PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Drug Alcohol Depend. Author manuscript; available in PMC 2010 June 1.
Published in final edited form as:
PMCID: PMC2751657
NIHMSID: NIHMS99383

Factors Associated with Event-Level Stimulant Use during Sex In a Sample of Older, Low-Income Men Who Have Sex with Men in Los Angeles

Abstract

Prior research shows that stimulant use is consistently associated with high-risk sexual behavior in samples of men who have sex with men (MSM), but few studies have explored factors associated with use of crack or methamphetamine during sex during specific sexual events among older, very low income MSM. This study examined stimulant use during the most recent sexual episodes in a sample of primarily older, very low-income MSM (n=779). Although crack use was more prevalent than methamphetamine use (33% v. 22%), findings suggest that methamphetamine users may be at greater risk for HIV transmission. HIV prevalence was higher among methamphetamine users (49%) than among crack users (24%). Having unprotected sex (OR 2.77, 95% CI 1.46 – 5.26), having sex in a public sex venue (OR 3.63, 95% CI 1.52 – 8.64), having sex with an HIV positive rather than with an HIV negative partner (OR 6.15, 95% CI 2.14 – 17.62), having exchanged sex for money or drugs (OR 4.16, 95% CI 1.78 –9.72), and having a higher number of sexual partners (OR 1.67, 95% CI 1.17 – 2.38) all were associated with increased odds of methamphetamine use during sex. Fewer high risk behaviors were associated with increased odds of using crack during sex. Having unprotected sex was associated with increased odds of crack use during sex only when sex partners were perceived to be HIV negative rather than to be HIV positive or of unknown status. Findings provide observations on associations between stimulant use during sex and risk behaviors that may be important to HIV prevention and drug treatment approaches for urban, older, very poor MSM.

Keywords: methamphetamine, crack, sex, HIV, respondent driven sampling (RDS)

1. INTRODUCTION

Although recent CDC surveillance data show that HIV incidence is decreasing among some segments of the U.S. population such as women, children, and injection drug users, incidence is increasing among men who have sex with men (MSM) (Centers for Disease Control and Prevention, 2007). Among MSM, there is a strong and consistent correlation between reported use of stimulant drugs and HIV incidence (Plankey et al., 2007). MSM who report using stimulant drugs during sex are more likely to report high risk behaviors that make them vulnerable to HIV, such as unprotected intercourse (UAI) (Colfax et al., 2005; Colfax et al., 2004; Drumright et al., 2006). Although studies demonstrating associations between stimulant use and UAI are abundant, less is known about the factors that influence specific sexual episodes or events that involve the use of stimulants during sex. Although some studies have examined the episode-specific relationship between drug use during sex and UAI (Colfax et al., 2004; Drumright et al., 2006), few have studied how event-level associations may apply to older, very low-income MSM, particularly those of minority racial/ethnic status. Although new HIV infections primarily are among younger MSM (Centers for Disease Control and Prevention, 2007), HIV prevalence rates are high among older MSM (Centers for Disease Control and Prevention, 2008) and among very poor MSM (Robertson et al., 2004). Identifying demographic, behavioral and contextual factors associated with event-level stimulant use during sex may assist development of HIV and drug treatment interventions specific to transmission risks among older, low-income MSM of diverse racial/ethnic backgrounds.

Demographic factors thought to be associated with stimulant use include race/ethnicity, age, HIV status, and socioeconomic status. Use of powder cocaine and club drugs such as methamphetamine tends to be associated with White individuals (Grov et al., 2006; Zule et al., 2007) and is less common among older, African American men (Halkitis and Palamar, 2008). In general, methamphetamine users tend to be White, HIV positive, and more likely to engage in high risk sexual behaviors (Copeland and Sorensen, 2001) than users of other drugs. Many studies of methamphetamine and other club drugs focus on younger, higher educated MSM, limiting understanding of factors associated with use of these drugs in older, less educated lower-income MSM. Indigent, marginally housed MSM also have high HIV incidence rates and involvement with stimulant drugs, particularly crack cocaine (Robertson et al., 2004).

Though global and event-level indices describe associations between drugs, sex risks and HIV prevalence, very little information exists to describe the ways these factors may increase risk for HIV transmission in any specific sexual episode in older, very low-income men and men of minority racial and ethnic status. Crack cocaine and methamphetamine-using MSM report high numbers of sexual partners (Campsmith et al., 2000; Molitor et al., 1998) and episodes of exchanging sex for money or drugs (Molitor et al., 1998; Newman et al., 2004). Location of a sexual event, such as a sex club (Halkitis et al., 2001), as well as having sex with an HIV positive or unknown serostatus partner (Colfax et al., 2005), are related to methamphetamine use in particular. Finally, although few studies describe the relationship between type of partner, such as a main or casual partner, and stimulant use during sex among MSM, MSM who have regular boyfriends are more likely to have unprotected sex than MSM who describe themselves as single (Hays et al., 1997).

