Article Text
PDF
Abstract
Objectives The authors examine whether young black and Latino men who have sex with men (MSM) who have older sex partners are more likely than those who do not have older sex partners to have unrecognised HIV infection. The authors examine whether the association stems from (1) increased sexual risk behaviour with male partners of any age, (2) heightened risk of being exposed to HIV infection by older partners or (3) a combination of these two factors.
Methods The analytical sample consisted of 723 black and Latino MSM, aged 18–35 years, who were HIV negative or of unknown serostatus at study entry. Participants completed a self-administered questionnaire and were tested for HIV infection. Men who reported having a male sex partner who was at least 4 years older than themselves were compared with those who did not. Outcomes included unprotected receptive anal intercourse (URAI) with male partners of any age (past 3 months) and having unrecognised HIV infection.
Results Men with older partners reported a higher prevalence of URAI (AOR=1.50, 95% CI 1.02 to 2.21). A second model found that men with older partners had increased odds of having unrecognised HIV infection (AOR=2.51, 95% CI 1.18 to 5.34) after controlling for the number of URAI partners of any age, which remained an independent predictor.
Conclusions Young black and Latino MSM who had older male sex partners were at increased risk of having unrecognised HIV infection. This heightened risk was associated with sexual risk behaviour with partners of any age as well as possible increased exposure to HIV infection from older partners.
- Behavioural science
- men
- epidemiology (general)
- sexual networks
Statistics from Altmetric.com
Introduction
Sexual mixing dynamics may affect one's risk of acquiring HIV infection.1–4 One such mixing pattern is age discordance between sexual partners. Older men who have sex with men (MSM) are more likely to be HIV positive,5 and thus when younger MSM have unprotected sex with older men, they may increase their risk of becoming infected. Several studies support this hypothesis.2 6 7 However, this association has not been examined specifically among MSM of colour.
Black and Latino MSM represent over half of the HIV infections reported among MSM in the USA.8 HIV prevalence among black MSM is five times and among Latino MSM almost three times that of white MSM.8 The frequency of unprotected sex does not explain the disparity in HIV prevalence between ethnic minorities and white MSM.9–12 If a sizeable percentage of young MSM of colour have older partners, this may partially account for the increased infection rate in this population.
We examine whether young black and Latino MSM who have older sex partners are more likely than their counterparts who do not have older sex partners to have unrecognised HIV infection. We focused on two explanatory hypotheses. First, young men who have older partners may engage in a generalised pattern of high-risk sexual behaviours with their male partners regardless of the age of those partners. Second, young men who have older sex partners may have increased probability of being exposed to HIV because the prevalence of infection is higher among older men compared with younger men. To test the first hypothesis, we examined whether men who had older male partners (compared with those who did not) were more likely to report sexual risk behaviours with male partners of any age. To test the second hypothesis, we examined whether men who had older male partners had an increased odds of unrecognised infection after statistically controlling for sexual risk behaviour with male partners of any age.
Methods
Procedures
The findings reported here are from the Brothers y Hermanos Study, which was sponsored by the US Centers for Disease Control and Prevention.13 14 The study examined socio-cultural, psychological and behavioural characteristics of black and Latino MSM HIV infection. From May 2005 to April 2006, 2235 MSM were enrolled in the study. Black MSM were recruited in New York City (n=614) and Philadelphia (n=540) and Latino MSM were recruited in New York City (n=516) and Los Angeles County (n=565). Men were eligible if they were at least 18 years of age, reported sex (oral sex, anal sex or mutual masturbation) with a man in the past 12 months and were a resident of the recruitment area. Men who were HIV positive, HIV negative or of unknown serostatus were eligible. Participants were recruited through respondent-driven sampling,15 and those who completed the study were given the opportunity to recruit up to three individuals from their social networks.
After providing written informed consent, participants completed an audio computer-assisted self-interview in English or Spanish. Next, men who did not disclose during eligibility screening that they had previously been diagnosed with HIV received an oral fluid rapid HIV antibody test (OraQuick Advance; Orasure Technologies, Inc, Bethlehem, Pennsylvania, USA). Trained HIV counsellors delivered results and counselling. Those who tested preliminarily positive on the rapid test and those who reported a previous HIV diagnosis provided a blood specimen for confirmatory testing with a Western blot assay. All participants received $50 for study participation and $15–$20 for each eligible individual they recruited. The protocol was approved by the Institutional Review Boards at the Centers for Disease Control and Prevention (protocol #4417) and at each study site. A more detailed description of the study methods and results of the respondent-driven sampling process has been published elsewhere.13 The analyses reported here focus on black and Latino MSM between the ages of 18 and 35 years who self-reported being HIV negative or of unknown HIV status.
