Secondary Logo

Share this article on:

HIV Sexual Risk Behavior Over 36 Months of Follow-Up in the World's First HIV Vaccine Efficacy Trial

Bartholow, Bradford N*; Buchbinder, Susan; Celum, Connie; Goli, Vamshidar*; Koblin, Beryl§; Para, Michael; Marmor, Michael; Novak, Richard M#; Mayer, Kenneth**; Creticos, Catherine††; Orozco-Cronin, Patti‡‡; Popovic, Vladimir‡‡; Mastro, Timothy D*for the VISION/VAX004 Study Team

JAIDS Journal of Acquired Immune Deficiency Syndromes: May 1st, 2005 - Volume 39 - Issue 1 - p 90-101
doi: 10.1097/01.qai.0000143600.41363.78
Epidemiology and Social Science

Increased risk behavior among participants in HIV vaccine efficacy trials has been a concern. This study evaluated HIV sexual risk behavior among 5095 HIV-negative men who have sex with men (MSM) and 308 women enrolled in a randomized, double-blind, placebo-controlled efficacy trial of a bivalent rgp120 vaccine at 61 sites, primarily in North America. Sexual risk behavior data were collected at baseline and semiannually for 36 months. Overall, sexual risk behavior did not exceed baseline levels during the trial. Among MSM, younger age (≤30 years), perceived assignment to vaccine, and nonblack race were associated with an increased probability of unprotected anal sex. Among women, unprotected vaginal sex initially decreased but was statistically equivalent to baseline by 24 months, whereas unprotected vaginal sex with HIV-infected partners decreased from baseline, where it remained throughout the trial. HIV sexual risk behavior did not increase among trial participants; however, it was substantial throughout the trial. Consistently high levels of risk behavior and the association of these behaviors to perceived assignment and demographic variables underscore the need for vigilant HIV risk reduction counseling, informed consent, and educational processes in the context of HIV vaccine efficacy trials.

From the *Centers for Disease Control and Prevention, Atlanta, GA; †San Francisco Department of Public Health, San Francisco, CA; ‡University of Washington, Seattle, WA; §New York Blood Center, New York, NY; ∥Ohio State University, Columbus, OH; ¶New York University School of Medicine, New York, NY; #University of Illinois, Chicago, IL; **Fenway Community Health Center, Boston, MA; ††Howard Brown Health Center, Chicago, IL; and ‡‡VaxGen, Inc., Brisbane, CA.

Received for publication January 23, 2004; accepted August 17, 2004.

The primary source of funding for this trial was VaxGen, Inc. Additional funding was provided by the US Centers for Disease Control and Prevention and the US National Institutes of Health.

Reprints: Bradford N. Bartholow, HIV Vaccine Section, Epidemiology Branch, Division of HIV/AIDS Prevention, Center for HIV, STD, and TB Prevention, CDC, 1600 Clifton Road, Mail Stop E-45, Atlanta, GA 30333 (e-mail: bnb1@cdc.gov).

Concerns about the potential risks of HIV vaccine efficacy trial participation have included speculation that participation might increase HIV risk behavior.1 These concerns are based, in part, upon the finding from hepatitis B vaccine efficacy trials that placebo recipients were more likely to become infected with hepatitis B after their final injection, suggesting that some placebo recipients may have assumed they were protected and subsequently increased their risk behavior.2 Just as optimism about the efficacy of highly active antiretroviral therapy has been implicated as a determinant of increased HIV risk among men who have sex with men (MSM),3-8 optimism about vaccine efficacy could hypothetically result in decreased perception of HIV risk and an associated increase in HIV risk behavior and incidence among trial participants.

Data from HIV vaccine preparedness studies have also raised concerns that efficacy trial participation may lead to increased risk behavior among trial participants. These studies have consistently shown that individuals at higher risk are more willing to participate in HIV vaccine efficacy trials than are those at lower risk.9-14 Additionally, a preliminary report indicated that 46% of participants in an efficacy trial gave protection against HIV as a reason for enrolling.15

To date, only limited information about HIV risk behavior within the context of HIV vaccine trials has been available. Chesney et al1 reported possible trial-related increases in risk behavior among phase 1/II HIV vaccine trial participants.1 However, unlike phase 3 efficacy trial participants, phase 1/II trial participants are generally at low risk of infection; thus, behavioral generalizations from phase 1/II trials to phase 3 trials may lack external validity. Furthermore, the risk behavior of phase 1/II trial participants reported by Chesney et al was significantly lower than what would be expected of phase 3 efficacy trial participants, and the observed increases in risk behavior could potentially be explained by regression toward the mean, rather than HIV vaccine trial participation.

Vaccine efficacy trials have not been designed to investigate whether trial participation may increase HIV risk behavior. Comparison groups not receiving an injection of vaccine or placebo have not been included in efficacy trial designs. Without such comparison groups, causal attributions as to the effect of trial participation on HIV risk behavior cannot be made. However, participants' perceptions of vaccine efficacy and assignment to vaccine or placebo arms of a trial could provide useful information for evaluating the association of trial participation and sexual risk behavior. For example, if individuals enroll in trials to gain protection from HIV infection, these individuals may be less likely to decrease their HIV risk behavior over time or may increase their risk behavior if they perceive the vaccine to be efficacious or that they have been assigned to the vaccine arm of a trial.

