In the United States, MSM represent the largest HIV transmission category , accounting for 53% of estimated incident infections in 2006 . Among MSM, blacks are disproportionately affected by HIV and AIDS [3–5], and HIV prevalence among black MSM has been found to be as high as 46% in some cities .
A literature review by Millett et al.  presented 12 hypotheses for the disparity in HIV infection between black and other MSM. These hypotheses included factors that affected likelihood of exposure to and acquisition of HIV infection (among HIV-negative MSM) and likelihood of HIV transmission (among HIV-infected MSM; Fig. 1) . Some of these hypotheses are not supported by evidence. For example, black MSM have comparable or lower numbers of sex partners and prevalence of unprotected anal intercourse [6–10]. Other hypotheses are supported by evidence. For example, higher rates of other sexually transmitted infections may contribute to racial disparities in HIV by increasing both acquisition and transmissibility of HIV [6,8]. Likewise, HIV-infected black MSM are less likely to be aware of their HIV status; because many persons reduce their risk behaviors after HIV diagnosis , this may contribute to HIV transmission among black MSM.
For several hypotheses described by Millett et al. , there was insufficient or conflicting evidence. Some of these hypotheses relate to the likelihood of being exposed to HIV and susceptibility to acquiring HIV. For instance, in the United States, black men are more likely to be incarcerated , and HIV prevalence is higher among inmates of prisons and jails than in the general population [13,14], yet studies have found no association between incarceration and HIV infection among black MSM . Likewise, differences in HIV status and other characteristics of sex partners (such as age and drug use) may reflect differences in HIV prevalence of one's sexual network and, therefore, affect likelihood of exposure to HIV. In addition, circumcision, which has been shown to reduce HIV acquisition among heterosexual populations in Africa [16–18], is less common among black than white men in the United States .
Other hypotheses relate to the possibility of increased HIV transmission from HIV-infected black MSM to their partners. For instance, use of antiretroviral therapy (ART) reduces viral load and infectiousness , yet is less common among black than white HIV-positive MSM . Because black MSM are more likely to have partners of the same race [7,21–23], increased duration of infectiousness due to lower ART use could contribute to increased HIV prevalence among this population.
The National HIV Behavioral Surveillance System (NHBS) is the largest and most geographically diverse surveillance system to monitor HIV risk among MSM in the United States . We used data from the second round of NHBS among MSM (NHBS-MSM2), conducted during 2008, to assess whether hypotheses described by Millett et al.  related to exposure to, acquisition of, or transmission of HIV may partially explain racial disparities in HIV infection among black and white MSM.
National HIV behavioral surveillance system
NHBS-MSM2 was conducted in 21 metropolitan statistical areas (MSAs), selected based on a high number of people living with AIDS.1 MSM were recruited using venue-based, time-space sampling . Activities included formative research to identify venues and times to recruit MSM ; development of sampling frames of eligible venues and day-time periods; random selection of venues and day-time periods; and recruitment, interviewing, and testing during sampled events.
The eligibility criteria included being male; at least 18 years of age; a resident of the MSA; able to complete the survey in English or Spanish; and able to provide informed consent. Trained interviewers used handheld computers to administer a standardized questionnaire. Anonymous HIV testing was offered to all participants regardless of self-reported HIV infection status. Blood or oral specimens were collected for either conventional laboratory testing or rapid testing in the field followed by laboratory confirmation. Activities for NHBS-MSM2 were approved by the Centers for Disease Control and Prevention Institutional Review Board (IRB) and local IRBs for each participating MSA.
Analysis inclusion criteria
Participants were included in this analysis if they had a completed, valid survey; reported at least one male sex partner in the past 12 months; had a positive or negative HIV test result, and reported being either black or white. MSM were considered white if they indicated they were not Hispanic and only selected ‘White’ to define their race. MSM were considered black if they selected ‘Black or African–American’ to define their race alone or in combination with any other racial or ethnic category. Data show that persons who select black and additional races or black race and Hispanic ethnicity have HIV prevalence and HIV risk behaviors similar to those who identify as black/African–American only . Reanalysis of our data after exclusion of participants that selected additional races or ethnicities did not substantially alter findings from this analysis.
