SINCE THE acquired immunodeficiency syndrome (AIDS) was first recognized in 1981, the epidemiology of this disease in the United States has changed gradually over time. Compared with earlier years, increasing proportions of AIDS and human immunodeficiency virus (HIV) infection cases are currently being diagnosed among drug injectors, women, nonwhites, and those who acquired the disease heterosexually.1 However, with HIV incidence rates as high as 0.7% to 2.4% per year,2,3 men who have sex with men (MSM) accounted for 50% of AIDS cases and 43% of non‐AIDS HIV cases among men reported to the Centers for Disease Control and Prevention (CDC) in 1996.4 As is true for other populations, subgroups of MSM at particular risk for HIV may be less likely to be reached by prevention messages aimed at the broader MSM community.5 These subgroups include MSM who inject drugs, MSM who trade sex for money or drugs (referred to hereafter as hustlers), and non‐gay‐identified MSM.
Drug‐injecting MSM are of particular concern because they may be exposed to HIV both through unprotected sex and the sharing of injection equipment.6–8 Of MSM who were reported with AIDS in 1996, almost 1 in 10 had a history of injection drug use.4 At HIV counseling and testing sites in Denver, HIV seroprevalence has consistently been highest among drug‐injecting MSM. In 1996, seroprevalence among drug‐injecting MSM in Denver was 13.3% compared with 6.7% among MSM who did not inject drugs and 3.8% among non‐MSM drug injectors (Rietmeijer CA, unpublished observation). Although several studies have examined differences in risk behavior between MSM and non‐MSM drug injectors,7–11 few have examined differences between MSM who inject drugs and those who do not.
MSM who engage in sex hustling have potentially higher levels of HIV risk. In a study of 235 actively working male hustlers in Atlanta, Elifson et al. found receptive anal intercourse with a nonpaying partner, serologic evidence of syphilis or hepatitis B, and a history of childhood physical abuse to be associated with HIV seroprevalence.12
Non‐gay‐identified MSM may also have risk reduction needs that are different from those of gay‐identified MSM. Although our earlier work has found that non‐gay‐identified MSM generally have lower levels of sexually related risk practices, they are less likely to have been HIV tested or to have been reached by HIV prevention programs.5,10
However, from these and other studies it remains unclear whether drug‐injecting MSM, hustlers, and non‐gay‐identified MSM represent demographically and behaviorally distinct subgroups of MSM. As we and others have shown previously, injection drug use, sharing of injection equipment, and sex hustling appear to be particularly prevalent among non‐gay‐identified MSM.8–10,13,14 Still, most of the studies to date have focused on comparing only one subgroup of MSM with MSM in general (e.g., non‐gay‐identified vs. gay‐identified MSM), and little is known about the complex overlap and interrelationships between these subpopulations and associated risk behaviors. Yet, a better understanding of these associations is important for the development and implementation of targeted HIV prevention interventions.
For our study, we used data collected from community‐recruited MSM in Denver and Long Beach as part of the Centers for Disease Control and Prevention AIDS Community Demonstration Projects (ACDP) to explore the demographic relationship between these three subgroups of MSM (i.e., drug‐injecting MSM, hustlers, and non‐gay‐identified MSM) and to assess subgroup differences in HIV risk practices.
The ACDP were community‐level behavioral intervention trials using a common protocol among five difficult‐to‐reach populations at high risk for HIV infection (non‐gay‐identified MSM, street‐recruited drug injectors, female partners of male drug injectors, street youth, and female sex workers) in five U.S. cities: Dallas, TX; Denver, CO; Long Beach, CA; New York, NY; and Seattle, WA. The methods used in the ACDP have been described elsewhere.15,16 For purposes of this study, the following methodological details are relevant. To guide and evaluate the ACDP interventions, data were collected in multiple cross‐sectional waves, during which interviewers administered detailed questionnaires to members of the target populations. Using a purposive sampling scheme, we contacted potential respondents in locations known from formative research to attract the target population. For MSM respondents, these locations included gay bars and bath houses, adult video arcades, and outdoor cruising areas. Two interviews were routinely conducted. First, a 5‐minute screening interview was administered to all MSM who reported ever having had anal or oral sex with a man. The screening questionnaire contained questions about demographics; recent HIV high‐risk practices, including oral, anal, and vaginal intercourse in the past 30 days; exchange of sex for money or drugs (hereafter referred to as hustling); and injection drug use during the past 6 months. Second, MSM who reported to have had oral or anal sexual intercourse with a man or to have injected drugs in the past 30 days were then invited to complete an in‐depth interview regarding HIV testing and results, number of sexual partners, and sexual and preventive behaviors, including the use of condoms for vaginal or anal sex with main and occasional partners. Respondents received $1 coupons for the screening interview and $10 coupons for the in‐depth interview.
