past 12 months, almost all sites reported 0% of MSM engaging in this behavior with the exception of New Orleans and Seattle (0.5% each). Finally, the percentage of MSM across sites that reported a male partner, female partner, and IDU behavior in the past 12 months was generally very low (≤1%) with the exceptions of New Orleans (2.3%) and Houston (1.9%) (Table 1).
Living HIV Cases—Case Reporting
Table 2 shows case counts and percentage of total counts for each behavioral risk group in each city. The largest percentage of living HIV cases across all sites was among MSM. The percentages of cases that were MSM ranged from 88.5% in San Francisco to 31.7% in Philadelphia. In contrast, the percentages of living male cases that were attributed to heterosexual contact were low across the sites and ranged from 15.2% in Miami to 1.0% in San Francisco. The percentages of living HIV cases attributed to heterosexual contacts were overall less than that of MSM cases across the sites but varied from 25% in Miami to 2.1% in San Francisco. Seven sites had more than 15% of their total living cases among females with heterosexual contact (Baltimore, Washington, DC, Philadelphia, Detroit, New York City, Miami, and Houston).
Race/Ethnicity and Poverty Indicators—US Census
Data were obtained from all sites calculating the percentage of their MSA that lives in poverty and the percentage of their MSA that are African American/Black from the 2010 Census. Five sites had a percentage of population who are African American/Black over 30% (Baltimore, New Orleans, Philadelphia, and Washington, DC). Five sites had a percent living in poverty over 20% (Baltimore, Detroit, New Orleans, New York City, and Philadelphia).
There was a high correlation between the percentage of cities' HIV cases that were attributed to female heterosexual and male heterosexual contacts (r = 0.936, P < 0.0001) as might be expected as heterosexual men and women are likely the most frequent sexual contacts of each other. However, there was also a moderate but not statistically significant correlation between the percentage of MSM who were classified as MSMW in NHBS sites and the percentage of cities' epidemics that were female and attributed to heterosexual contact (r = 0.49, P = 0.1), whereas the correlation between female heterosexual cases and male IDU cases across jurisdictions was low (r = 0.35, P = 0.26) (data not shown). Moreover, there were 2 notable clusters of cities when we plotted the correlation between the NHBS MSMW percentage and of case report data that were female heterosexual cases. These clusters were different in terms of mean percentage of residents who were African American/Black and percentage of people living in poverty defined by the US Census. Seven cities (San Francisco, Seattle, Los Angeles, Denver, Washington, DC, New York City, and Houston; Cluster A in Fig. 1) had a mean percentage in poverty of 16.1% compared with 5 cities (Baltimore, Miami, Philadelphia, New Orleans, and Detroit; Cluster B in Fig. 1) where the mean percentage in poverty was 21.8%. Moreover, there were differences in the mean percentage of the population that was African American/Black; 17.5% vs. 42.9% in Cluster A and Cluster B, respectively. Finally, the mean percentages of men in NHBS who were MSMW were 9.4% and 17.9%, in Cluster A and Cluster B, respectively.
We found clear differences in transmission HIV risk behaviors among MSM in 12 US cities. Although the majority of MSM across sites only engaged in MSM behaviors, there were notable differences across sites in terms of the percentage of MSM who also had female partners. Previous research has noted that MSMW report fewer HIV risk-taking behaviors than MSM only, suggesting that overlapping risk from MSM to heterosexual women may be responsible for a small but nontrivial percentage of HIV infections among women.5,6 Potential bridging of HIV from MSM to heterosexual females is not likely uniform across the United States and appears, based on the current analysis, to be concentrated in cities with higher poverty and higher percentages of African American/Black residents, a group with already higher rates of HIV than other race/ethnicity groups. These cities are also characterized by higher percentages of MSMW, and overall, the combination of MSM and IDU behaviors was low in the 12 NHBS sites included in the current analysis. In studies conducted among IDU, in these same cities, only 10% of male IDU reported same sex behavior in the past 12 months.7 Furthermore, the distribution of HIV cases across the behavioral risk groups is not uniform across the United States despite MSM comprising the largest percentage consistently across all study sites.8
There was a high correlation between the percentage of reported living HIV cases from male heterosexual contact and female heterosexual contact, as might be expected because these populations are most commonly engaging in sexual risk with each other. However, there was also a moderate but nonsignificant correlation between the percentage of MSM who were MSMW in NHBS and the percentage of reported cases that were from female heterosexual contact suggesting some potential overlap that bears further analysis across larger samples of cities and with additional data on potential differences among cities that can be used as control variables in the analysis.