Exposure and acquisition analysis
We limited our analysis of risk factors for HIV exposure and acquisition to MSM who were not previously diagnosed with HIV infection. Hypothesis-related predictor variables assessed were incarceration during the past 12 months (limited to participants incarcerated for >1 day at last incarceration); HIV status of last male partner; age of last male partner; last male partner probably or definitely had concurrent relationships; last male partner known to be at increased risk (ever imprisoned, used crack, or injected drugs); having an exchange partner during the past 12 months (i.e., giving something like money or drugs in exchange for sex); and circumcision status.
We determined prevalence of these factors among black and white MSM with newly diagnosed infection (those who first tested positive during NHBS). Additionally, to assess differences between participants with newly diagnosed HIV infection and HIV-negative participants, we used logistic regression . A univariable logistic regression model was fit for each independent variable to determine the unadjusted association with the outcome. All variables were included in a multivariable logistic regression model to determine associations with the outcome after controlling for all other covariates. Because the importance of partner's age can be expected to vary with the respondent's age, an interaction term for respondent's age and partner's age was included. We controlled for number of unprotected anal sex partners and whether the participant had been recently tested for HIV infection. Other control variables included age, education, income, and injection drug use. To determine whether associations were influenced by race, we assessed for interactions between race and all hypothesis variables. A factor was considered to potentially explain a portion of the disparity if the factor was associated with HIV infection, and there was an interaction indicating that the effect of that factor was stronger for black than white MSM; or the factor was associated with HIV infection, and there was not an interaction, but the factor was more prevalent among black than white MSM.
To assess duration of infectiousness, we first assessed factors that might influence the duration of time from HIV acquisition to diagnosis among MSM with newly diagnosed HIV infection. The variables evaluated were health insurance; seeing a healthcare provider during the past 12 months; being offered an HIV test by a healthcare provider during the past 12 months; being tested for HIV during the past 12 months; the number of tests received during the past 2 years; and receiving all test results. Next, to assess duration of infectiousness among previously diagnosed HIV-positive MSM, we assessed factors that might influence time from HIV diagnosis to viral suppression on ART, including health insurance; seeing a healthcare provider for HIV infection within 3 months of diagnosis (necessary for timely initiation of ART); seeing a healthcare provider for HIV during the past 6 months (necessary for continued provision of ART); and being on ART. We also tested associations between various factors and current ART use using a multivariable model. After fitting a univariable model for current ART use and race, we explored the changes in this effect after additional covariates were added to the model.
For bivariate analyses, we used Pearson χ 2 to test for differences between groups. Responses to some covariates were missing, and more black than white participants had missing responses (8.5 vs. 5.7%, P < 0.0001). As a result, for all multivariate logistic regression analyses, we used a multiple imputation approach to account for the uncertainty in these missing responses . Ten imputations were analyzed. Adaptive rounding was used for dichotomous and categorical variables . Model fit for each imputation was evaluated with the Hosmer–Lemeshow goodness-of-fit test and diagnostic statistics . Additionally, we explored accounting for clustering at the venue level. Differences between the logistic regression model and one with venue-specific random effects were trivial. Therefore, we present the former. Analyses were completed using the LOGISTIC and MIANALYZE procedures in SAS (version 9.2; SAS Institute Inc., Cary, North Carolina, USA). All tests and confidence intervals (CIs) are two-sided and based on the 5% level of significance.