Measures
Sexual risk behaviour
Men were asked about anal intercourse and condom use in the past 3 months. We provided descriptive data on the occurrence (yes/no) of the following behaviours with any male partners: (1) unprotected anal intercourse (insertive or receptive) (UAI), (2) unprotected insertive anal intercourse (UIAI) and (3) unprotected receptive anal intercourse (URAI). Our regression models focused on URAI because it has the highest per-contact risk of acquiring HIV infection.5 This variable is the dependent measure in the analysis examining whether having older partners is associated with high-risk sexual behaviour. Additionally, we created a three-level categorical variable for the number of URAI partners in the past 3 months (‘0’, ‘1’ or ‘2 or more’) that was examined as a predictor of unrecognised HIV infection.
Partner age
Age of male sex partners was ascertained by asking participants about the age of (1) the most recent partner with whom they had unprotected anal sex, (2) the most recent partner with whom they had protected anal sex and (3) the most recent main partner. Partner's age was measured using the following categories: ‘less than 16’, ‘16–20’, ‘21–25’, ‘26–30’, ‘31–35’, ‘36–40’, ‘41–50’ and ‘over 50’. Participants were classified as having an older partner if they indicated that any of these partners were at least 4 years older than themselves. That is, the lower boundary of the partner age category had to be at least 4 years more than the participant's age. The age difference between partners was based on previous research showing heightened risk for infection among MSM who had male partners who were, on average, 4 years older than themselves.7
Unrecognised HIV infection
Participants were asked if they had ever been tested for HIV outside of this study. Previous testers reported the date and result of the most recent test (‘positive’, ‘negative’, ‘did not get result’ or ‘indeterminate’). The analytical sample for our analysis is limited to respondents who reported that they had never been tested for HIV, that their last HIV test result was negative or indeterminate, or that they did not get the results of their last test. Participants were classified as having unrecognised infection if they had a confirmed positive HIV test result in the study.
Covariates
The covariates examined here were chosen based on other research indicating an association with sexual risk behaviour or risk of HIV infection. Participant age was divided into two groups based on a median split (18–24 vs 25–35 years). Education was dichotomised by high school graduation (yes/no). Men were categorised as being gay if they responded being ‘homosexual or gay’ when asked about their sexual orientation. Men who reported consuming more than five alcoholic beverages within a couple of hours in the past 3 months were classified as having engaged in binge drinking. Men who reported using crystal methamphetamine, cocaine, crack, heroin, poppers or ecstasy in the past 3 months were classified as recent illicit substance users. Based on prior research,16 we used a dichotomous variable to indicate whether or not participants were less than 19 years old when they first had anal sex with a man. Men were categorised as having engaged in exchange sex if they reported any anal sex with male sex partners in exchange for drugs, money, a place to stay or other things. Finally, social network size was measured as the number of MSM the participants knew by name and had seen in the past 6 months in their respective recruitment cities.
Analysis
The analytical sample included 723 men between the ages of 18 and 35 years who reported being HIV negative or of unknown serostatus (ie, never tested, did not receive results of last test or indeterminate results). Our intent was to focus on younger MSM; the 18–35 years age range included men who fell below the mean age (37 years) of the full Brothers y Hermanos sample and is consistent with the cut points of prior MSM-focused studies.17 18 In descriptive analysis, we used χ2 statistics for categorical variables and t tests for continuous variables to compare men who reported having an older partner with those who did not in terms of demographics, substance use, sexual history, HIV testing and sexual behaviour variables.
Unadjusted and adjusted logistic regression analyses were conducted to examine the hypotheses. One model examined the odds of engaging in URAI and a second model examined the odds of having unrecognised HIV infection. Variables that reached p≤0.10 in the unadjusted analysis were included in the multivariate model. To account for variation in social network size, this variable was included in the multivariate models regardless of the magnitude of its unadjusted effect. We present unadjusted ORs and adjusted ORs and 95% CIs. We also examined whether having an older partner interacted with participant age and with participant race/ethnicity in predicting URAI and unrecognised HIV infection. These two-way interaction terms were added, one at a time, to the multivariate models, which were modified to include the main effect component if that term had not already qualified for inclusion in the model.
All analyses were conducted with unweighted data, because the intent was to examine associations in the sample rather than develop population estimates. Analyses were performed using SPSS V.17.0.