Each data collection wave lasted approximately 10 weeks, with 2 to 4 weeks between waves. From February 1991 through June 1994, a total of 10 data collection waves was completed. For this study, we analyzed MSM data from the two most recent waves of the ACDP (waves 9 and 10) in Denver and Long Beach, the only two cities and the only time period in which data were collected both from non‐gay‐identified and gay‐identified MSM. During these waves, which spanned 9 months (September 1993 through June 1994), MSM who were intercepted at the above‐described locations in Denver and Long Beach and who reported oral or anal sex with a man in the past year received the screening questionnaire. All qualifying MSM (i.e., MSM who reported oral or anal sex with a man, or having injected drugs in the past 30 days) who did not gay‐identify were invited to answer the in‐depth questionnaire immediately after they completed the screening questionnaire. Although the intervention targeted non‐gay‐identified MSM, the survey offered the opportunity to evaluate and compare the risk behaviors of non‐gay‐identified MSM with those of gay‐identified MSM. Therefore, during the 9‐month period, all qualifying gay‐identified MSM in Denver were also offered the in‐depth instrument; in Long Beach, only a subset of gay‐identified MSM were sampled for the in‐depth instrument. To avoid duplicates, an identifier was constructed from the respondent's birth date, city and state of birth, and race/ethnicity. In the analysis, only data from the first interview for each participant were used. Because few gay‐identified MSM reported sex with a woman, analyses related to vaginal sex and anal sex with women were limited to non‐gay‐identified MSM.
Yates' corrected chi‐square and Student's t test were used in univariate analysis; logistic regression was used to calculate adjusted odds ratios, 95% confidence intervals, and significance. A P value of 0.10 was chosen as a cutoff for entry in the multivariate model.17 In the multivariate analyses, we dichotomized the use of condoms as to whether respondents had used condoms consistently for more than 6 months. Cutoff values for other variables were determined empirically from the data to best distinguish between subpopulations (i.e., age, 33 years; sex partners in past month, 5.
Data From Screening Interview
For the 9‐month study, a total of 1,290 unduplicated responses to the screening questionnaire were available for analysis: 531 (41%) from Denver and 759 (59%) from Long Beach. Of this group of MSM, 417 (32%) were non‐gay‐identified, 86 (7%) had injected drugs in the past 6 months, and 117 (9%) had exchanged sex for money or drugs. Of non‐gay‐identified men, 10% straight‐identified. Of drug‐injecting MSM and hustlers, 19% and 13%, respectively, were straight‐identified.
Compared with respondents in Long Beach, Denver's respondents to the screening interview were somewhat older, more were white, and more had hustled in the 6 months before the interview (Table 1). The proportions of non‐gay‐identified MSM were similar in both cities (34% and 32%, P = NS).
As can be seen from Figure 1, there seemed to be extensive overlap between non‐gay‐identified MSM, those who injected drugs, and hustlers. Of drug‐injecting MSM, 55% (47/86) reported sex hustling, and 40% (47/117) of hustlers reported injection drug use. Of all hustlers and drug‐injecting MSM, 61% (96/156) did not gay‐identify. In contrast, however, only 23% (96/417) of non‐gay‐identified and 7% (60/873) of gay‐identified MSM reported either sex hustling or injection drug use.
To further explore the relationship between these three categories, we created a series of multivariate models in which the three categories were tested separately as the dependent variable against the other two categories as independent variables, while controlling for age, ethnicity, and city of enrollment (Table 2). In all three analyses, the three categories were strongly and independently associated. The highest odds ratio (14.3, P < 0.001) was calculated for the association between injection drug use and sex hustling. Regardless of associations between categories, Hispanic and “other” ethnicity was independently associated with non‐gay‐identification. Hispanic ethnicity was inversely associated with injection drug use. Finally, Denver recruitment was directly associated with sex hustling, and age older than 33 years was inversely associated with sex hustling.