A total of 28 468 persons were approached for participation at 626 venues; 12 325 (43%) persons were screened for participation, 10 493 (84%) of whom completed the survey with valid responses. Of these, 9355 (89%) consented to HIV testing and had a valid test result (59 of those excluded reported that they were HIV-positive but tested negative), 8166 (87%) of whom reported male–male sex during the past 12 months. Among this group, 5855 were black or white MSM; 4675 (80%) were HIV-negative, 508 (9%) had newly diagnosed HIV infection, and 672 (11%) had previously diagnosed HIV infection.
Median age was 36 years for white MSM and 28 years for black MSM. Black MSM had lower levels of education and income (Table 1). Larger proportions of black MSM identified as bisexual, and fewer reported injection drug use. Similar percentages of black and white MSM (11 vs. 12%) reported that they were HIV-positive at the time of interview, but substantially more black than white MSM had a positive NHBS test result (27 vs. 16%).
Exposure and acquisition analysis
Table 2 presents characteristics of black and white MSM who were newly diagnosed with HIV infection. Notably, fewer black MSM reported having an HIV-positive last male partner or being circumcised, whereas more black MSM reported that they did not know the HIV status of their last partner.
Multivariable analysis demonstrated that the control variables of race, age, education, income, number of unprotected anal sex partners, and not having an HIV test during the past year were associated with newly diagnosed HIV infection (Table 3). Having a last partner who was HIV-positive or of unknown HIV status was associated with HIV infection. Interactions between race and hypothesis variables were not statistically significant.
Factors that might influence duration of time from HIV acquisition to HIV diagnosis were similar between black and white MSM newly diagnosed with HIV infection (Table 4). However, black MSM were significantly more likely to have been offered an HIV test by a provider. Among previously diagnosed HIV-positive MSM, significantly fewer black than white MSM reported having health insurance, being seen by a healthcare provider for HIV infection within 3 months of diagnosis, or being on ART at the time of interview.
We performed additional modeling to understand possible reasons for differences in ART use. Univariate logistic regression demonstrated that black MSM were less likely than white MSM to be on ART [adjusted odds ratio (AOR) = 0.5, 95% CI = 0.4–0.8]. Adding education, health insurance, and whether seen by a provider for HIV care to the model had little effect on the association between race and ART (AOR = 0.6, CI = 0.4–0.8). We did not have CD4 cell count or other clinical measures, but used time since diagnosis as a surrogate for disease progression to the point at which ART would be clinically indicated. Median time since diagnosis was 6 years for black MSM and 10 years for white MSM (Wilcoxon P < 0.0001). Adding time since HIV diagnosis to the model attenuated the association between race and ART use (AOR = 0.7, CI = 0.5–0.9995).
We assessed whether hypotheses related to exposure to, acquisition of, or transmission of HIV may partially explain racial disparities in HIV infection among black and white MSM. In general, we confirmed the findings of a number of smaller or more geographically limited studies. We found that incarceration, several characteristics of sex partners (partner's age, concurrency, and risk behaviors), and circumcision status were not independently associated with HIV infection, suggesting that these factors do not explain the disparity in HIV infection between black and white MSM. We also found that reporting that one's most recent partner was of unknown HIV status was associated with HIV infection, and this characteristic was more common among black MSM. Finally, we found that black MSM who had been previously diagnosed with HIV infection were less likely to be on ART. This, in combination with the high proportion unaware of their HIV infection, likely contributes to increased duration of infectiousness and ongoing HIV transmission among black MSM.
The prevalence of infection within one's sexual network may have greater influence than individual risk behavior on the risk of HIV acquisition among MSM . We found that having a partner known to be HIV-infected or a partner of unknown status was independent risk factor for HIV infection. Black MSM with newly diagnosed infection were less likely than white MSM to report having a partner known to be HIV-infected; a previous meta-analysis found no difference . However, this is limited by the extent to which MSM know their HIV status and openly and honestly communicate with their partners about HIV status. We found, as have others, that black MSM are less likely to know the HIV status of their partners . Moreover, in our analysis, 59% of HIV-infected black MSM and 25% of HIV-infected white MSM were unaware of their HIV status, suggesting that our estimate of the proportion of MSM with HIV-positive partners is artificially low, particularly for black MSM.