Measuring drug use and sexual behaviors at the event-level, during a single instance of a behavior, is thought to better represent individuals’ “typical” behaviors and to better assess associations between sex and contextual variables than aggregated data (Leigh and Stall, 1993). Although some studies have examined associations between event-level substance use and unprotected UAI (Colfax et al., 2004; Drumright et al., 2006; Schroder et al., 2007), few have examined event-level correlates of stimulant use during sex in a population of older, very low-income MSM of diverse racial and ethic backgrounds.

This article addresses the following questions: 1) Are there significant associations between demographic variables, such as race, age, HIV status, and homelessness, and the reported use of stimulant drugs while having sex with recent sexual partners; 2) Are there significant associations between behavioral and contextual variables (e.g., number of sexual partners, location of a sexual event, a partner’s HIV status) and the use of stimulant drugs during sex with recent sexual partners; and 3) Do these associations vary by type of stimulant. We hypothesized that event-level methamphetamine use during sex in this sample of low-income MSM would be associated with being White, younger, HIV positive, and with having an HIV positive partner, a high number of sex partners, having sex in a sex club or other public venue, having sex with male partners, and with having sex with a main or primary partner. We hypothesized that event-level crack use during sex would be associated with being older, Black, homeless, and with having a high number of sex partners.

2. METHODS

2.1 Participants

Participants were MSM who participated in the Los Angeles site of the NIDA Sexual Acquisition and Transmission of HIV Cooperative Agreement Program (SATHCAP). NIDA SATHCAP is an international, multi-site cooperative agreement between NIDA and the University of California, Los Angeles; University of Illinois, Chicago; Research Triangle Institute; Yale University and The Biomedical Center in St. Petersburg, Russia. The primary goal of the SATHCAP study was to examine the role of drugs in the sexual diffusion of HIV from high-risk individuals (MSM and drug users) to the general population.

The present sample (n=779) was drawn from a total of 1,646 participants (MSM, heterosexual men, and women) who enrolled in either phase of the cross sectional study at the UCLA site, between September, 2005 and December, 2006 and between November, 2006 and April, 2008. MSM participants were defined as any male who reported having anal intercourse with men in the past six months who may or may not have also reported sexual intercourse with women. All participants in the subsample reported on at least one sexual event in the past six months. Those reporting no sexual events in the last six months were omitted.

2.2 Procedures

All study procedures were overseen by the UCLA Human Subjects Protection Committee. Participants were recruited using respondent-driven sampling (RDS), a variant of chain-referral sampling that delivers incentives to individuals both to participate in the study and to recruit others in their social network who share common characteristics (Abdul-Quader et al., 2006; Heckathorn, 1997). RDS can efficiently recruit large samples of traditionally hard to reach populations such as drug users and MSM (Deiss et al., 2007; Frost et al., 2006; Johnston et al., 2008; Millett et al., 2007; Wang et al., 2007). In this study, a dual “core” group was defined as an individual who was an MSM and/or a drug user. The study also asked core group members to recruit their sexual partners, both males and females, and asked these sexual partners to recruit their partners.

Recruitment began with the enrollment of seeds, that is, MSM and/or drug using individuals who saw a study flyer posted at locations known to be frequented by this population, who phoned the clinic to arrange a consent visit, and who participated in study procedures. Upon completion of all study procedures, staff members instructed participants to give “core” coupons to people they knew who used heroin, methamphetamine, powder cocaine, or crack, or injected any other drugs in the past six months, or to any man they knew who had anal intercourse with another man in the past six months. A separate set of “sex partner” coupons was used to encourage participants to recruit their sex partners, defined as any person with whom they had sex in the past six months. Coupons were coded to link recruits with their recruiters. When recruits redeemed core or sex partner coupons and were deemed eligible to participate in the study, referring recruiters received compensation. Upon completing all study activities, participants were provided with coupons to recruit additional core members and their sexual partners into the study to continue the chain referral recruitment process. All participants were compensated for study participation, regardless of whether or not they became recruiters.