Of all men who responded to the screening questionnaire, 75% reported oral sex with a male partner and 43% reported anal sex with a male partner in the past 30 days. In multiple logistic regression, sex hustling was associated with higher rates of oral sex (adjusted odds ratio [AOR], 5.0; 95% confidence interval [CI]; 2.4‐10.4) and anal sex (AOR 1.5; 95% CI, 1.0‐2.3); non‐gay‐identified MSM had significantly lower rates for oral sex (AOR 0.5; 95% CI, 0.4‐0.7) and anal sex (AOR 0.4; 95% CI, 0.3‐0.5). Among non‐gay‐identified MSM, sex hustling was also associated with having anal sex with a female partner (AOR, 4.0; 95% CI, 1.7‐10.1).
Data From In‐Depth Interview
Detailed information on HIV testing and serostatus, number of partners, and sexual practices, including condom use, was available for the 482 men who had reported anal or oral sex with a man or who had injected drugs in the past 30 days and were thus eligible for the full interview. As can be seen from Table 1, men who completed the in‐depth survey were comparable to men who completed the screening questionnaire except that more of the men who completed the indepth survey were enrolled at the Denver site and more were non‐gay‐identified. This partly reflects the undersampling of gay‐identified MSM at the Long Beach site. Overall, 248 (51%) men in this group were non‐gay‐identified MSM; 73 men (15%) reported sex hustling, and 61 (13%) reported injection drug use in the past 6 months.
In the subsample, 422/482 (87%) reported having had an HIV test; of those, 61 (14%) reported a positive test result. Fewer of the non‐gay‐identified MSM had been tested for HIV infection than had gay‐identified MSM (82% vs. 97%; OR, 0.24; 95% CI, 0.1‐0.5; p < 0.0001). Fewer non‐gay‐identified MSM reported an HIV positive test result than did gay‐identified MSM (8% vs. 20%, OR, 0.4; 95% CI, 0.2‐0.7; P < 0.001); more drug‐injecting MSM reported an HIV‐ positive test result than did those who did not report injection drug use (27% vs. 12%, OR, 2.6; 95% CI, 1.3‐5.1; P < 0.01). Equal numbers of hustlers and nonhustlers reported a positive test result (12% vs. 15%).
The subsample of MSM reported a mean of 3.5 sex partners in the past 30 days (median 2). In multivariate analysis, sex hustling was associated with reporting more than five partners in the past 30 days (OR, 7.7; 95% CI, 3.7‐16.3, P < 0.0001).
Of respondents in the subsample, 135/237 (57%) reported consistent (longer than 6 months) condom use for anal sex with an occasional partner, 67/175 (38%) for anal sex with a main partner, 34/89 (38%) for vaginal sex with an occasional partner, and 21/92 (23%) for vaginal sex with a main partner. In an analysis of mutually exclusive groups (i.e., persons who reported having engaged in only one of the four sexual behaviors), consistent condom use was reported significantly more often by persons who had had anal sex with occasional partners (60%) than by persons who reported anal sex with a main partner (36%, P < 0.001), vaginal sex with occasional partners (33%, P < 0.01), or vaginal sex with a main partner (23%, P < 0.0001).
Finally, we analyzed the association between consistent condom use for longer than 6 months for each of the four sexual behaviors and the following independent variables: age, ethnicity, city of recruitment, non‐gay‐identification, injection drug use, sex hustling, number of partners in the past 30 days, and self‐reported HIV status. As before, the analyses of condom use for vaginal sex were limited to non‐gay‐identified MSM. The bivariate analysis is presented in Table 3. In multivariate analysis, sex hustling remained the only factor negatively associated with consistent condom use for anal sex with male occasional partners; 36% of hustlers reported consistent condom use, compared with 61% of nonhustlers (AOR, 0.4; 95% CI, 0.2‐0.9, P < 0.05). Consistent condom use for vaginal intercourse with occasional partners was associated with black ethnicity (AOR, 8.1; 95% CI, 1.9‐33.4, p < 0.01); sex hustling and injection drug use remained associated with less consistent condom use for vaginal sex with main partners. In fact, none of the hustlers or drug‐injecting MSM reported consistent condom use for vaginal intercourse with a main partner, compared with 26% of the nonhustlers and 25% of MSM who did not inject drugs (OR not calculable, p < 0.05 for both). None of the factors investigated in this study showed association with consistent condom use for anal sex with male main partners in the multivariate analysis.