Black MSM are more likely to select partners of their own race than other MSM [7,21–23]. As a result, factors that increase the likelihood that HIV-infected black MSM will transmit HIV to their partners are likely to disproportionately increase HIV acquisition among black MSM. We assessed factors associated with duration of infectiousness, including the time from HIV infection to diagnosis and the time from HIV diagnosis to viral suppression on ART. Black MSM are less likely to be aware of their HIV infection than white MSM [4,5,31], and this was the case in our data as well. Similar proportions of black and white MSM with newly diagnosed HIV infection had health insurance, had seen a provider, and had received HIV testing. This suggests that universal recommendations for HIV testing among MSM may not adequately address the disparities in HIV incidence among different populations. Recommending more frequent HIV testing for populations with high HIV incidence, including black MSM, may reduce the time from HIV infection to diagnosis for HIV-infected black MSM and, therefore, reduce the duration of infectiousness.
Several measures suggested an increased duration of infectiousness after HIV diagnosis for black MSM. Black MSM previously diagnosed with HIV infection were significantly less likely to have seen a provider for HIV care within 3 months of their HIV diagnosis, which suggests that they may be less likely to initiate ART in a timely manner; they were also substantially less likely to be on ART than white MSM. Logistic regression analysis demonstrated that the difference in ART use is likely partly explained by time since HIV diagnosis, suggesting that black MSM in our sample, who were more recently infected, may not yet have progressed to stages of infection requiring ART. Nonetheless, regardless of the reason for lower prevalence of ART, this difference may be partially responsible for the disparity in ongoing HIV transmission between black and white MSM, as ART has been shown to reduce viral load, and reduced viral load is associated with markedly diminished risk of transmission to one's partners . Observational data suggest that expanded HIV treatment and resultant decreases in community viral load may lead to decreased HIV incidence at the population level . If these findings prove to be robust, expanded HIV treatment may help to reduce HIV acquisition among black MSM.
This analysis had several limitations. First, because of the sensitive nature of HIV status, some participants who had previously been diagnosed with HIV infection may not have reported their positive HIV status, resulting in our analysis considering them newly diagnosed when they were not. Moreover, all of our survey data were self-reported. This may have been particularly important with respect to partner's characteristics, and our analysis may, therefore, underestimate the extent to which certain partner's characteristics were present. Additionally, because the NHBS-MSM2 questionnaire did not collect partner's race/ethnicity, we were not able to assess the role that partner race/ethnicity may play in racial disparities in HIV infection. Finally, participants were recruited at venues in 21 US cities with high AIDS prevalence and are not representative of all MSM.
Strengths of this analysis include the size and geographic diversity of the sample and the use of venue-based, time-space sampling for recruitment. The breadth of the NHBS questionnaire allowed us to present data on a wide variety of variables related to HIV infection, and the large sample size allowed us to investigate the correlation between these factors and newly diagnosed HIV infection while controlling for possible confounders.
In summary, our analysis demonstrated that partner HIV status and duration of infectiousness may partially explain the disparity in HIV infection between black and white MSM. Efforts to increase the proportion of HIV-infected black MSM who are aware of their infection and encourage open and honest discussions about HIV status between MSM and their partners may decrease HIV transmission among black MSM. Additional studies are needed to determine reasons for the disparity in receipt of ART between black and white MSM and to identify and address barriers to provision of, acceptance of, and adherence to ART. Finally, although they were considered control variables for the purposes of this analysis, differences between black and white MSM with respect to education and income, which are associated with HIV infection, are important in their own right. Evaluation of additional hypotheses for the disparity in HIV infection between black and white MSM, including socioeconomic and cultural factors, is needed .