Results
Among the 723 participants between 18 and 35 years of age, 172 (24%) were of black race and 551 (76%) were of Latino ethnicity. In the full study, the Latino participants were younger on average than the black participants, which resulted in a larger number of Latino men in this analytical sample. The mean age of the analytical sample was 25.1 years. Twenty-four per cent reported having older partners in the past 3 months (27% of black and 23% of Latino MSM, p=0.23). Half of the participants (50.1%) reported any UAI, 35.7% reported UIAI and 29.4% reported URAI in the past 3 months. Thirty-seven men (5.1%) had unrecognised HIV infection (13.4% among black and 2.5% among Latino men). As seen in table 1, men with older partners were more likely to report binge drinking, substance use and exchange sex compared with men who did not have older partners. Men with older partners were more likely to report sexual risk behaviours, including any UAI, any URAI and a larger number of URAI partners. Unrecognised infection was more likely among those with an older partner.
Demographic and behavioural characteristics of 18–35-year-old black and Latino men who have sex with men (MSM), stratified by whether or not participants had male sex partners who were at least 4 years older than themselves, the Brothers y Hermanos Study, 2005–2006 (n=723)
As shown in table 2, unadjusted logistic regression analyses indicated that gay identity, being less than 19 years old at first anal sex with a man, exchange sex and having an older partner were associated (p≤0.10) with increased odds of engaging in URAI with any man in the past 3 months. These variables (as well as social network size) were included in the multivariate model of this behavioural outcome. The adjusted odds of engaging in URAI were higher among men with older partners, those who were gay identified and those who engaged in exchange sex. Having an older partner did not interact with participant age (p=0.22) or with participant race/ethnicity (p=0.85) in predicting URAI.
Unadjusted and adjusted odds of unprotected receptive anal intercourse with male partners among 18–35-year-old black and Latino men who have sex with men (MSM), the Brothers y Hermanos Study, 2005–2006
With respect to unrecognised HIV infection, the results of the unadjusted logistic regression analyses indicated that participant age, black race, not being a high school graduate, not having had an HIV test within the last 12 months, greater number of URAI partners and having an older partner were associated (p≤0.10) with increased odds of unrecognised infection and, therefore, were included in the multivariate model of this outcome (table 3). Network size was also included. The adjusted odds of having unrecognised infection were greater among men who had older sex partners compared with those who did not. Older participant age, black race, no HIV test in the last 12 months and greater number of URAI partners were also associated with unrecognised HIV infection in the adjusted model. Having an older partner did not interact with participant age (p=0.18) or with participant race/ethnicity (p=0.35) in predicting unrecognised infection.
Unadjusted and adjusted odds of unrecognised HIV infection among 18–35-year-old black and Latino men who have sex with men (MSM), the Brothers y Hermanos Study, 2005–2006
Discussion
Our findings add to the growing recognition that age-discordant sexual partnerships may play an important role in heightening the risk for HIV infection among young MSM.2 7 19 A unique contribution of our analysis is the focus on 18–35-year-old sexually active black and Latino MSM, groups that are of particular importance given their high burden of HIV. The 2008 National HIV Behavioral Surveillance System indicates that HIV prevalence was 28% among black MSM and 18% among Hispanic MSM. Even more concerning is the proportion who are unaware of their infection, which was the highest among young (18–30 years) black men (71% of those testing positive) followed by Latino men (63%).20 Partner age discordance may be more common among black MSM compared with other MSM, although evidence has been inconsistent.21 In our sample, 27% of the black and 23% of the Latino MSM reported having partners who were at least 4 years older than themselves in the last 3 months.
We found that men with older partners were more likely to engage in sexual risk behaviour with male partners of any age, even after controlling for gay identity, age at first sex and exchange sex. This finding supports the hypothesis that MSM who have older partners may be engaging in heightened sexual risk with any of their partners. We were able to locate only one other unpublished report that addressed this relationship. That study among black and white MSM ages 15–29 years also found that having older partners was associated with engaging in UAI.22 Importantly, we demonstrated that after statistically controlling for age, race, education, HIV testing and the number of URAI partners, having an older partner was associated with having unrecognised HIV infection. While several of the traditional risk factors also played a significant role in predicting unrecognised infection, the relationship between unrecognised infection and older partner age persisted. Taken as a whole, these findings suggest that there are dual sources of increased risk for HIV acquisition among young MSM who have older partners—namely, through increased prevalence of unprotected sex with partners of any age and greater probability of being exposed to HIV from older partners.