This analysis is subject to limitations. First, our analysis is an ecologic analysis and is subject to the potential of ecologic fallacy in that the patterns observed in the present analysis may not represent the individual level. Moreover, our findings may be confounded by unmeasured variables that cannot be adjusted for. Future work on this question would benefit from more extensive measures of potential confounders that can be used in analysis. Second, the data for the present analysis come from 12 US cities and thus may not represent what is occurring in other areas of the United States. Nonetheless, we feel that the geographic and socioeconomic diversities across the 12 cities lend itself to potentially understanding a phenomenon that may be occurring elsewhere. Moreover, our analysis presents a testable hypothesis. It should be possible to test whether other cities that have high poverty and high concentrations of African American/Black populations also have moderate correlations between the percentages of MSMW and female heterosexual HIV cases. Third, data from NHBS are collected at venues attended by MSM and are collected with an interviewer-administered survey. These venues may not capture some segment of MSMW as readily as others. Social desirability bias may have had an impact on behaviors reported by MSM. Fourth, HIV case reporting data may not be 100% complete at any given time because of lags in reporting and identification of transmission risk groups. Fifth, slight variations in defining risk groups among reported HIV cases across jurisdictions may have influenced the percentage of each city's cases that were NIR. Despite these variations, most cities' percentages of NIR cases were similar. Lastly, as with every cross-sectional analysis, we cannot ascribe causality to our conclusion.
Despite these limitations, the current study suggests, as have other studies in the literature,9,10 that there is a relationship between poverty, African American/Black populations, MSMW, and the level of female heterosexual HIV. Factors such as financial instability, incarceration, internalized homonegativity, societal gay stigma, gender role conflict, and childhood sexual abuse among African American/Black communities and African American/Black MSM, in particular, have been suggested as contributory to HIV risk taking and contextual to the spread of HIV among heterosexual women because of increased sexual overlap between MSM and women in these communties.11–16 Addressing these profound health issues among African American/Black MSM would have the dual effect of improving the health of a segment of the African American/Black MSM population that has experienced a high burden of HIV and to improve the health of other members of their larger communities. Further research is warranted to better understand these factors and how they contribute to the potential bridging of HIV from MSM to female heterosexuals. Greater understanding of these factors could inform more effective HIV interventions in 2 populations at high risk for HIV transmission and acquisition.
1. Centers for Disease Control and Prevention. Diagnoses of HIV Infections in the United States
and Dependent Areas, 2014. Available at: http://http://www.cdc.gov
/hiv/pdf/library/reports/surveillance/cdc-hiv-surveillance-report-us.pdf. Accessed April 4, 2016.
2. Oster AM, Wertheim JO, Hernandez AL, et al. Using molecular HIV surveillance data to understand transmission between subpopulations in the United States
. J Acquir Immune Defic Syndr. 2015;70:444–451.
3. MacKellar DA, Gallagher KM, Finlayson T, et al. Surveillance of HIV risk and prevention behaviors of men who have sex with men
–a national application of venue-based, time-space sampling. Public Health Rep. 2007;122(suppl 1):39–47.
4. US Census. American Fact Finder. Available at: http://factfinder.census.gov/faces/nav/jsf/pages/index.xhtml. Accessed June 15, 2016.
5. Maulsby C, Sifakis F, German D, et al. HIV risk among men who have sex with men
only (MSMO) and men who have sex with men
and women (MSMW) in Baltimore. J Homosex. 2013;60:51–68.
6. Friedman MR, Wei C, Klem ML, et al. HIV infection and sexual risk among men who have sex with men
and women (MSMW): a systematic review and meta-analysis. PLoS One. 2014;9:e87139.
7. Centers for Disease Control and Prevention. HIV infection and HIV-associated behaviors among persons who inject drugs—20 cities, United States
, 2012. MMWR Morb Mortal Wkly Rep. 2015;64:270–275.
8. Centers for Disease Control and Prevention. HIV Surveillance Report, 2014. Vol. 26. Available at: http://www.cdc.gov
/hiv/library/reports/surveillance/. Accessed November 29, 2016.
9. Dinenno EA, Oster AM, Sionean C, et al. Piloting a system for behavioral surveillance among heterosexuals at increased risk of HIV in the United States
. Open AIDS J. 2012;6:169–176.
10. Latkin C, Yang C, Tobin K, et al. Differences in the social networks of African American men who have sex with men
only and those who have sex with men and women. Am J Public Health. 2011; 101:e18–e23.
11. Harawa NT, Obregon NB. McCuller WJ.Partnerships between black women and behaviorally bisexual men: implications for HIV risk and prevention. Sex Cult. 2014;18:570–891.
12. Harawa NT, Williams JK, Ramamurthi HC, et al. Sexual behavior, sexual identity, and substance abuse among low-income bisexual and non-gay-identifying African American men who have sex with men
. Arch Sex Behav. 2008;37:748–762.
13. Bingham TA, Harawa NT, Williams JK. Gender role conflict among African American men who have sex with men
and women: associations with mental health and sexual risk and disclosure behaviors. Arch Sex Behav. 2013;103:127–133.
14. Benoit E, Downing MJ Jr. Childhood sexual experiences among substance-using non-gay identified Black men who have sex with men
and women. Child Abuse Negl. 2013;37:679–690.
15. Quinn K, Dickson-Gomez J, DiFranceisco W, et al. Correlates of internalized homonegativity among black men who have sex with men
. AIDS Educ Prev. 2015;27:212–226.
16. Arnold EA, Rebchook GM, Kegeles SM. 'Triply cursed': racism, homophobia and HIV-related stigma are barriers to regular HIV testing, treatment adherence and disclosure among young Black gay men. Cult Health Sex. 2014;16:710–722.