In this study of MSM recruited in public sex and other high‐risk environments, we found a strong association between drug injection and hustling, and a weaker, yet significant, association between these two risk behaviors and non‐gay‐identification. More important, we found the highest risks for potential HIV transmission to be particularly associated with the exchange of sex for money or drugs. Hustlers in our study had the highest number of sexual partners in the past 30 days, and more of them engaged in anal and oral sex with other men, as well as anal sex with female partners, yet fewer hustlers used condoms during anal sex with occasional male partners or during vaginal sex with female partners.
Injection drug use, sex hustling, and associated high‐risk behaviors have been shown to be particularly prevalent among non‐gay‐identified MSM.8,10,13,14 In fact, Doll and Beeker suggested four contexts in which bisexual behavior is more likely to occur and that are associated with greater HIV risk: sex hustling, injection drug use, sexual identity exploration, and culturally specific gender roles and norms.13 Our study confirms the relationship between sex hustling, injection drug use, and non‐gay‐identification and also shows that sex hustling and injection drug use seem to be highly correlated among MSM, regardless of sexual self‐identification. More than 60% of non‐gay‐identified MSM in our study did not report sex hustling or injection drugs use and did not seem to have higher HIV risks than gay‐identified MSM. Indeed, as has been reported elsewhere, non‐gay‐identified MSM who did not hustle or inject drugs were less likely to have engaged in anal or oral sex in the past 30 days than were gay‐identified men and were less likely to report that they were HIV positive; they reported condom use at similar rates to those for gay‐identified men.18,19
The strong association found in our study between injection drug use and sex hustling suggests a causal relationship between the two. Although our study was essentially descriptive and not designed to make causal inferences, we still want to offer some speculative explanations of our findings, as they may have implications for future research. Recently, Koester et al. demonstrated that non‐sex hustling (including shoplifting and robbery) by drug injectors is mainly an economic response driven by the need to obtain drugs.20 Sex hustling can be seen as an extension of this in situations where a demand for sexual services exist. Indeed, in studies of prostitutes in the Netherlands, De Graaf et al. found that the financing of their addiction was the most important reason for hard‐drug‐using male and female prostitutes to enter prostitution and that for male prostitutes an additional motivation was to abandon the criminal activities by which they had originally financed their habit.21,22 Sex hustling by men may thus represent a pragmatic solution to the need to support a drug habit, especially for those drug users who do not object to same‐sex encounters or whose objections are overcome by the price paid. This simple economic model would also predict that as long as “the price is right,” even heterosexually identified men would engage in same‐sex activities. Indeed, in our study, 19% of drug injectors and 13% of hustlers straight‐identified.
According to this hypothesis then, the high prevalence of injection drug use and hustling by behaviorally bisexual men in this study could be explained by the high prevalence in our recruitment locations of drug injectors who have included same‐sex transactions in their hustling repertoire and are looking for a market in which to sell sex. Alternatively, or in addition, a subgroup of behaviorally bisexual men may have a psychological propensity for high‐risk behaviors, including injection drug use. In this context, Ross and Rosser have pointed to a relationship between non‐gay‐identification and lack of homosexual self‐acceptance, referred to as self‐homophobia.11 Self‐homophobia and a perceived lack of acceptance of same‐sex behaviors by others have also been identified as the only consistent and significant differences in psychosocial variables between gay‐identified and non‐gay‐identified MSM in a recent study by Stokes et al.23 High‐risk sexual behaviors seem to be associated with self‐homophobia among MSM, and, by analogy, (injection) drug use may also be associated with self‐homophobia and thus linked to bisexual behavior (Ross MW, personal communication, November, 1997).
Obviously, these hypotheses are largely speculative, and more research is needed to better understand the underlying psychological and social dynamics among hustlers and drug‐injecting MSM to provide the necessary underpinnings for the development and implementation of effective prevention programs targeting these groups.