The growing evidence that age-discordant partnerships are a risk factor for HIV transmission suggests that partner selection should be incorporated into HIV prevention messages for young MSM. As proposed by Coburn and Blower19 and Hurt et al,7 prevention programmes should consider moving beyond discussing only traditional risk factors and safer sex and should tailor programmes according to age and serostatus. Young MSM need to know that age discordance of sex partners is a risk factor for HIV infection and understand how this risk factor compares with other risk factors, such as having multiple partners.6 At the same time, such efforts must avoid stereotyping and promoting the misconception that sex without condoms with same-age partners is safe. Additionally, those delivering prevention counselling should recognise that MSM who have older partners may also be engaging in greater sexual risk behaviour more generally and should thus tailor risk-reduction information as appropriate. The findings can also be used to guide the targeting of prevention outreach; for example, as Bingham et al2 suggest, it may be appropriate to focus on the places, such as bars and clubs, where men meet new partners of varying ages. Finally, further exploration of the motives for seeking older sexual partners is warranted. Young MSM in one qualitative study cited stability, emotional maturity and mentorship as reasons to have older partners.23 Sex in exchange for money, drugs or other necessities may also drive this behaviour. Alternatively, partnering choices could be influenced by social and sexual network characteristics that make seeking same-age partners more challenging for certain populations.24 25 Understanding these partner availability dynamics can help shape prevention interventions concerned with partner selection.
The results from this study should be interpreted in the light of several limitations. Latino MSM comprised approximately 75% of the analytical sample, and the percentage with unrecognised infection was lower among Latino MSM than black MSM. We did not find any significant interaction effects involving race/ethnicity or age of participants (18–24 vs 25–35 years). Nevertheless, caution should be used in generalising our findings beyond our sample. The behavioural risk data were based on self-reports, which may be subject to social desirability bias. However, the use of audio computer-assisted self-interview probably lessened the likelihood of response bias. Also, participants may have assessed their partners' age inaccurately. Due to the use of response categories for assessing partner age, our measure for older partners may have misclassified some participants; however, misclassification would be more likely to miss older partners where they exist, rather than count partners as older when they were not. Our analysis did not link specific sexual behaviours with specific partners, thus our findings do not establish that sexual risk behaviour was more likely to occur in the context of age-discordant partnerships compared with age-concordant partnerships. Instead, this analysis focused on identifying a general pattern of behaviour among men with older partners. Finally, these are cross-sectional data, and we are not able to draw conclusions about cause and effect relationships among the variables in our analysis. For example, in terms of unrecognised infection, we cannot rule out the possibility that transmission is occurring in the reverse direction, that is, younger men may be infecting older men. However, because seroprevalence is generally higher among older men compared with younger men20 and our multivariate model included many known risk factors for unrecognised infection, the general pattern is supportive of the interpretation that having older partners may place young men at elevated risk. The possibility that the transmission dynamic may be bidirectional should also be acknowledged. Additional studies with appropriate designs are necessary to answer this question definitively.
We have provided evidence of the relevance of having older partners to both sexual risk behaviours and risk for unrecognised HIV infection in a sample of black and Latino MSM of ages 18–35. Other explanatory variables not included in this analysis, such as attitudes towards risk taking or the power dynamics between sexual partners, may influence both the selection of older partners and willingness to engage in high-risk sexual behaviours. Factors related to the social environment, such as partner availability, characteristics of sexual networks and social norms within networks, may also heighten the risk for infection among young black and Latino MSM. Additional studies are needed to examine these explanatory variables and to test behavioural interventions to reduce HIV infection risk among young MSM.
Key messages
In the USA, black and Latino men who have sex with men (MSM) are disproportionately affected by HIV; differences in sexual risk behaviour do not explain this disparity. We examine whether older partners' age may play a role.
This cross-sectional study examines whether young black and Latino MSM who have older sex partners are more likely to have unrecognised HIV infection compared with those who do not have older partners.
We found that MSM who had older male sex partners were at increased risk of having unrecognised HIV infection.
Heightened HIV risk was associated with sexual risk behaviour with partners of any age as well as possible increased exposure to HIV infection from older partners.
Acknowledgments
The authors thank the entire Brothers y Hermanos research team, the study participants and the collaborating community-based partners who contributed to the design and implementation of the study.
References
Footnotes
Disclaimer: The findings and conclusions of this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
Funding This study was funded through a cooperative agreement from the US Centers for Disease Control and Prevention.
Competing interests None.
Ethics approval The protocol was approved by the Institutional Review Boards at the US Centers for Disease Control and Prevention (protocol #4417) and at each study site.
Provenance and peer review Not commissioned; externally peer reviewed.