Perceptions of assignment status and associated behavior change are of concern given the potential for personal harm (eg, becoming infected with HIV) and the potential biasing of efficacy estimates. If participants accurately guess their assignment to placebo or vaccine arms and decrease or increase, respectively, their risk behavior, the calculation of efficacy will be biased.16 Furthermore, increased risk behavior by those who receive vaccine might subject these participants to increased risk for infection, especially with a low-efficacy vaccine.

There is concern beyond efficacy trials, that HIV vaccination programs using moderately efficacious vaccines might lead to increased risk behavior and HIV incidence, increasing the severity of the epidemic.17 Modeling studies suggest that even with extremely effective HIV vaccines, failure to simultaneously implement effective HIV risk behavior interventions may render HIV prevention less effective.17 Because of these ethical, empirical, and public health concerns related to HIV prevention programs and HIV vaccine efficacy trials, it is important to evaluate HIV risk behavior in the context of actual efficacy trials so that any adverse behavioral manifestations may be identified, understood, and prevented in current and future trials and in eventual HIV vaccination programs.

In this paper, we describe HIV sexual risk behavior over 36 months in MSM and women at high risk enrolled in the VaxGen phase 3 efficacy trial of a bivalent rgp120 HIV vaccine. Additionally, we evaluate the relationship of demographic variables, motivation for joining trials, perceived vaccine efficacy, and perceived assignment status to reported sexual risk behavior.

Back to Top | Article Outline

METHODS

Efficacy Trial Protocol

The VAX004 trial was a randomized, double-blind, placebo-controlled efficacy trial of a bivalent rgp120 HIV-1 subtype B vaccine developed using the MN and GNE8 strains of the predominant virus clade circulating in North America. Both the vaccine and the placebo contained 600 μg of alum adjuvant, which is considered to be very reactogenic. The trial was conducted at 61 sites in the United States (n = 57), Canada (n = 3), and the Netherlands (n = 1). Two-thirds of participants were randomly assigned to the vaccine arm and one-third to the placebo arm of the trial. Seven doses of vaccine were administered during the 36-month trial at the baseline, 1-, 6-, 12-, 18-, 24-, and 30-month visits. Participant-centered HIV counseling and testing based upon the Centers for Disease Control and Prevention model18 and trial-related education were provided at the baseline visit and every 6 months throughout the trial. Trial-related education stressed to all participants that the vaccine being tested may not be effective and that they may have received a placebo rather than the vaccine. Additional details about the trial may be found elsewhere.19,20

Back to Top | Article Outline

Participants

MSM and women at risk for HIV infection were enrolled from June 1998 through November 1999. To be eligible, participants must have been 18-60 years of age, HIV seronegative, healthy, and able to provide informed consent. In addition to these general eligibility criteria, women who reported any of the following behaviors were eligible to enroll: smoking “crack,” engaging in sex for drugs or money, having >5 male sex partners, or having a sexual relationship with an HIV-seropositive male partner during the previous 12 months. Eligibility requirements for MSM were having engaged in anal sex with a male partner during the previous 12 months. MSM were excluded from enrollment if they had been involved in a monogamous relationship with an HIV-seronegative partner for ≥12 months. Also excluded were individuals reporting a history of injection drug use any time during the 3 years prior to the baseline visit.

Back to Top | Article Outline

HIV Antibody Testing

Volunteers were screened for HIV-1 antibodies using commercially available, enzyme-linked immunosorbent assay kits approved by the Food and Drug Administration (FDA), and positive results were confirmed using FDA-approved commercially available immunoblot kits. At screening, volunteers determined to be HIV antibody positive were not enrolled in the study and were referred to medical care. Participants found to be HIV-seronegative at screening were followed up using an uninfected protocol while those who became HIV-seropositive during the trial (confirmed by a central laboratory) were followed up using an infected protocol to assess secondary endpoints such as viral load, CD4 cell count, and progression to clinically defined AIDS. The sexual risk behavior data presented in this paper were reported by participants while in the uninfected protocol.

Back to Top | Article Outline

HIV Infection and Retention Status

Because the sexual risk behavior of trial participants who remain uninfected could be substantially different from those who become infected or who were lost to follow-up, a categorical variable was created to indicate those participants who remained uninfected, became infected, and were lost to follow-up.

Back to Top | Article Outline

Structured Interview

A face-to-face questionnaire administered at baseline included questions regarding participant demographics, medical history, sexually transmitted diseases, motivations for joining the trial, and sexual and drug-related risk behavior. Questions about HIV risk behavior and sexually transmitted diseases were asked every 6 months throughout the trial.

Back to Top | Article Outline

Sexual Risk Behavior

Participants were asked about the number of times they had engaged in insertive and receptive anal sex both with and without using condoms during the previous 6 months. These questions were asked for each of 3 partner types: known HIV-seropositive and -seronegative partners and partners of unknown HIV serostatus. Two dependent variables were created using these composite variables: engaging in any unprotected anal sex and unprotected anal sex with known HIV-seropositive partners.