2.3 Measures

All participants completed an audio, self-administered computerized interview (ACASI) that asked about demographics, health and health behavior, HIV risk-behaviors such as drug and alcohol use and sexual partnerships and behaviors, and about structural and geographic factors related to drug use and sexual behaviors. Drug use and sexual risk behaviors were asked both globally and at the event level. Global questions asked if participants ever engaged in certain behaviors and if so, if they had engaged in these behaviors recently, for example, “have you ever used crack during sex,” and, if so, “have you used crack during sex with any sexual partners in the past six months.” Event-level questions asked about specific behaviors in which they engaged with each of their recent sexual partners whose initials they entered into the ACASI program at the start of the event-level questions. For example, “the last time you had sex with [AA], did you use crack/methamphetamine/etc.” Respondents were asked about their last sexual acts with their three most recent partners in order to capture the last sexual act of a variety of partners. If these partners were not a drug injecting partner, a main partner, or a female partner and respondents indicated that they had one or more of these types of partners, they were also asked to provide information on their last sex act with each of these partners. Upon completion of the questionnaires, participants provided biological samples to be tested for HIV and other sexually transmitted infections. HIV testing consisted of an oral fluid sample tested for HIV antibodies (Oraquik, Orasure Technologies, Bethlehem, PA) and blood specimens (10mL each) to confirm HIV infection (Western Blot).

The two dependent variables were defined as being high on methamphetamine or crack during the last sexual event with a sexual partner. Because reported use of other stimulants such as powder cocaine and other club drugs was minimal in this sample, these drugs were not included as dependent variables in this study.

Explanatory variables included the following participant characteristic: 1) Age, derived from a continuous variable and categorized into four levels – younger than 30, 30–39, 40–49, 50 and older; 2) Race/ethnicity – Non-Hispanic Black, Non-Hispanic White, Hispanic, and Other Race; 3) HIV status, determined from oral fluid rapid test and confirmatory Western Blot; 4) Homelessness – participants were asked if they considered themselves to be homeless at any time during the past year; and 5) Number of sex partners in the past six months (treated as continuous and collapsed from the original continuous variable into 5 categories: 1 to 3 partners, 4 to 6, 7 to 10, 11 to 20, and more than 20 partners). We also controlled for study phase (phase 1 or phase 2).

Explanatory variables also included the following event-level variables: 1) Gender of sexual partner; 2) Type of partner – primary or main partner; regular partner, friend, or acquaintance; one-time partner or stranger; or trade partner; 3) Location of sexual venue –someone’s home; a public venue; a crack house or abandoned building; an hourly rate motel; or some other place; 4) Perceived HIV status of sexual partner – negative, positive, or unknown; 5) Exchange of sex for money or drugs during the sexual event; and 6) Unprotected anal or vaginal intercourse with the partner.

2.4 Statistical Analysis

Subjects reported on stimulant use for up to six different sex partners from the previous six months. The outcomes are binary indicators of “methamphetamine use during sex” and “crack use during sex” during the last sexual episode with each of up to 6 different sex partners. We first fit single predictor logistic random intercept models for each outcome variable, methamphetamine use during sex and crack use during sex, to explore the relationship between each explanatory variable and each outcome variable. Predictors included those at the individual-level -- age, race/ethnicity, HIV status, homelessness, and number of sexual partners, and those at the event-level -- partner type, place of sexual event, partner gender, partner HIV status, exchange of sex for money or drugs, and unprotected sex. Next, we fit multivariate models to the data for each outcome variable. We included all predictors that were significant (p<.05) in the single predictor models for either outcome variable. For single-predictor and multivariate models, we fit multi-level random effects logistic regression models to the data using STATA version 10.0 (xtlogit, random effects) to properly accommodate the multiple observations from a single subject (Rabe-Hesketh and Srkondal, 2008). Random effects models are appropriate when observations are correlated, as when we collect information on multiple events for each subject. Further, random effects models can estimate effects for both subject-level and event-level covariates (Rabe-Hesketh and Srkondal, 2008). We also tested for interactions between three variables – partner type, participant HIV status, and perceived partner HIV status -- and condom use to determine the potential moderating role of these variables in the relationship between condom use and stimulant use during sex. We added significant interaction terms (p<.05) into the final multivariate models. Non-significant interactions were not included.

