Individual sexual risk taking has traditionally been measured by questions that summarize recent sexual experience across all sex partners by gender, HIV serostatus, and frequency of unprotected sex. It is now recognized that more specific inquiries must be made to define risk more precisely and that sexual practices and condom use patterns may vary greatly between partners and partner types and by perceived use of highly active antiretroviral therapy (HAART) by HIV-infected partners.1-4 Collection of partner-specific behavioral data is necessary to describe and interpret risk levels accurately at this level of detail.
The first phase 3 HIV vaccine efficacy trial (VAX004) of a bivalent recombinant envelope subunit vaccine (AIDSVAX B/B; VaxGen, Brisbane, CA) enrolled 5108 HIV-negative men who have sex with men (MSM) and 309 women from 61 sites in North America and Europe during 1998 through 1999. The trial was completed in February 2003 with 84% subject retention,5 which affirms the feasibility of recruiting, enrolling, and retaining high-risk participants in a vaccine trial, as suggested by vaccine preparedness cohort studies.6,7
Concerns have been raised about the feasibility of conducting such trials among high-risk persons,8 because investigators are obligated to provide risk reduction counseling, to reinforce repeatedly that vaccine or placebo assignment is masked, and that the efficacy of the candidate vaccine is unknown.9 Counseling could reduce risk behaviors sufficiently to decrease HIV incidence, and thus reduce the power of the study to ascertain vaccine efficacy. Alternatively, increased risk taking by participants because of assumed protection from the candidate vaccine, leading to higher HIV incidence, could represent social harm from vaccine trial participation.
Using partner-specific behavioral data to define risks precisely, we compared risk behaviors of MSM enrolled in the VAX004 HIV vaccine efficacy trial with those of HIV-negative MSM enrolled in a control group from 5 North American sites to assess the impact of vaccine trial participation on sexual behavior.
HIV-negative vaccine trial participants enrolled in the VAX004 trial and a comparison group of HIV-negative MSM were recruited into the Project Vaccine Substudies Network (VISION) study between January 2001 and June 2002 at 5 of the 61 VAX004 sites: Fenway Community Health Center (Boston), Howard Brown Health Center (Chicago), Ohio State University (Columbus), the San Francisco Department of Public Health (San Francisco), and the University of Washington (Seattle). VAX004 participants were recruited during semiannual vaccine trial visits, and 526 (70%) of 750 agreed to participate. More than half (58%) of the 526 VAX004 participants were enrolled in Project VISION at their 12-month visit. The control group of 732 HIV-negative MSM met the same eligibility criteria as VAX004 participants (anal sex with a man in the prior year, not in a monogamous relationship of ≥12 months' duration with an HIV-negative partner, and willing to participate in an HIV vaccine efficacy trial) and was recruited by the same staff using the same recruitment strategies as for the VAX004 trial.
At baseline, all participants made 2 visits to the study site. The first included an interviewer-administered questionnaire, pretest counseling, and HIV testing; VAX004 participants also received placebo or vaccine if that visit included immunization. Two weeks later, all participants completed audio computer-assisted self-interviews (A-CASI) about a maximum of 3 sexual partnerships in the past 6 months before receipt of HIV test results and counseling. Participating institution Institutional Review Boards approved the VAX004 and Project VISION protocols, and participants provided written informed consent.
Demographic and Behavioral Measures
Age, race, and education were collected at the original vaccine trial baseline visit for VAX004 participants and at the Project VISION baseline visit for control participants. Age for VAX004 participants was adjusted to the date of enrollment in Project VISION. At Project VISION enrollment, all participants were asked their number of sex partners in the prior 6 months, by gender and HIV serostatus.