Back to Top | Article Outline

Protection Motivation for Trial Participation

At the baseline visit, participants were asked to report their motivations for joining the trial, using a 5-point Likert-type scale (1 = agree strongly, 2 = agree, 3 = neither agree nor disagree, 4 = disagree, 5 = disagree strongly). This analysis included “protection motivation” (ie, gaining protection from the vaccine) as a reason for joining the trial.

Back to Top | Article Outline

Perceived Vaccine Efficacy

At the baseline visit, participants were asked how effective they thought the vaccine was likely to be (1 = not effective at all, 2 = <10% effective, 3 = 10%-25% effective, 4 = 26%-50% effective, 5 = 51%-75% effective, 6 = 76%-99% effective, 7 = 100% effective, and 8 = “don't know”). For multivariate analyses, this variable was collapsed into 3 categories: 0%-50% effective, 51%-100% effective, and “don't know.”

Back to Top | Article Outline

Perceived Receipt of Vaccine or Placebo

At the 12-, 24-, and 36-month visits, participants were asked whether they thought that they had received the vaccine or the placebo, with response options being vaccine, placebo, or “don't know.” Using the participants' responses at each of these 3 visits, a single variable was constructed indicating that they consistently thought that they had received vaccine; consistently thought that they had received placebo; consistently didn't know; or inconsistently perceived assignment.

Back to Top | Article Outline

Enrollment Time Covariate

Because enrollment spanned 18 months, it is possible that historical events occurring during trial enrollment differentially affected individuals over time. Furthermore, variation in enrollment strategies across time and trial site could also influence HIV risk behavior. Due to these potential confounders, a sequential day of enrollment variable was calculated for each participant to control for the potential effects of history and variation in enrollment procedures. Although enrollment spanned 18 months, there were 360 actual enrollment days; thus the range of the enrollment time covariate was 1-360 days.

Back to Top | Article Outline

Data Analyses

Generalized estimating equations (GEE) were used to model sexual risk behaviors using SAS version 8.02.21 GEE provides a within-subject, repeated-measures procedure for assessing categorical outcome variables that accounts for the correlated nature of repeated observations. Separate models were fit for MSM and women.

Overall trends in risk behavior were first modeled using only demographic, enrollment time, and infection/retention status covariates to maximize sample size and statistical power. Subsequently, demographic variables significantly related to risk behavior were retained and trial-related covariates (ie, perceived vaccine efficacy, perceived assignment status, and protection motivation) were included in a 2nd model to evaluate the association of both demographic and trial-related covariates to HIV risk behavior. This approach was considered optimal because perceived assignment was not assessed until the 12-month visit; thus, participants who became infected or were lost to follow-up prior to the 12-month visit would have been dropped from GEE analyses, resulting in less information to evaluate the relationship between demographic variables and risk behavior.

Back to Top | Article Outline

RESULTS

Participant Characteristics

Enrolled in the trial were 5095 MSM and 308 women at risk for HIV infection. Male and female participants differed with regard to demographic and risk characteristics (Table 1). Men were more likely to be white, college educated, and younger than women. At the baseline visit, men reported having had more male sex partners than did women; however, there were no differences in the proportion of men and women who reported having had HIV-seropositive male sex partners. Women were more likely than men to report having had hepatitis C or chlamydial infection during the past 6 months. No gender differences were observed with regard to recent self-reported hepatitis B or gonorrhea infections. Of the 5403 participants, 3924 men (77%) and 229 women (74%) remained HIV-uninfected, 362 men (7%) and 6 women (2%) were infected with HIV, and 809 men (16%) and 73 women (24%) were lost to follow-up.

TABLE 1

TABLE 1

Back to Top | Article Outline

Efficacy Results

Overall, the rgp120 vaccine did not prevent HIV-1 acquisition. Details of the efficacy analysis can be found elsewhere.20

Back to Top | Article Outline

Perceived Study Arm Assignment and Vaccine Efficacy

Of 4879 participants with data at the 12-month visit, 67% of MSM and 63% of women reported that they did not know to which trial arm they had been assigned; 19% of MSM and 21% of women perceived assignment to vaccine; and 14% of MSM and 16% of women guessed that they had been assigned to placebo. Overall, 46.6% of 4178 participants with data from the 12-, 24-, and 36-month visits consistently stated that they did not know their assignment; 9.5% consistently thought that they had been assigned to the vaccine arm; and 6.0% consistently thought that they had been assigned to the placebo arm. Perceptions were not consistent for the remaining 1585 participants (37.9%) at these 3 visits. Participants did not correctly perceive their assignment status at the 12-, 24-, or 36-month visits as indicated by kappa coefficients of 0.12, 0.11, and 0.09, respectively.

Most trial participants did not guess as to the efficacy of the vaccine; 53% of men and 50% of women stated that they did not know. Women, however, were more optimistic about vaccine efficacy than were men; 21% vs. 13%, respectively, indicated that they thought the vaccine was 76%-100% effective, 17% vs. 16% guessed effectiveness to be 51%-75%, 8% vs. 10% guessed 26%-50%, and 5% vs. 7% guessed efficacy to be <25% (χ2 = 92.4, P ≤ 0.0001).