Our study has several limitations that may influence the generalizibility of the results. First, the purposive sampling scheme used in the ACDP oversampled non‐gay‐identified MSM. Second, as mentioned, recruitment in public sex locations may have oversampled hustlers, drug‐injecting MSM, and other MSM at increased risk for HIV, hence, the reported prevalence of high‐risk behaviors cannot be generalized outside these locations. Third, data from only two metropolitan areas may not be representative of other cities. However, in our analyses, we did not find significant site‐specific differences. In addition, many of our findings are similar to those from studies of MSM in other cities.6,7,13,14,19,24 Nonetheless, our findings should be interpreted with these limitations in mind and should be used as a starting point for further studies of MSM at highest risk for HIV.
Finally, what may be the implications of our findings for prevention programs? Although HIV testing rates for non‐gay‐identified MSM were somewhat lower than for gay‐identified MSM in our study, most had had an HIV test. HIV counseling and testing services therefore are still an important opportunity for individual prevention efforts aimed at these high‐risk populations. The significance of HIV counseling and testing has recently been confirmed in a controlled study demonstrating a 20% reduction in the incidence of newly acquired sexually transmitted diseases among persons receiving a two‐session, client‐centered counseling intervention, compared with persons receiving standard education about sexually transmitted diseases.25 Although these findings may not be generalizable to our study population, it seems prudent to offer client‐centered, individual interventions to those seeking prevention services.
However, many persons at highest risk for HIV are unlikely to use these services; even if they do, the opportunity for the reinforcement of preventive behaviors is usually small. As a result, the impact of these types of interventions, while effective, is likely to be limited. By contrast, community‐level interventions may be particularly suited to reaching high‐risk populations, as they expose all at‐risk persons to prevention messages regardless of their individual motivation to seek prevention services. Community‐level interventions aim to change the attitudes, norms, values, and environmental factors of a community through health communications, prevention marketing, community mobilization and empowerment, and environmental facilitation of safer behaviors through the provision of appropriate materials.26–28 Experience with these interventions for HIV prevention has been very encouraging,29–31 particularly among populations that are traditionally difficult to reach, such as drug injectors and non‐gay‐identified MSM.32
In conclusion, hustlers and drug‐injecting MSM constitute a small, but important portion of the MSM community. High‐risk behaviors and a relatively high prevalence of HIV put these men and their sex partners at increased risk for transmission or the acquisition of HIV infection. Innovative programs are needed to effectively intervene with this population.
1. Centers for Disease Control and Prevention. Update: Trends in AIDS incidence, deaths, and prevalence: United States, 1996. MMWR 1997; 46:165-173.
2. Holmberg SD. The estimated prevalence and incidence of HIV in 96 large US metropolitan areas. Am J Public Health 1996; 86:642-654.
3. Buchbinder S, Judson F, McKirnan D, Douglas J, Scheer S, MacQueen K. The feasibility of conducting preventive HIV vaccine trials in gay men in the United States. XIth International Conference on AIDS. Vancouver, July, 1996 [abstract Mo.C. 1592].
4. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report, 1997; 8(No. 2):10-30.
5. Goldbaum G, Perdue T, Wolitski R, et al. Differences in risk behavior and sources of AIDS information among gay, bisexual, and straight-identified men who have sex with men. AIDS and Behavior 1998;2:13-21.
6. Mandell W, Vlahov D, Latkin C, Oziemkowska M, Cohn S. Correlates of needle sharing among injection drug users. Am J Pub Health 1994; 84:902-923.
7. Lewis DK, Watters JK. Sexual behavior and sexual identity in male injection drug users. J Acquir Immune Defic Syndr 1994; 7:190-198.
8. Centers for Disease Control and Prevention. HIV risk practices of male injecting-drug users who have sex with men-Dallas, Denver, and Long Beach, 1991-1994. MMWR 1995; 44:767-769.
9. Wolitski RJ, Humfleet GL, Lee JA, Corby NH. HIV risk-related practices of male homosexual and heterosexual injection drug users. VIIIth International Conference on AIDS. Amsterdam, July, 1992 [abstract poC4267].
10. Wolitski RJ. Comparison of HIV risk practices of non-gay-identified and gay-identified men who have sex with men. 121st Annual Meeting of the American Public Health Association. San Francisco, October, 1993.