The authors would like to thank all of the NHBS-MSM2 participants. They would also like to acknowledge Amy Lansky, Damian Denson, and Elizabeth DiNenno for their work on the early analysis and all members of the project sites participating in the National HIV Behavioral Surveillance System.
A.M.O., G.A.M., R.E.W., C.S., I.J.M., P.E.T., and L.M-M. designed the analysis. R.E.W., B.C.L., and A.M.O. analyzed the data. A.M.O., C.S., I.J.M., and R.E.W. drafted the manuscript. All authors critically revised the manuscript.
The National HIV Behavioral Surveillance System is funded by the Centers for Disease Control and Prevention.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
Previous presentations of these data were presented at the International AIDS Conference, Vienna, Austria, 21 July 2010.
1. Centers for Disease Control and Prevention. Trends in HIV/AIDS diagnoses among men who have sex with men: 33 States, 2001–2006. MMWR Morb Mortal Wkly Rep
2. Hall HI, Song R, Rhodes P, Prejean J, An Q, Lee LM, et al
. Estimation of HIV incidence in the United States. JAMA 2008; 300:520–529.
3. Centers for Disease Control and Prevention. Subpopulation estimates from the HIV incidence surveillance system: United States, 2006. MMWR Morb Mortal Wkly Rep
4. Centers for Disease Control and Prevention. Prevalence and awareness of HIV infection among men who have sex with men: 21 cities, United States, 2008. MMWR Morb Mortal Wkly Rep
5. Centers for Disease Control and Prevention. HIV prevalence, unrecognized infection, and HIV testing among men who have sex with men: five U.S. cities, June 2004–April 2005. MMWR Morb Mortal Wkly Rep
6. Millett GA, Peterson JL, Wolitski RJ, Stall R. Greater risk for HIV infection of black men who have sex with men: a critical literature review. Am J Public Health 2006; 96:1007–1019.
7. Berry M, Raymond HF, McFarland W. Same race and older partner selection may explain higher HIV prevalence among black men who have sex with men. AIDS 2007; 21:2349–2350.
8. Millett GA, Flores SA, Peterson JL, Bakeman R. Explaining disparities in HIV infection among black and white men who have sex with men: a meta-analysis of HIV risk behaviors. AIDS 2007; 21:2083–2091.
9. Flores SA, Bakeman R, Millett GA, Peterson JL. HIV risk among bisexually and homosexually active racially diverse young men. Sex Transm Dis 2009; 36:325–329.
10. Harawa NT, Greenland S, Bingham TA, Johnson DF, Cochran SD, Cunningham WE, et al
. Associations of race/ethnicity with HIV prevalence and HIV-related behaviors among young men who have sex with men in 7 urban centers in the United States. J Acquir Immune Defic Syndr 2004; 35:526–536.
11. Marks G, Crepaz N, Senterfitt JW, Janssen RS. Meta-analysis of high-risk sexual behavior in persons aware and unaware they are infected with HIV in the United States: implications for HIV prevention programs. J Acquir Immune Defic Syndr 2005; 39:446–453.
12. West HC. Jail inmates at midyear 2009
, Bureau of Justice Statistics Statistical Tables. Washington DC: US Department of Justice; 2010.
13. Maruschak LM. HIV in prisons, 2007–2008
. Washington, DC: US Department of Justice; 2009.
14. McQuillan GM, Kruszon-Moran D, Granade T, Feldman JW. Seroprevalence of human immunodeficiency virus in the US household population aged 18–49 years: the National Health and Nutrition Examination Surveys, 1999–2006. J Acquir Immune Defic Syndr
15. Wohl AR, Wohl AR, Johnson D, Jordan W, Lu S, Beall G, Currier J, Kerndt PR. High-risk behaviors during incarceration in African-American men treated for HIV at three Los Angeles public medical centers. J Acquir Immune Defic Syndr 2000; 24:386–392.