Back to Top | Article Outline

Unprotected Anal Sex Among Men

Overall, the proportion of men who reported engaging in unprotected anal sex decreased from the baseline to the 36-month visit. However, there were differences in risk behavior among men who remained HIV uninfected, those who became infected, and those who were lost to follow-up (Fig. 1A). Relative to uninfected men, a greater proportion of men who became infected reported engaging in unprotected anal sex (see Table 2, model 1 for odds ratios [ORs]).

TABLE 2

TABLE 2

FIGURE 1

FIGURE 1

The proportion of men remaining uninfected who reported engaging in unprotected anal sex decreased from the baseline to the 6-month visit (OR = 0.75, 95% CI = 0.70-0.81), increased slightly from the 6- to 30-month visits (OR(30 vs. 6 mo) = 1.15, 95% CI = 1.07-1.23), and then returned to the 6-month level at 36 months (OR(36 vs. 6 mo) = 1.05, 95% CI = 0.98-1.13) (Fig. 1A). Among men who became infected while in the trial, fewer reported engaging in unprotected anal sex at the 6-month visit than at baseline (OR = 0.69, 95% CI = 0.54-0.89); however, this behavior returned to the baseline level at 12 months and did not significantly differ from baseline for the remainder of the trial. Men who were lost to follow-up reported a consistent decrease in unprotected anal sex from baseline to 12 months (OR = 0.60, 95% CI = 0.49-0.74) and remained at the 12-month level until being lost to follow-up.

In addition to study visit and HIV infection/retention status, an increased probability of self-reported unprotected anal sex was associated with younger age, nonblack race/ethnicity, and greater educational level (Table 2, model 2). Furthermore, there were significant visit × race, and visit × age interaction effects, indicating that the probability of reporting unprotected anal sex over time was contingent upon race and age (Fig. 2A, B). When trial-related covariates were included in the model, protection motivation and perceived vaccine efficacy were not related to unprotected anal sex; however, perceived assignment to vaccine or placebo was (Table 2). Men who consistently perceived assignment to vaccine or who had inconsistent perceptions of assignment were more likely than men who perceived assignment to placebo to report engaging in unprotected anal sex. Men who consistently stated that they did not have an idea as to their assignment status were no more likely than those who perceived assignment to placebo to report engaging in unprotected anal sex (Fig. 3).

FIGURE 2

FIGURE 2

FIGURE 3

FIGURE 3

Back to Top | Article Outline

Unprotected Anal Sex With HIV-Infected Partners Among Men

The proportion of men reporting unprotected anal sex with HIV-infected partners decreased from the baseline to the 36-month visit (Fig. 1B). However, there were differences in this behavior among men who remained HIV uninfected, those who became infected, and those who were lost to follow-up. Relative to uninfected men and those lost to follow-up, a greater proportion of men who became infected reported this behavior (see Table 3 for OR). Men who remained uninfected and those not retained in the trial reported decreases in unprotected anal sex with HIV-seropositive partners from the baseline to the 6-month visit, where the rates remained throughout the duration of the trial.

Men who became infected reported a decrease in unprotected anal sex with HIV-seropositive partners from the baseline to the 6-month visit (OR = 0.75, 95% CI = 0.59-0.95) and then reported increases from the 6- to 36-month visits (OR = 1.67, 95% CI = 1.13-2.47). Because infected men were discontinued from the uninfected protocol, the n's decreased over time as follows: 0 = 362, 6 = 362, 12 = 291, 18 = 226, 24 = 161, 30 = 98, 36 = 37. Thus, unprotected anal sex with HIV-positive partners reported at later visits represents a decreasing subset of men who became infected late in the trial. These men (ie, those infected at 30 and 36 months) reported increases in unprotected anal sex with HIV-positive partners from the 18- to 36-month visits (OR = 2.5, 95% CI = 1.2-5.0); however, the reported behavior at 36 months was not significantly greater than that reported at baseline (OR = 1.72, 95% CI = 0.82-3.59).

In addition to study visit and HIV infection/retention status, an increased probability of reporting unprotected anal sex with HIV-infected partners was associated with time of enrollment (men enrolled later in the trial were less likely to report this behavior than those enrolled earlier), older age, nonblack race/ethnicity, and a visit × race interaction (Table 3). Similar to the model of unprotected anal sex, this interaction effect indicates that change in unprotected anal sex with HIV-infected partners is contingent on race, with black men reporting a greater decline relative to other racial/ethnic groups (plot not shown). Also, similar to the model of unprotected anal sex, when unprotected anal sex with HIV-infected partners was modeled using trial-related covariates, perceived assignment status was the only trial-related covariate significantly associated with this behavior (Table 3).

TABLE 3

TABLE 3

Back to Top | Article Outline

Unprotected Vaginal Sex Among Women

Because few women were enrolled in the trial, fitting multivariate models with interaction terms was not possible; thus, only multivariate main effects models could be fit. No differences in unprotected vaginal sex or unprotected vaginal sex with HIV-infected partners were observed among women who remained uninfected, those who became infected, and those who were lost to follow-up; therefore, behavioral trends are presented in the aggregate for these groups. Overall, unprotected vaginal sex among women decreased from the baseline to the 12-month visit and then increased from the 12- to 24-month visits (OR = 1.32, 95% CI = 1.10-1.71), returning at 24 months to the level reported at baseline (Fig. 4). Women who enrolled later in the trial were more likely to report unprotected vaginal sex over time than were women who enrolled earlier. Race, age, education, and trial-related covariates were not related to unprotected vaginal sex (Table 4).