11. Ross M, Wodak A, Gold J, Miller M. Differences across sexual orientation on HIV risk behaviours in injecting drug users. AIDS Care 1992; 4:139-148.
12. Elifson K, Boles J, Sweat M. Risk factors associated with HIV infection among male prostitutes. Am J Pub Health 1993; 83:79-83.
13. Doll L, Beeker C. Male bisexual behavior and HIV risk in the United States: Synthesis of research with implications for behavioral interventions. AIDS Educ Prev 1996; 8:205-225.
14. McKirnan DJ, Stokes JP, Doll L, Burzette RG. Bisexually active men: Social characteristics and sexual behavior. J Sex Res 1995; 32:65-76.
15. O'Reilly KR, Higgins DL. AIDS community demonstration projects for HIV prevention among hard-to-reach groups. Pub Health Rep 1991; 106:714-720.
16. Centers for Disease Control and Prevention. Community-level prevention of human immunodeficiency virus infection among high-risk populations: The AIDS Community Demonstration Projects. MMWR 1996; 45(RR-6):1-24.
17. Hosmer DW, Lemeshow S. Applied Logistic Regression. New York: John Wiley & Sons, 1989.
18. Center for Disease Control and Prevention. Condom use among men who have sex with men and the relationship to sexual identity-Dallas, 1991. MMWR 1993; 42:13-14.
19. Stokes JP, McKirnan DJ, Burzette RG. Sexual behavior, condom use, disclosure of sexuality, and stability of sexual orientation in bisexual men. J Sex Res 1993; 30:1-10.
20. Koester SK. Copping, running, and paraphernalia laws: contextual variables and needle risk behavior among drug users in Denver. Hum Organization 1994; 53:287-295.
21. De Graaf R, Vanwesenbeeck I, Van Zessen G, Straver CJ, Visser JH. Male prostitutes and safe sex: Different settings, different risks. AIDS Care 1994; 6:277-288.
22. De Graaf R, Vanwesenbeeck I, Van Zessen G, Straver CJ, Vissers JH. Alcohol and drug use in heterosexual and homosexual prostitution, and its relation to protection behaviour. AIDS Care 1995; 7:35-47.
23. Stokes JP, Vanable P, McKirnan DJ. Comparing gay and bisexual men on sexual behavior, condom use, and psychosocial variables related to HIV/AIDS. Arch Sex Behav 1997; 26:383-397.
24. Lever J, Kanouse DE, Rogers WH, Carson S, Hertz R. Behavior patterns and sexual identity of bisexual males. J Sex Res 1992; 29:141-167.
25. Kamb M, Douglas J, Rhodes F, et al: A multi-center, randomized controlled trial evaluating HIV prevention counseling (Project RESPECT): Preliminary results. XIth International Conference on AIDS. Vancouver, July, 1996 [abstract Th.C.4380].
26. Farquhar JW, Fortmann SP, Flora JA, et al. Effects of community wide education on cardiovascular disease risk factors: The Stanford Five-City Project. JAMA 1990; 264:359-365.
27. Luepker RV, Murray DM, Jacobs DR, et al. Community education for cardiovascular disease prevention: risk factor changes in the Minnesota Heart Health Program. Am J Pub Health 1994; 84:1383-1393.
28. Puska P, Nissines A, Tuomilehto J, et al. The community-based strategy to prevent coronary heart disease: Conclusions from the ten years of the North Karelia Project. Ann Rev Pub Health 1985; 6:147-193.
29. Kelly JA, St. Lawrence JS, Diaz YE, et al: HIV risk behavior reduction following intervention with key opinion leaders of population: An experimental analysis. Am J Pub Health 1991; 81:168-171.
30. Kegeles SM, Hays RB, Coates TJ: The Mpowerment project: A community-level HIV prevention intervention for young gay men. Am J Public Health 1996; 86:1129-1136.
31. Rietmeijer CA, Kane SM, Simons PZ, et al. Increasing the use of bleach and condoms among injecting drug users in Denver: Outcomes of a targeted, community-level HIV prevention program. AIDS 1996; 10:291-298.
32. Fishbein M, Johnson WD, AIDS Community Demonstration Projects. The AIDS Community Demonstration Projects: A successful multi-site community-level behavioral intervention. XIth International Conference on AIDS. Vancouver, July, 1996 [abstract Mo.D.580].