Information about a participant's 3 most recent male sex partners in the last 6 months included partner type (primary, occasional, single-time), partner HIV serostatus, specific sexual behaviors (receptive and insertive anal and oral sex with and without use of a condom), discussion of serostatus, and the participant's perception of their HIV-positive partners' use of antiretroviral treatment (ART). Primary partners were defined as “sexual partners, with whom you may or may not live, but to whom you have a strong emotional commitment,” and occasional partners were defined as sexual partners with whom “you had sex 2 or more times in the past, but whom you do not consider as your primary partner.” Partnership characteristics included (1) settings where a partner was first met, categorized as social (eg, work, school, friends), bars, anonymous (eg, Internet, personal advertisements), or public (eg, baths, parks, street); (2) having had first sex within minutes of meeting the partner; and (3) the dates of the first and most recent sexual encounters with that partner.
Analyses were restricted to receptive and insertive unprotected anal intercourse (UAI) with the 3 most recent partners and constructed variables that incorporated partners' HIV serostatus and partnership type. Receptive and insertive UAI was defined by any acts during the entire relationship with this partner and acts performed at the last sexual encounter.
Parametric and nonparametric tests were used as appropriate for comparison of continuous variables, and χ2 tests were used to evaluate differences in categoric variables. Logistic regression was used to evaluate associations between the 2 study cohorts, while controlling for demographic and behavioral covariates. Variables were entered into models if they were associated (P < 0.1) with the outcome in univariate analyses and retained in the model if they were significantly associated (P < 0.05) with the outcome.
VAX004 participants were somewhat older and less racially diverse and reported higher educational levels than controls. Distribution of cities of enrollment differed between VAX004 and control participants. In the partner-specific questions, most (82%) study participants reported at least 2 partners in the 6 months before enrollment, but fewer VAX004 participants described the maximum of 3 partners (57.9% vs. 70.9% of controls). VAX004 participants reported fewer male partners, were more likely to report an HIV-positive male sex partner, and were less likely to report a female sex partner than controls (Table 1; P < 0.05 for each comparison).
VAX004 participants were significantly more likely to describe a primary male partner than controls (52.6% vs. 42.9%) and less likely to describe an occasional partner (42.5% vs. 50.1%) or a single-time partner (65.7% vs. 78.0%; P < 0.01 for each comparison). Although nearly all participants reported discussing their HIV status with their primary partner(s) (99% of VAX004 participants and 97% of controls), many did not discuss their HIV status with their occasional or single-time partners (47.9% of VAX004 participants vs. 59.7% of controls; P < 0.05). Similar proportions of VAX004 participants and controls met a new partner through an anonymous setting (27.0% vs. 29.6%) or in a public venue (33.0% vs. 36.9%). Regardless of the venues where partners were met, more than 40% of all participants reported having sex within minutes of meeting at least 1 sex partner in the prior 6 months (see Table 1).
A greater proportion of VAX004 participants (24.7%) described any HIV-positive partner than did controls (14.1%; P < 0.01), and there was a similar difference between groups in describing an HIV-positive primary partner (16.1% vs. 6.8%; P < 0.01). VAX004 participants also reported relationships of longer duration with HIV-positive primary partners before enrollment than did controls (mean: 51.5 vs. 21.4 months; P < 0.01). Among participants with any HIV-positive partners, 86.7% of VAX004 participants and 73.5% of controls reported ART use by their HIV-positive partner(s). Fewer VAX004 participants than controls reported occasional or single-time partners of unknown HIV serostatus (51.6% vs. 63.2%; P < 0.01).
More than one third (38.6% VAX004 participants and 36.3% controls) of all participants reported ever engaging in receptive UAI with at least 1 of their 3 most recent male partners. VAX004 participants were more likely to report ever having insertive UAI with these partners than controls (46.2% vs. 40.3%; P < 0.05). During their most recent sexual encounters, VAX004 participants were more likely to report insertive UAI than were controls (Table 2; 30.5% vs. 24.1%; P < 0.05). Both groups reported similar levels of receptive UAI during their most recent sexual encounters (19.9% vs. 22.3%).