TABLE 4

TABLE 4

FIGURE 4

FIGURE 4

Unprotected vaginal sex with known HIV-seropositive male partners decreased from the baseline to the 12-month visit (OR = 0.50, 95% CI = 0.32-0.81) and then remained steady from the 12- to 36-month visits (OR = 0.73, 95% CI = 0.42-1.27). At 36 months, unprotected vaginal sex with HIV-seropositive partners remained below that reported at the baseline visit (OR = 0.42, 95% CI = 0.23-0.75). Women who enrolled later in the trial were less likely to report unprotected vaginal sex with HIV-infected partners than were women who enrolled earlier. Race, age, and education were not related to unprotected vaginal sex with HIV-infected partners; nor were perceived vaccine efficacy or joining the trial to protect oneself from HIV. Women who consistently did not have an idea as to their assignment status were more likely to report unprotected vaginal sex with HIV-infected partners than were women who perceived assignment to placebo (Table 5).

TABLE 5

TABLE 5

Back to Top | Article Outline

DISCUSSION

Participants enrolled in the phase 3 efficacy trial of an rgp120 vaccine containing 2 subtype B antigens (MN and GNE8) reported substantial HIV risk, both at the baseline visit and during 36 months of follow-up, underscoring the critical need for HIV vaccines. Contrary to speculation that participating in HIV vaccine efficacy trials might increase HIV sexual risk behavior,1 sexual risk behavior did not increase beyond that reported at baseline among participants in this trial. The highest-risk behaviors, unprotected anal sex with HIV-infected partners among the men and unprotected vaginal sex with HIV-infected partners among the women, decreased over the duration of the trial. Unprotected anal sex among the men and unprotected vaginal sex among the women initially decreased and later returned to baseline levels but did not exceed baseline levels at any point during the trial.

The reported decrease in unprotected anal sex among MSM in this trial approximates the observed reductions in risk behavior reported in a recent meta-analysis evaluating 1- to 6-month postintervention effects of HIV risk reduction interventions among MSM (OR = 0.69, 95% CI = 0.56-0.86).22 Also, the reduction in unprotected vaginal sex from the baseline to the 12-month visit exceeds that observed in similar meta-analyses of HIV risk reduction interventions among drug users (OR = 0.86, 95% CI = 0.76-0.98)23 and heterosexuals (OR = 0.81, 95% CI = 0.69-0.95).24 Lastly, the overall decrease in unprotected anal sex from the baseline to 36-month visit among all men in the trial (7%) is almost identical to that observed among participants randomized to the enhanced intervention arm (∼10%) of a large-scale randomized controlled behavioral intervention trial that enrolled >4000 MSM in 6 US cities.25 The results from these studies demonstrate that the reductions in HIV risk behavior observed among the participants in this efficacy trial are similar to those observed across multiple HIV risk reduction intervention studies.

However, although risk behavior did not exceed that reported at baseline at any time during follow-up and generally declined over time, both younger MSM and MSM who perceived that they had received vaccine reported levels of unprotected anal sex at follow-up that were statistically indistinguishable from those reported at baseline; older men and men who perceived that they had received placebo or who did not have a clear perception of their assignment did not return to risk levels reported at baseline. The effect of age on risk behavior change over time is not unanticipated because younger age has been associated with greater HIV risk in other studies of MSM.26,27

The MSM who perceived that they received the vaccine consistently reported at baseline and over time a greater amount of unprotected anal sex than did men who did not perceive that they received vaccine which suggests that this may be an individual-level behavioral phenomenon unrelated to trial participation per se. This conclusion is supported by the lack of an observed perceived assignment × visit interaction effect on risk behavior. Thus, the difference in risk behavior between participants who perceived that they received vaccine and those who did not is, on average, about the same over time. If an increased probability of risk behavior were attributable to trial participation, the hypothesis would be that risk behavior would not differ by perceived assignment status at baseline but that participants who perceived that they received vaccine would increase their risk behavior over time relative to those who perceived that they received placebo.

Psychologic theory may help explain why the participants at highest risk were consistently more likely to perceive that they received vaccine than were participants at lower risk. Cognitive dissonance theory28-30 suggests that when a person engages in a behavior that is inconsistent with their beliefs, that person is motivated to reduce the uncomfortable state of arousal (dissonance) that results. Because beliefs are often easier to change than behavior, a dissonance-reducing change in beliefs may be more likely to occur than behavior change. One way that trial participants could reduce dissonance associated with risk behavior is to establish a belief that they have received vaccine and that the vaccine is efficacious. In fact, among these trial participants, perceptions of HIV vaccine efficacy and assignment status were significantly correlated. This phenomenon may be suggestive of additional counseling strategies and underscores the need for continued counseling emphasizing that the product being tested is of unknown efficacy, that participants may receive placebo, that infections have occurred in the trial, and that participants should maintain lower risk behavior.