16. Auvert B, Taljaard D, Lagarde E, Sobngwi-Tambekou J, Sitta R, Puren A. Randomized, controlled intervention trial of male circumcision for reduction of HIV infection risk: the ANRS 1265 trial. PLoS Med 2005; 2:e298.
17. Bailey RC, Moses S, Parker CB, Agot K, Maclean I, Krieger JN, et al
. Male circumcision for HIV prevention in young men in Kisumu, Kenya: a randomised controlled trial. Lancet 2007; 369:643–656.
18. Gray RH, Kigozi G, Serwadda D, Makumbi F, Watya S, Nalugoda F, et al
. Male circumcision for HIV prevention in men in Rakai, Uganda: a randomised trial. Lancet 2007; 369:657–666.
19. Xu F, Markowitz LE, Sternberg MR, Aral SO. Prevalence of circumcision and herpes simplex virus type 2 infection in men in the United States: the National Health and Nutrition Examination Survey (NHANES), 1999–2004. Sex Transm Dis 2007; 34:479–484.
20. Granich R, Crowley S, Vitoria M, Smyth C, Kahn JG, Bennett R, et al. Highly active antiretroviral treatment as prevention of HIV transmission: review of scientific evidence and update. Curr Opin HIV AIDS
21. Bingham TA, Harawa NT, Johnson DF, Secura GM, MacKellar DA, Valleroy LA. The effect of partner characteristics on HIV infection among African American men who have sex with men in the Young Men's Survey, Los Angeles, 1999–2000. AIDS Educ Prev 2003; 15(1 Suppl A):39–52.
22. Laumann EO, Youm Y. Racial/ethnic group differences in the prevalence of sexually transmitted diseases in the United States: a network explanation. Sex Transm Dis 1999; 26:250–261.
23. Raymond HF, McFarland W. Racial mixing and HIV risk among men who have sex with men. AIDS Behav 2009; 13:630–637.
24. Gallagher KM, Sullivan PS, Lansky A, Onorato IM. Behavioral surveillance among people at risk for HIV infection in the U.S.: the National HIV Behavioral Surveillance System. Public Health Rep 2007; 122(Suppl 1):32–38.
25. MacKellar DA, Gallagher KM, Finlayson T, Sanchez T, Lansky A, Sullivan PS. 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.
26. Allen DR, Finlayson T, Abdul-Quader A, Lansky A. The role of formative research in the National HIV Behavioral Surveillance System. Public Health Rep 2009; 124:26–33.
27. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. New York: John Wiley & Sons; 2000.
28. Bernaards CA, Belin TR, Schafer JL. Robustness of a multivariate normal approximation for imputation of incomplete binary data. Stat Med 2007; 26:1368–1382.
29. Morris M, Goodreau S, Moody J. Sexual networks, concurrency, and STD/HIV.
In: Holmes KK, Sparling P, Stamm W, Piot P, Wasserheit J, Corey L, et al
., editors. Sexually transmitted diseases.
New York: McGraw-Hill; 2007. pp. 109–126.
30. Eaton LA, Kalichman SC, Cherry C, Sexual partner selection and HIV risk reduction among black and white men who have sex with men. Am J Public Health
31. MacKellar DA, Valleroy LA, Secura GM, Behel S, Bingham T, Celentano DD, et al
. Unrecognized HIV infection, risk behaviors, and perceptions of risk among young men who have sex with men: opportunities for advancing HIV prevention in the third decade of HIV/AIDS. J Acquir Immune Defic Syndr 2005; 38:603–614.
32. Das M, Chu PL, Santos GM, Scheer S, Vittinghoff E, McFarland W, Colfax GN, et al. Decreases in community viral load are accompanied by reductions in new HIV infections in San Francisco. PLoS One
33. Kraut-Becher J, Eisenberg M, Voytek C, Brown T, Metzger DS, Aral S. Examining racial disparities in HIV: lessons from sexually transmitted infections research. J Acquir Immune Defic Syndr 2008; 47(Suppl 1):S20–S27.