Approximately 10% of both groups reported insertive UAI with a partner of unknown serostatus (see Table 2), but VAX004 participants were less likely than controls to report receptive UAI with a partner of unknown serostatus (odds ratio [OR] = 0.5, 95% confidence interval [CI]: 0.3 to 0.9). Fewer men in both groups reported insertive UAI with an HIV-positive partner (5.9% and 3.6% of VAX004 and control participants, respectively; P = 0.07). The riskiest exposure, receptive UAI with an HIV-positive partner, was reported by a much smaller minority of both groups; VAX004 participants were somewhat more likely to report this than were controls (2.6% vs. 1.0%; OR = 2.7, 95% CI: 1.0 to 7.9).
Receptive UAI with an HIV-positive partner during the last sexual episode was significantly associated with perceived use of ART by the partner (P < 0.001) but not with having a primary male partner, age, race, or city of recruitment. In a multivariate model controlling for perceived ART use, study arm was no longer significantly associated with receptive UAI with a positive partner (OR = 1.5, 95% CI: −0.6 to 3.9; see Table 2). In final models containing age, VAX004 participants remained less likely to report receptive UAI with an unknown serostatus partner and more likely to report insertive UAI with any partner.
High-risk sexual behaviors were prevalent among HIV-negative MSM participating in the first phase 3 trial of a candidate HIV vaccine and a control group recruited in 5 geographically diverse United States cities. Risk levels of the VAX004 and control participants differed modestly based on traditional measures, such as the number of sexual partners in the prior 6 months or reporting any HIV-positive partner. Partner-specific data based on participants' 3 most recent sexual partners in the prior 6 months confirmed high levels of risk across both groups, however; more than two thirds of participants had a single-time partner, more than one half reported at least 1 partner of unknown HIV status, and 42% reported sex with a new partner within minutes of meeting him. This level of risk taking is consistent with recent reports of sexual behavior among some MSM, although previous studies have not ascertained partner-specific behaviors.10,11
The partner-specific data provided additional insight into risk taking and differences in risk behaviors between the 2 groups. Compared with controls, VAX004 participants were more likely to report having a male primary partner or an HIV-positive primary partner, and those primary partnerships were of longer duration compared with those of controls. During their last sexual encounter, VAX004 participants were more likely than controls to have engaged in receptive UAI with an HIV-positive partner, but that difference was attenuated after adjusting for perceived ART use by their HIV-positive partner. In addition, they were less likely than controls to report receptive UAI with an unknown serostatus partner. These findings suggest more precise calculations of risk among MSM and indicate that sexual behavior with HIV-positive partners is influenced by the partners' perceived ART use, which is more often known and likely a greater consideration in long-term HIV-discordant partnerships.12
Observed behavioral differences between the VAX004 and control participants should be interpreted cautiously. Although more vaccine study participants reported receptive UAI with HIV-positive partners, the proportion of participants reporting these behaviors was low across both groups and, for some men, may have been associated with lower perceived infectiousness of their known HIV-positive partners who were thought to be taking ART.3,4,13 Overall, adverse effects of vaccine trial participation on sexual behavior were not suggested by our findings or by a longitudinal analysis of behavioral data among VAX004 participants from all 61 vaccine study sites.14
Limitations of this study include the initial enrollment of vaccine trial participants before their enrollment in Project VISION. Thus, vaccine trial participants had experienced risk-reduction counseling several times before enrollment in the Project VISION study. Such behavioral interventions may be only modestly effective,15 however, and we were able to measure behaviors during the same period for both groups of men. Also, the study was not able to examine several factors that could affect perception of risk and probability of HIV acquisition, such as perceived and actual viral load in their HIV-positive partners or frequency of UAI with known HIV-positive partners. Finally, the generalizability of these findings must be viewed cautiously, because MSM who agreed to participate in an HIV vaccine efficacy trial and a control cohort who agreed to participate in studies with multiple follow-up visits may not be representative of high-risk HIV-negative MSM.