In RDS, inclusion probability weights can be created for each dependent variable analyzed to adjust for possible differential effects of recruitment in order to generalize findings to targeted networks (Heckathorn, 2002). These effects can include participants recruiting a higher or lower number of peers than they have in their networks, or primarily recruiting individuals like themselves. However, several authors describe limitations in RDS weights as they currently are prescribed (Frost et al., 2006; Heimer, 2005; Wang et al., 2007). Specifically, these authors suggest that RDS weighting cannot fully account for bias introduced by non-random selection from personal networks (Heimer, 2005; Wang et al., 2007), and that estimates of network size may be compromised by high variances, the small size of subpopulations, the inability of subjects to estimate the size of their networks, as well as the way in which network questions are asked (Frost et al., 2006). Further, our analysis of the data showed that statistical independence of the seeds from recruits, a key assumption of RDS, was never achieved. Application of RDS weights would have introduced distortion into the data that could not be measured or accounted for by current RDS weighting methods. Therefore, findings in this report are based on unweighted data and the sample is treated as a convenience sample.

Data were missing on each outcome variable for 88 (5%) events and on predictor variables for 252 (13%) events. Exploration of missing data indicated no association between missing outcome or predictor variables and participant demographics. Missing observations appeared to be missing at random and were removed from regression analysis.

3. RESULTS

3.1 Sample Characteristics

All 779 participants were men who reported sex with men only or with both men and women, of whom 312 (40%) reported having sex with both men and women in the six months preceding the study. Participants were 51% non-Hispanic Black, 24% Hispanic, 20% non-Hispanic White and 5% of another race (“Other”) (see Table 1). The mean age was 42 (range 18–68). More than one-half (63%) earned less than $500 in the past month or reported being homeless during the past year (55%). A total of 299 (38%) of the men were HIV infected and of these, 38% were Black, 35% were Hispanic, 22% were White, and 5% were men of other races/ethnicities. Of the HIV-negative men, 59% were Black, 17% were Hispanic, 19% were White and 5% were another race.

Table 1
Demographic, Behavioral, and Contextual Variables by Stimulant Drug

In response to questions about being high on stimulants while having sex with the most recent partners, 22% (n=168) reported methamphetamine use during sex and 33% (n=255) reported using crack during sex (see Table 1). Of the men who reported being high on methamphetamine during sex, most were non-Hispanic White (44%) or Hispanic (30%), and over half (61%) reported having sex with men only in the past six months. Of men reporting use of crack during sex, most were non-Hispanic Black (80%) and over two thirds (68%) reported having sex with both men and women in the past six months. About half (49%) of the men who used methamphetamine during sex were HIV positive, compared to 24% of those who used crack during sex. A greater percentage of men who used crack during sex (73%) than those who used methamphetamine during sex (54%) reported being homeless in the past year.

Study participants reported behaviors on a total of 1,898 specific sexual encounters. The average number of partners reported upon by study participants was 2.2 with a minimum of 1 and a maximum of 5. Participants used methamphetamine during 320 sexual events with their recent partners and crack during 456 events. Out of all reported sexual events during which methamphetamine was used, 279 (87%) of the events were with male sexual partners, 26 (8%) were with female partners, and 9 (3%) were with transgendered (biologically male) partners. Of events during which crack was used, 350 (77%) of the events were with male partners, 90 (20%) were with female partners, and 12 (3%) were with transgendered (biologically male) partners. About half of the events during which methamphetamine or crack were used (53% and 54%) were with partners described as a regular partner, friend, or as an acquaintance. Participants reported that 19% and 24% of sexual events involving methamphetamine and crack use, respectively, were with main or primary partners. About one quarter (24%) of methamphetamine events ad 17% of crack events were with one-time partners or strangers. Most methamphetamine events took place in someone’s home (66%) or at a public sex venue (22%). Crack events took place at a variety of locations, including someone’s home (43%), a public sex venue (16%), a crack house or abandoned building (18%) and at an hourly rate motel (20%). Almost half of all methamphetamine events (43%) and over half of all crack events (58%) were with partners whose HIV status was unknown to study participants. Almost one-third (32%) of methamphetamine events were with partners who participants believed to be HIV positive, compared with 12% of crack events. Participants exchanged sex for money or drugs during 30% of methamphetamine events and 51% of crack events. Almost half (43%) of participants who used methamphetamine during sex and 38% of participants who used crack during sex had unprotected sex with partners. Of methamphetamine events where men had unprotected sex with a partner believed to be HIV positive (n=44), most (86%) of the men were HIV positive themselves. Of crack events during which men had unprotected sex with a partner believed to be HIV positive (n=12), 92% of the men were HIV positive themselves.