Some have suggested that individuals whose motivation to participate in efficacy trials is to gain protection from HIV should possibly be excluded from participating.31 Given that about half of the participants in this trial endorsed protection as a reason for participating,15 excluding such individuals from trials could significantly increase the time and cost of enrollment, decrease HIV incidence of trial cohorts, and increase the sample size and follow-up requirements to determine efficacy. Furthermore, excluding such individuals from trials may raise ethical issues regarding fair and equal access to trials.31 The data from this trial indicate that the motivation of protection is not associated with increased risk among trial participants and, thus, would not support excluding such participants from future trials.

Finally, that MSM who became infected later in the trial reported increases in unprotected anal sex with HIV-infected partners from the 18- to 36-month visits may suggest that avoidance of infection up to 18 months of trial participation could potentially result in increasing optimism about vaccine efficacy and increased risk behavior. While this is speculation, the possibility suggests an additional counseling message to consider for future trials. In addition to the messages that low-risk behavior should be maintained during the trial because the vaccine may not be effective and participants may receive placebo, future trials should consider incorporating an additional message to mitigate the potentially negative impact of increasing participant optimism. For example, in addition to the standard messages, participants might also be told that infections have occurred in the trial, that we do not know in which treatment arm the infections have occurred, and that participants should continue safer sex practices. Such a message could serve to counterbalance optimistic perceptions among trial participants not only potentially related to their avoidance of infection, but also to contextual influences such as news reports of successful trial enrollment, Data Safety Monitoring Board (DSMB) reviews, and retention. Care should be taken when delivering such a message, however, as there is a chance that it could affect study compliance and retention.

This study has several limitations. Notably, we were unable to attribute risk behavior change to trial participation because there was no comparison group that did not receive an injection of either vaccine or placebo. Without such a comparison group, interpretation of either upward or downward trends in risk behavior is limited. Similarly, without such a comparison group we cannot make comparative statements about the magnitude of behavior change we observed in this study. For example, it may be that although trial participation did not cause an increase in HIV sexual risk behavior, it could attenuate the declining slope of risk behavior over time. In other words, the rate of decrease for trial participants could be less than that for a similar cohort of participants who do not receive an injection of vaccine or placebo. Unless contrast groups that do not receive an injection are included in future trial designs, quasi-experimental nonequivalent group designs may be the only way to evaluate the relationship of trial participation to HIV risk behavior. Despite the possible confounds associated with these designs,32 comparing vaccine efficacy trial cohorts to those enrolled using similar eligibility criteria could provide convergent validity as to the relationship between trial participation and HIV sexual risk behavior. The Centers for Disease Control and Prevention has enrolled such a contrast group at a subset of VaxGen trial sites and results are expected in the near future.

We were also limited with regard to sample size of both minority MSM and female participants. The small sample of women constrained our ability to calculate parameter estimates with reasonable CIs and to evaluate interaction effects related to unprotected vaginal sex and perceived vaccine efficacy and assignment over time. Also, the small sample of women participating in the trial limited our ability to evaluate heterogeneity among these women with regard to motivations to participate in trials, risk behavior, education, and trial-related perceptions of vaccine efficacy and assignment. In general, these same limitations apply to our ability to understand the risk behavior dynamics of people in high-risk communities of color who will be recruited for enrollment in future HIV vaccine efficacy trials.

Our ability to assess changing perceptions of vaccine efficacy and perceived assignment was limited by infrequent measures. It was not possible to evaluate how changing perceptions may be related to risk behavior change over time. Future trials should measure these perceptions at each visit to assess and monitor changing perceptions and associated risk behavior change.

Despite limitations, this study provides the first descriptive data on HIV sexual risk behavior among phase 3 HIV vaccine efficacy trial participants. These data may alleviate some concerns about increasing HIV sexual risk behavior by trial participants; however, they also suggest areas for future evaluation. There is a need for additional analyses to evaluate how substance use may vary as a result of trial participation and to identify mediators and moderators of risk behavior of trial participants. It will also be important to assess perceived assignment to vaccine or placebo and the extent to which these perceptions may lead to unblinding and differential risk behavior change. Finally, despite the methodologic challenges, quasi-experimental nonequivalent control group analyses, such as that being conducted by the Centers for Disease Control and Prevention, will provide a more complete picture of HIV sexual risk behavior within the context of HIV vaccine efficacy trials.

Back to Top | Article Outline

ACKNOWLEDGMENTS

The authors thank the study coordinators at each of the trial sites for their perseverance and dedication to this trial. Furthermore, the authors thank Marta Ackers, Alan Greenberg, John Williamson, Eleanor McLellan, Markus Durham, Bill Heyward, and Marc Gurwith for their support and review of the manuscript and Marlene Chernow, Gina Rossen, John Jermano, and Nzeera Ketter for their critical contributions to the design and implementation of the trial. Most importantly, the authors thank the VAX004 volunteers for their participation and commitment.