Results of this study suggest that the behaviors of MSM during a vaccine trial were not markedly different than the contemporary behaviors of controls. Detailed partner-specific risk behavior data have been useful in describing and partially explaining differences between the 2 groups. In particular, the data provide greater insight into behaviors with the highest risk of exposure to HIV-receptive anal sex with HIV-discordant partners12-which seems to be partially mediated by perceived partners' use of ART. These results indicate that HIV vaccine and prevention intervention trials may benefit from ascertainment of detailed partner-specific information so as to understand behavioral differences and longitudinal changes in risk behaviors.
The authors appreciate the contributions of the study staff, recruiters, and project coordinators in the 5 cities, particularly Joseph Picciano, Robert Pickard, Kellie Dyslin, Dr. Jonathan Fuchs, Rose Quinones, Christopher Ahrens, Mark Hite, and James Price; their colleagues at the Centers for Disease Control and Prevention, especially Drs. Tim Mastro, Marta Ackers, and Eleanor McLellan; their colleagues at VaxGen, particularly Drs. Marc Gurwith and Vladimir Popovic; the local community advisory boards for their suggestions about study design and implementation; and the study participants for their time and commitment.
1. Binson D, Woods WJ, Pollack L, et al. Differential HIV risk in bathhouses and public cruising areas. Am J Public Health
2. Picciano JF, Roffman RA, Kalichman SC, et al. A telephone based brief intervention using motivational enhancement to facilitate HIV risk reduction among MSM: a pilot study. AIDS Behav
3. Suarez TP, Kelly JA, Pinkerton SD, et al. Influence of a partner's HIV serostatus, use of highly active antiretroviral therapy, and viral load on perceptions of sexual risk behavior in a community sample of men who have sex with men. J Acquir Immune Defic Syndr
4. Dukers NH, Goudsmit J, de Wit JB, et al. Sexual risk behaviour relates to the virological and immunological improvements during highly active antiretroviral therapy in HIV-1 infection. AIDS
5. Harro CD, Judson FN, Gorse GJ, et al. Recruitment and baseline epidemiologic profile of participants in the first phase 3 HIV vaccine efficacy trial. J Acquir Immune Defic Syndr
6. Seage GR, III, Holte SE, Metzger DS, et al. Are US populations appropriate for trials of human immunodeficiency virus vaccine? The HIVNET Vaccine Preparedness Study. Am J Epidemiol
7. Buchbinder SP, Douglas JM, Jr, McKirnan DJ, et al. Feasibility of human immunodeficiency virus vaccine trials in homosexual men in the United States: risk behavior, seroincidence, and willingness to participate. J Infect Dis
8. Guenter D, Esparza J, Macklin R. Ethical considerations in international HIV vaccine trials: summary of a consultative process conducted by the Joint United Nations Programme on HIV/AIDS (UNAIDS). J Med Ethics
9. Chesney MA, Chambers DB, Kahn JO. Risk behavior for HIV infection in participants in preventive HIV vaccine trials: a cautionary note. J Acquir Immune Defic Syndr Hum Retrovirol
10. Whittington WLH, Collis TK, Dithmer-Schreck D, et al. Sexually transmitted diseases and human immunodeficiency virus-discordant partnerships among men who have sex with men. Clin Infect Dis
11. Stall RD, Hays RB, Waldo CR, et al. The Gay '90s: a review of research in the 1990s on sexual behavior and HIV risk among men who have sex with men. AIDS
12. Vittinghoff E, Douglas JM, Jr, Judson FN, et al. Per-contact risk of human immunodeficiency virus transmission between male sexual partners. Am J Epidemiol
13. Vanable PA, Ostrow DG, McKirnan DJ, et al. Impact of combination therapies on HIV risk perceptions and sexual risk among HIV-positive and HIV-negative gay and bisexual men. Health Psychol
14. Bartholow BN, Buchbinder SP, Celum CL, et al. HIV sexual risk behavior over 36 months of follow-up in the world's first HIV vaccine efficacy trial. J Acquir Immune Defic Syndr
15. Koblin BA, Chesney MA, Coates TJ, for the EXPLORE Study Team. Effects of a behavioural intervention to reduce acquisition of HIV infection among men who have sex with men: the EXPLORE randomised controlled study. Lancet