3.2 Bivariate Associations of Demographic, Behavioral and Contextual Variables with Use of Stimulants during Sex

In bivariate analyses, being an MSM older than 30 (age groups 30–39, 40–49, and 50 and over) was associated with increased odds of using crack during sex compared with MSM younger than 30 (OR 11.07, 95% CI 2.65 – 46.32; OR 28. 83, 95% CI 7.22 – 115.21; 27.31, 95% CI 6.21 – 120.02). The odds of using methamphetamine during sex were lower for men between 40 and 49 and for men over 49 than for men under 30 (OR .18, 95% CI .05 –.64; OR .17, 95% CI .04 – .71). The odds of White and Hispanic MSM being high on crack during any of their last sexual encounters were much lower than those for non-Hispanic Black MSM (OR .02, 95% CI .006 –.05; OR .03, 95% CI .01 – .07), and the odds of White and Hispanic MSM being high on methamphetamine during sex were much higher than those for non-Hispanic Black MSM (OR 149.73, 95% CI 50.18 – 446.81; OR 22.24, 95% CI 8.17 – 60.57). MSM who reported homelessness in the past year had greater odds of being high on crack during sex compared with men who did not consider themselves to be homeless (OR 9.46, 95% CI 4.83 – 18.54), while being homeless was not significantly associated with using methamphetamine during sex. Being HIV positive was associated with increased odds of using methamphetamine during sex (OR 3.98, 95% CI 1.80 – 8.58) and decreased odds of using crack during sex (OR .13, 95% CI .07 –.26). The odds of using methamphetamine increased multiplicatively as numbers of sexual partners increased (OR 1.72, 95% CI 1.26 – 2.34).

Among event-level variables, having a sexual partner described as a regular partner/friend/acquaintance, compared with a main/primary partner, was associated with increased odds of using crack during sex (OR 1.7, 95% CI 1.06 – 2.75), while the likelihood of using methamphetamine during sex was not associated with any particular type of partner. Having sex in a public sex venue, such as a bar, bathhouse, or other public place, compared with having sex in someone’s home, was associated with higher odds of using methamphetamine during sex (OR 2.12, 95% CI 1.03 – 4.37). Having sex in a crack house or abandoned building was associated with lower odds of using methamphetamine during sex (OR .17, 95% CI .04 –.70) and with higher odds of using crack during sex (OR 5.84, 95% CI 2.85 – 11.95). Sex in an hourly rate motel also was associated with higher odds of using crack during sex (OR 4.07, 95% CI 2.07 – 8.01). Men who perceived that sexual partners were HIV positive had 10 times higher odds of using methamphetamine during sex than those who perceived that their partners were HIV negative (OR 9.52 95% CI 3.96 – 22.84). Men who did not know their partners’ HIV status also had higher odds of using methamphetamine during sex compared with men who perceived that their partners were hiV negative (OR 2.18, 95% CI 1.11 – 4.29). Having unprotected sex was associated with higher odds of using methamphetamine (OR 2.12, 95% CI 1.20 – 3.72). The relationship between unprotected sex and the odds of using crack during sex were dependent upon the participant’s HIV status and on his perception of his partner’s HIV status. Unprotected sex was associated with lower odds of crack use during sex for men who were HIV positive (OR .21, 95% CI .10 – .47) and for those who thought their partner was HIV positive (OR .08, 95% CI .02 – .36) or who didn’t know their partner’s HIV status (OR .29, 95% CI 0.11 – 0.80). Interactions between partner type and condom use were not significant for either drug, and the interaction between participant HIV status and condom use was not significant for methamphetamine use. Finally, exchanging sex for money or drugs was associated with increased odds of using methamphetamine during sex and using crack during sex (OR 3.03, 95% CI 1.55 – 5.92; OR 8.08, 95% CI 4.88 – 13.36). Whether a sexual event was with a male or female sexual partner did not influence the odds of using either drug during the event.

3.3 Multivariate Analysis

In multilevel random effects logistic models for methamphetamine and crack use during sexual events with up to five recent partners, White and Hispanic MSM were much more likely to use methamphetamine during sex than Black men and less likely to use crack during sex than Black men, holding constant other demographic and behavioral factors (see Table 2). Older age was associated with increased odds of reporting crack use and with decreased odds of reporting methamphetamine use during sex, independent of other factors. Reports of being homeless during the past year increased the odds of reporting crack use during sex by a factor of 3.04 but did not increase the odds of using methamphetamine during sex, independent of covariates. Being HIV positive was associated with decreased odds of using crack during sex, but was not associated with increased or decreased odds of using methamphetamine during sex. Having sex in a public sex venue was associated with increased odds of using methamphetamine during sex, while having sex in a crack house or abandoned building was associated with using crack during sex. Participants were more likely to use methamphetamine with HIV positive partners than with HIV negative partners. Exchanging sex for money or drugs was associated with increased odds of reporting methamphetamine or crack use during sex. Having sex without a condom during the sexual event increased the odds of reporting methamphetamine use during the event, but only increased the odds of using crack when a partner’s HIV status was perceived to be negative rather than positive or unknown. The relationship between condom use and the odds of using either drug was not moderated by the participant’s HIV status in the multivariate models.