Back to Top | Article Outline

VAX004 Study Group Members and Affiliations

Franklyn Judson, MD, Denver Department of Public Health, Denver, CO; Geoffrey J. Gorse, MD, Saint Louis University, St. Louis, MO; Clayton D. Harro, MD, Johns Hopkins School of Public Health, Baltimore, MD; Kenneth Mayer, MD, Fenway Community Health, Boston, MA; Jay Kostman, MD, Philadelphia Fight, Philadelphia, PA; Harold Kessler, MD, Chicago Center for Clinical Research, Chicago, IL; Stephen J. Brown, MD, AIDS Research Alliance, West Hollywood, CA; Richard DiCarlo, MD, Louisiana State University Health Sciences Center, New Orleans, LA; Michael C. Keefer, MD, University of Rochester, Rochester, NY; Bienvenido G. Yangco, MD, Infectious Disease Research Institute, Inc., Tampa, FL; Javier O. Morales, MD, Clinical Research Puerto Rico, Inc., San Juan, PR; Donald Forthal, MD, University of CA-Irvine Medical Center, Orange, CA; Susan Buchbinder, MD, San Francisco Department of Public Health, San Francisco, CA; Connie Celum, MD, University of Washington/Seattle HPTU, Seattle, WA; Keith Henry, MD, Hennepin County Medical Center, Minneapolis, MN; Jeffrey M. Jacobson, MD, Beth Israel Medical Center, New York, NY; Jerry Cade, MD, Wellness Center, Las Vegas, NV; Elaine Thomas, MD, University of New Mexico Health Sciences Center, Albuquerque, NM; Richard Novak, MD, University of Illinois at Chicago, Chicago, IL; Frank Rhame, MD, Abbott Northwestern Hospital, Minneapolis, MN; Ralph W. Richter, MD, Clinical Pharmaceutical Trials, Inc., Tulsa, OK; Michael C. Caldwell, MD, Dutchess County Department of Health, Poughkeepsie, NY; Beryl Koblin, PhD, New York Blood Center, New York, NY; Michael Marmor, PhD, NYU School of Medicine, New York, NY; Catherine Creticos, MD, Howard Brown Health Center, Chicago, IL; Michael Para, MD, Ohio State University, Columbus, OH; Stephen K. Tyring, MD, UTMB Center for Clinical Studies, Houston, TX; David McKinsey, MD, Antibiotic Research Associates, Kansas City, MO; Neil Flynn, MD, UC Davis Medical Center, Infectious Diseases, Sacramento, CA; Edwin DeJesus, MD, IDC Research Initiative, Altamonte Springs, FL; Steven Santiago, MD, Care Resource, Inc., Coral Gables, FL; Barry M. Miskin, MD, Palm Beach Research Center, West Palm Beach, FL; Cecilia Shikuma, MD, Hawaii AIDS Clinical Trials Unit, Honolulu, HI; Melanie Thompson, MD, AIDS Research Consortium of Atlanta, Atlanta, GA; James O. Kahn, MD, UCSF-Positive Health Program, San Francisco, CA; Ross G. Hewitt, MD, Erie County Medical Center, Buffalo, NY; Michael S. Somero, MD, Office of Michael S. Somero, MD, Palm Springs, CA; Roel A. Coutinho, MD, GG&GD/Municipal Health Service Amsterdam, Amsterdam, NL; Martin Fenstersheib, MD, Crane Center, San Jose, CA; M. Angeli Adamczyk, MD, ACRC/Arizona Clinical Research Center, Inc., Tucson, AZ; Peter Piliero, MD, Albany Medical College, Albany, NY; Ronald Poblete, MD, North Jersey Community Research Initiative, Newark, NJ; Michael Sands, MD, University of Florida-Jacksonville, Jacksonville, FL; Barbara Gripshover, MD, University Hospitals of Cleveland, Cleveland, OH; David Brand, MD, North Texas Center for AIDS & Clinical Research, Dallas, TX; Patrick Daly, MD, Nelson-Tebedo Health Resource Center, Dallas, TX; Mark J. Mulligan, MD, University of Alabama at Birmingham, Birmingham, AL; Robert L. Baker, MD, Community Medical Research Institute, Indianapolis, IN; Peter S. Vrooman, Jr., MD, ALL TRIALS Clinical Research, Winston Salem, NC; Robert Hogg PhD, BC Center for Excellence in HIV/AIDS, Vancouver, BC, Canada; Jean Vincelette, MD, Hopital Saint-Luc du CHUM, Montreal, PQ, Canada; Brian P. Buggy, MD, Wisconsin AIDS Research Consortium, Milwaukee, WI; James H. Sampson, MD, Research & Education Group, Portland, OR; Sharon Riddler, MD, University of Pittsburgh, Pittsburgh, PA; Robert A. Myers, MD, Body Positive, Inc., Phoenix, AZ; Michelle Lally, MD, The Miriam Hospital, Providence, RI; Joseph Jemsek, MD, Jemsek Clinic, Huntersville, NC; Howard Grossman, MD, Polaris Medical Group, New York, NY; Ken Logue, MD, CascAids Research, Toronto, ON, Canada.