Table 2
Multivariate Multi-Level Random Effects Logistic Regression Models of Factors Associated with Use of Stimulant Drugs During Sex

4. DISCUSSION

Findings from this study indicate that a higher proportion of MSM in this very-low-income sample report using crack during their last sexual episode (33%) than men who reported using methamphetamine during sex (22%). Yet, to the extent that event-specific behaviors represent typical sexual behavior, it seems the methamphetamine users may face greater risks for encountering HIV than MSM who incorporate crack use with their sexual behavior. Not only was HIV prevalence higher among MSM who used methamphetamine during sex (49%) compared to those who used crack (24%), but having an HIV positive partner, having a high number of partners, and having unprotected sex all were associated with increased odds of using methamphetamine during sex. Unprotected sex was associated with increased odds of crack use during sex only when a partner was thought to be HIV negative. Compounding these risks is the association between use of methamphetamine during sex and exchanging sex for money or drugs and having sex in a public sex venue. These findings are compelling in describing the functional differences for these two stimulants in transacting sexual risks for HIV transmission in active drug users. Together with cohort data showing higher seroincidence for methamphetamine using MSM (Plankey et al., 2007), our findings describe potential behavioral linkages between methamphetamine use and extreme sexual behaviors facilitated by the drug in increasing risks for HIV transmission.

Although observed rates of reported crack and methamphetamine use during sex are concentrated in small subgroups in this sample, the prevalence of HIV for MSM is high, particularly among men who use methamphetamine during sex. The event-level findings emphasize behavioral disinhibition during sex that occurs under the influence of stimulants, methamphetamine in particular, and corresponding high potential for HIV transmission between MSM. Whether crack or methamphetamine facilitate unprotected sexual behavior between MSM and women, thus creating a “behavioral bridge” (Laumann and Youm, 1999) by which HIV is transmitted from MSM to heterosexual networks, requires further study.

Methamphetamine using MSM in this sample reported on sexual episodes with a relatively small number of female partners (n=26) and, of these, unprotected sex was reported during only 12 events. Ten out of these 12 unprotected sexual events were between women believed to be HIV negative or of unknown HIV status and men negative for HIV. However, of MSM who reported using methamphetamine during sex, 38% (n=122) reported that they had sex with a female partner in the last six months despite providing information primarily on events with male partners. Although most (53%) of the men who used methamphetamine identified with being “gay or homosexual,” 33% identified with the term “bisexual” and a small proportion identified as “heterosexual” (5%), “on the down low or messing around with the other team” (4%), or another term (5%) (data not shown). It is not clear whether MSM who used methamphetamine during sex misrepresented the number of women with whom they had sex in the past six months. Nevertheless, the risk for transmission for women via sex with methamphetamine using men in this sample was quite low. In all of the unprotected methamphetamine events reported, only one was between an HIV positive MSM and a woman, Although data from this study suggest that methamphetamine-using MSM may not pose a significant risk to women, further study is needed to assess behaviors of behaviorally bisexual methamphetamine users.

MSM who used crack reported on a higher number (n=90) of sexual events with women and tended to identify as bisexual (42%) or heterosexual (18%) rather than as gay (21%). Although unprotected sex was associated with increased odds of crack use during sex when the MSM perceived a partner (male or female) was HIV negative, only 35 crack events with women were reported as unprotected. As well, none of the unprotected crack events with women was between an HIV positive MSM and a woman believed to be HIV negative or positive and only one was between an HIV positive MSM and a woman of unknown HIV status. Although data from this study suggest that the risks of HIV transmission to women when MSM use either stimulant during sex are weak within this group of older, low-income MSM, further study is needed, perhaps within younger, higher income MSM to assess the potential role for these drugs in the sexual transmission of HIV to female partners.

Associations between sexual risk-taking and stimulant use during sex observed in the episodes of the men’s last sexual episode with recent partners provide evidence for HIV prevention programs to focus on drug-associated sexual transmission risks for HIV in similar groups in Los Angeles. While findings are most relevant to the local group of MSM in Los Angeles, they may also apply to urban communities with similarly established HIV epidemics and primary transmission behaviors involving male-to-male sexual behaviors in the presence of high rates of illicit stimulant use.

Although several studies describe associations between stimulant use and high-risk sexual behaviors in resourced communities of MSM, findings from the present study shed light on event-level HIV risk factors in an understudied group of very poor, urban, older MSM with a remarkably high prevalence of HIV. Due to the high prevalence of HIV and risk behaviors associated with methamphetamine use during sex, epidemiologists and interventionists alike must take care not to overlook study, intervention, and treatment within this group.

These findings are important as they measure drug use at the time of the sexual behavior, rather than correlating measures of recall for overall drug use in the past 6 months with measures of recall for overall sexual behaviors. While we did not ask about sequencing of drug use and sexual behavior, we did query sexual behaviors that occurred under the influence of illicit stimulants. The strategy of not requiring individuals to report on whether stimulant use preceded high risk sex (or the reverse) is based on recognition that drug-associated sexual episodes are experienced as unitary phenomena that cannot be parsed reliably in terms of timing of events. Thus, the focus on event-level data when reporting on associations between drug use during sex and other episode-level risks such as condom use is a strength of the study. Findings based on over 1,500 reported unique sexual episodes provide greater opportunities for correlating factors with event level behaviors than studies of global estimates of these behaviors.

Limitations to this study’s findings include self-report bias, selection bias due to our application of RDS, and small cell sizes that resulted from disproportionate use of crack by older, Black MSM. Bias due to self-reporting is a potential problem in all survey studies, but bias was minimized by our use of audio computer self-administered interviews (ACASI). In addition, our implementation of RDS yielded a sample that was overwhelmingly very poor, which may reflect unique challenges to the sampling method that arise when using a dual core (MSM and/or drug users) to define the referral approach. This study described drug and risk behaviors of a unique group of very low-income urban men who have sex with both men and women, a largely neglected group in the prevention literature. It is not known whether the associations observed in this group would also be seen in younger MSM or MSM of other income strata. Finally, because of the some small cell sizes and corresponding wide confidence intervals, care must be taken in interpreting findings about age and race. Limitations notwithstanding, findings provide an uncommon event-level observation into a largely hidden group of men who face significant risks for both drug dependence and for HIV transmission.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • Abdul-Quader AS, Heckathorn DD, McKnight C, Bramson H, Nemeth C, Sabin K, Gallagher K, Des Jarlais DC. Effectiveness of respondent-driven sampling for recruiting drug users in New York City: findings from a pilot study. J Urban Health. 2006;83:459–476. [PMC free article] [PubMed]
  • Campsmith M, Nakashima A, Jones J. Association between crack cocaine use and high-risk sexual behaviors after HIV diagnosis. J Acquir Immune Defic Syndr. 2000;25:192–198. [PubMed]
  • Centers for Disease Control and Prevention. HIV/AIDS surveillance report, 2005. Vol. 17. Atlanta: U.S. Department of Health and Human Services; 2007. Revised edition.
  • Centers for Disease Control and Prevention. HIV/AIDS among African Americans Fact Sheet. 2008. [Accessed on October 11, 2008]. http://www.cdc.gov/hiv/topics/aa/resources/factsheets/aa.htm.
  • Colfax G, Coates TJ, Husnik MJ, Huang Y, Buchbinder S, Koblin B, Chesney M, Vittinghoff E. Explore Study Team. Longitudinal patterns of methamphetamine, popper (amyl nitrite), and cocaine use and high-risk sexual behavior among a cohort of San Francisco men who have sex with men. J Urban Health. 2005;82:i62–70. [PMC free article] [PubMed]
  • Colfax G, Vittinghoff E, Husnik MJ, Mckirnan D, Buchbinder S, Koblin B, Celum C, Chesney M, Huang Y, Mayer K, Bozeman S, Judson FN, Bryant KJ, Coates TJ. Explore Study Team. Substance use and sexual risk: a participant-and episode-level analysis among a cohort of men who have sex with men. Am J Epidemiol. 2004;159:1002–1012. [PubMed]
  • Copeland AL, Sorensen JL. Differences between methamphetamine users and cocaine users in treatment. Drug Alcohol Depend. 2001;62:91–95. [PubMed]