Back to Top | Article Outline

REFERENCES

1. 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. 1997;16:266-271.
2. Szmuness W, Stevens CE, Harley EJ, et al. Hepatitis B vaccine: demonstration of efficacy in a controlled clinical trial in a high-risk population in the United States. N Engl J Med. 1980;303:833-841.
3. Katz MH, Schwarcz SK, Kellogg TA, et al. Impact of highly active antiretroviral treatment on HIV seroincidence among men who have sex with men: San Francisco. Am J Public Health. 2002;92:388-394.
4. Kelly JA, Hoffman RG, Rompa D, et al. Protease inhibitor combination therapies and perceptions of gay men regarding AIDS severity and the need to maintain safer sex. AIDS. 1998;12:F91-F95.
5. Ostrow DE, Fox KJ, Chmiel JS, et al. Attitudes towards highly active antiretroviral therapy are associated with sexual risk taking among HIV-infected and uninfected homosexual men. AIDS. 2002;16:775-780.
6. Remien RH, Wagner G, Carballo-Dieguez A, et al. Who may be engaging in high-risk sex due to medical treatment advances? AIDS. 1998;12:1560-1561.
7. Van de Ven P, Rawstorne P, Nakamura T, et al. HIV treatments optimism is associated with unprotected anal intercourse with regular and with casual partners among Australian gay and homosexually active men. Int J STD AIDS. 2002;13:181-183.
8. Van de Ven P, Crawford J, Kippax S, et al. A scale of optimism-scepticism in the context of HIV treatments. AIDS Care. 2000;12:171-176.
9. Koblin BA, Avrett S, Taylor PE, et al. Willingness to participate in HIV-1 vaccine efficacy trials and the effect of media events among gay and bisexual men in New York City. Project ACHIEVE. J Acquir Immune Defic Syndr Hum Retrovirol. 1997;15:165-171.
10. 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. 1996;174:954-961.
11. Gross M, Seage GR, Mayer KH, et al. Interest among gay/bisexual men in greater Boston in participating in clinical trials of preventive HIV vaccines. J Acquir Immune Defic Syndr Hum Retrovirol. 1996;12:406-412.
12. Bartholow BN, MacQueen KM, Douglas JM Jr, et al. Assessment of the changing willingness to participate in phase III HIV vaccine trials among men who have sex with men. J Acquir Immune Defic Syndr Hum Retrovirol. 1997;16:108-115.
13. Koblin BA, Heagerty P, Sheon A, et al. Readiness of high-risk populations in the HIV Network for Prevention Trials to participate in HIV vaccine efficacy trials in the United States. AIDS. 1998;12:785-793.
14. Seage GR III, Holte SE, Metzger D, et al. Are US populations appropriate for trials of human immunodeficiency virus vaccine? The HIVNET Vaccine Preparedness Study. Am J Epidemiol. 2001;153:619-627.
15. Colfax G, McKirnan D, Neidig J, et al. Motivations for trial participation and risk behavior among participants in an HIV vaccine efficacy trial. Paper presented at: National HIV Prevention Conference; August 12-15, 2001; Atlanta, GA.
16. Halloran ME, Longini IM, Haber MJ, et al. Exposure efficacy and change in contact rates in evaluating prophylactic HIV vaccines in the field. Stat Med. 1994;13:357-377.
17. Blower SM, McLean AR. Prophylactic vaccines, risk behavior change, and the probability of eradicating HIV in San Francisco. Science. 1994;265:1451.
18. Centers for Disease Control. HIV Counseling, Testing, and Referral Standards and Guidelines. Atlanta, GA: US Department of Health and Human Services, Public Health Service, CDC; 1994.
19. 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. In press.
20. Cohen J. AIDS vaccine trial processes disappointment and confusion. Science. 2003;299:1290-1291.
21. SAS. Cary, NC: SAS Institute, Inc.; 2002.
22. Johnson WD, Hedges LV, Ramirez G, et al. HIV prevention research for men who have sex with men: a systematic review and meta-analysis. J Acquir Immune Defic Syndr. 2002;30(Suppl 1):S118-S129.
23. Semaan S, Des J, Sogolow E, et al. A meta-analysis of the effect of HIV prevention interventions on the sex behaviors of drug users in the United States. J Acquir Immune Defic Syndr. 2002;30(Suppl 1):S73-S93.
24. Neumann MS, Johnson WD, Semaan S, et al. Review and meta-analysis of HIV prevention intervention research for heterosexual adult populations in the United States. J Acquir Immune Defic Syndr. 2002;30(Suppl 1):S106-S117.
25. The EXPLORE Study Team. Effects of a behavioural intervention to reduce acquisition of HIV infection among men who have sex with men: the EXPLORE randomized controlled study. Lancet. 2004;364:41-50.
26. Crepaz N, Marks G, Mansergh G, et al. Age-related risk for HIV infection in men who have sex with men: examination of behavioral, relationship, and serostatus variables. AIDS Educ Prev. 2000;12:405-415.
27. Mansergh G, Marks G. Age and risk of HIV infection in men who have sex with men. AIDS. 1998;12:1119-1128.
28. Festinger L. A Theory of Cognitive Dissonance. Evanston, IL: Row, Peterson & Co.; 1962.
29. Festinger L. Cognitive dissonance. Sci Am. 1962;207:93-107.
30. Festinger L. Conflict, Decision, and Dissonance. Palo Alto, CA: Stanford University Press; 1964.
31. Kerns TA. Ethical Issues in HIV Vaccine Trials. New York: St. Martins Press; 1997.
32. Cook TD, Campbell DT. Quasi-experimentation: Design and Analysis Issues for Field Settings. Chicago: Rand McNally; 1979.
Keywords:

HIV vaccine efficacy trials; HIV risk behavior; clinical trials; sexual risk behavior; MSM; women at risk

© 2005 Lippincott Williams & Wilkins, Inc.