  • Deiss RG, Brouwer KC, Loza O, Lozada RM, Ramos R, Firestone Cruz MA, Patterson TL, Heckathorn DD, Frost S, Strathdee S. High-risk sexual and drug using behaviors among male injection drug users who have sex with men in 2 Mexico-US border cities. Sex Transm Dis. 2007;35:243–249. [PMC free article] [PubMed]
  • Drumright LN, Little SJ, Strathdee SA, Slymen DJ, Araneta MR, Malcarne VL, Daar ES, Gorbach PM. Unprotected anal intercourse and substance use among men who have sex with men with recent HIV infection. J Acquir Immune Defic Syndr. 2006;43:344–350. [PubMed]
  • Frost SD, Brouwer KC, Firestone Cruz MA, Ramos R, Ramos ME, Lozada RM, Magis-Rodriquez C, Strathdee SA. Respondent-driven sampling of injection drug users in two U.S.-Mexico border cities: recruitment dynamics and impact on estimates of HIV and syphilis prevalence. J Urban Health. 2006;83:i83–97. [PMC free article] [PubMed]
  • Grov C, Bimbi D, Nanin J, Parsons J. Exploring racial and ethnic differences in recreational drug use among gay and bisexual men in New York City and Los Angeles. J Drug Educ. 2006;36:105–123. [PubMed]
  • Halkitis PN, Palamar JJ. Multivariate modeling of club drug use initiation among gay and bisexual men. Subst Use Misuse. 2008;43:871–879. [PubMed]
  • Halkitis PN, Parsons JT, Stirratt MJ. A double epidemic: crystal methamphetamine drug use in relation to HIV transmission among gay men. J Homosex. 2001;41:17–35. [PubMed]
  • Hays RB, Kegeles SM, Coates TJ. Unprotected sex and HIV risk taking among young gay men within boyfriend relationships. AIDS Educ Prev. 1997;9:314–329. [PubMed]
  • Heckathorn D. Respondent driven sampling: a new approach to the study of hidden samples. Soc Probl. 1997;44:174–199.
  • Heckathorn D. Respondent driven sampling II: deriving valid population estimates from chain-referral samples of hidden populations. Soc Probl. 2002;49:11–34.
  • Heimer R. Critical issues and further questions about respondent-driven sampling: comment on Ramirez-Valles, et al (2005) AIDS Behav. 2005;9:403–408. [PubMed]
  • Johnston LG, Khanam R, Reza M, Khan SI, Banu S, Alam MS, Rahman M, Azim T. The effectiveness of respondent driven sampling for recruiting males who have sex with males in Dhaka, Bangladesh. AIDS Behav. 2008;12:294–304. [PubMed]
  • Laumann E, 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. [PubMed]
  • Leigh BC, Stall R. Substance use and risky sexual behavior for exposure to HIV: issues in methodology, interpretation, and prevention. Am Psychol. 1993;48:1035–1045. [PMC free article] [PubMed]
  • Millett GA, Ding H, Lauby J, Flores S, Stueve A, Bingham T, Carballo-Diequez A, Murrill C, Liu KL, Wheeler D, Liau A, Marks G. Circumcision status and HIV infection among Black and Latino men who have sex with men in 3 US cities. J Acquir Immune Defic Syndr. 2007;46:643–650. [PubMed]
  • Molitor F, Truax SR, Ruiz JD, Sun RK. Association of methamphetamine use during sex with risky sexual behaviors and HIV infection among non-injection drug users. West J Med. 1998;168:93–97. [PMC free article] [PubMed]
  • Newman PA, Rhodes F, Weiss RE. Correlates of sex trading among drug-using men who have sex with men. Am J Public Health. 2004;94:1998–2003. [PubMed]
  • Plankey MW, Ostrow DG, Stall R, Cox C, Li X, Peck JA, Jacobson LP. The relationship between methamphetamine and popper use and risk of HIV seroconversion in the multicenter AIDS cohort study. J Acquir Immune Defic Syndr. 2007;45:85–92. [PMC free article] [PubMed]
  • Rabe-Hesketh R, Srkondal A. Multilevel and Longitudinal Modeling Using Stata. 2. College Station; StatCorp LP: 2008.
  • Robertson MJ, Clark RA, Charlebois ED, Tulsky J, Long HL, Bangsberg DR, Moss AR. HIV seroprevalence among homeless and marginally housed adults in San Francisco. Am J Public Health. 2004;94:1207–1217. [PubMed]
  • Schroder KE, Johnson CJ, Wiebe JS. An event-level analysis of condom use as a function of mood, alcohol use, and safer sex negotiations. Arch Sex Behav. 2007 [epub ahead of print] [PubMed]
  • Wang J, Falck RS, Li L, Rahman A, Carlson RG. Respondent-driven sampling in the recruitment of illicit stimulant drug users in a rural setting: findings and technical issues. Addict Behav. 2007;32:924–937. [PubMed]
  • Zule WA, Costenbader E, Coomes CM, Meyer WJ, Jr, Riehman K, Poehlman J, Wechsberg WM. Stimulant use and sexual risk behaviors for HIV in rural North Carolina. J Rural Health. 2007;23(Suppl):73–78. [PubMed]