Sexual Seroadaptation: Lessons for Prevention and Sex Research from a Cohort of HIV-Positive Men Who Have Sex with Men

Abstract

Background

Surveillance data on sexually transmitted infections (STIs) and behavioral characteristics identified in studies of the risk of seroconversion are often used as to track sexual behaviors that spread HIV. However, such analyses can be confounded by “seroadaptation”—the restriction of unprotected anal intercourse (UAI), especially unprotected insertive UAI, to seroconcordant partnerships.

Methods

We utilized sexual network methodology and repeated-measures statistics to test the hypothesis that seroadaptive strategies reduce the risk of HIV transmission despite numerous partnerships and frequent UAI.

Principal Findings

In a prospective cohort study of HIV superinfection including 168 HIV-positive men who have sex with men (MSM), we found extensive seroadaptation. UAI was 15.5 times more likely to occur with a positive partner than a negative one (95% confidence interval [CI], 9.1–26.4). Receptive UAI was 4.3 times more likely in seroconcordant partnerships than with negative partners (95% CI, 2.8–6.6), but insertive UAI was 13.6 times more likely with positives (95% CI, 7.2–25.6). Our estimates suggest that seroadaptation reduced HIV transmissions by 98%.

Conclusion

Potentially effective HIV prevention strategies, such as seroadaptation, have evolved in communities of MSM before they have been recognized in research or discussed in the public health forum. Thus, to be informative, studies of HIV risk must be designed to assess seroadaptive behaviors rather than be limited to individual characteristics, unprotected intercourse, and numbers of partners. STI surveillance is not an effective indicator of trends in HIV incidence where there are strong patterns of seroadaptation.

Introduction

Surveillance data on sexually transmitted infections (STIs) and results from behavioral studies of the risk of seroconversion have frequently been used as surrogate markers of trends in sexual behaviors that spread new HIV infections and signal changing trends in the HIV epidemic. The incidence of rectal gonorrhea among men who have sex with men (MSM) dropped precipitously with the onset of the AIDS epidemic in the 1980s. STIs, such as gonorrhea, usually become symptomatic soon after infection, and testing and treatment were simple and widely available, making STI surveillance an accessible indicator of HIV risk. At the same time, the rates of high-risk sexual behaviors and the incidence of HIV decreased [1][4]. Behavioral surveillance focusing on risk of seroconversion is relatively inexpensive and can be implemented with a wide variety of sampling strategies that can lead to findings that can be generalized. The concomitant decreases in the incidence of rectal gonorrhea, rates of high-risk sexual practices, and HIV incidence in the 1980s in San Francisco suggested that the former two measures were good surrogate indicators of potential trends in HIV incidence.

By the mid-1990s, behavioral studies revealed increasing rates of sex unprotected by condoms among MSM, and STIs rose accordingly [5][6]. For example, from 1993 to 1999 in San Francisco, the proportion of MSM reporting multiple partners and unprotected intercourse increased while those reporting always using condoms dropped, and the rates of rectal gonorrhea rose [7]. However, more direct measures of HIV incidence in San Francisco did not show a concurrent increase [6][10]. This apparent discrepancy may have reflected changes in the infectivity of HIV due to widespread treatment [11] or other factors that specifically inhibited new HIV infections independently of other STIs.

However, changes in sexual mixing based on HIV-1 status could also partially explain discrepancies between trends in HIV and other STIs. Sexual behaviors adapted to the risk of HIV developed quickly among MSM in San Francisco, including abstinence, reduction in numbers of partners, avoiding anal sex, and condom use, and was credited with dramatic reductions in HIV incidence density before the end of the 1980s [1][2].

The HIV epidemic has engendered an unprecedented sexual behavior surveillance infrastructure; a likewise unprecedented and large body of research on specific sexual practices, especially unprotected anal intercourse (UAI) among MSM has resulted. However, research that narrowly trained its sights on practices that could spread the epidemic was poorly suited to measure the role of behavior change in inhibiting the spread of the epidemic [12] (p71). Special adaptive strategies might be particularly relevant in communities in which the members correctly perceive themselves to be at risk of acquiring or transmitting infection. Newer trends in behaviors, discussed by French activists and intellectuals under the term “seroadaptation,” might reverse trends toward fewer partners and more condom use without concomitant increases in HIV incidence [13].

Data from a population-based sample of MSM living in California in 2002 led researchers in 2006 to conclude that knowledge of sexual partner’s serostatus was associated with sexual practices frequently indicating an apparent seroadaptive strategy to reduce the risk of transmission without necessarily limiting numbers of partnerships or ruling out UAI. Serosorting, generally understood to refer to the selection of only concordant-serostatus partners, at least for UAI, was only one tactic characteristic of this seroadaptive strategy. While MSM often chose partners of discordant or unknown status those partnerships were less likely to involve anal intercourse or more likely to include condom use [14].

First described as a harm-reduction strategy in an Australian study, “strategic positioning” was another seroadaptive tactic in which an HIV-positive individual assumed the receptive role during UAI with a negative or unknown-status partner [15]. Because the risk of transmission from a receptive to an insertive partner during unprotected anal intercourse is minimal compared to the obverse [16][18], strategic positioning was another adaptive behavior that could mediate HIV incidence.

Other evidence of seroadaptation had been documented before the cohort study reported here although not in these terms. Limited serosorting was reported for Austrailian MSM between 1986 and 1991 [19]. In 1991–1992, 9% of a sample of HIV-positive MSM in Los Angeles reported insertive UAI and were more likely to do so with HIV-positive partners (odds ratio, 3.27) [20]. During the same period in the United Kingdom, one study found that in 15% of partnerships the serostatus of both partners was known to the respondent. Among this group, HIV-positive MSM were more likely to have UAI with HIV-positive MSM and less so with HIV-negative MSM (odds ratio, 1.64 vs. 0.24, p<.05) than with unknown-status partners [21]. During the latter half of the 1990s, patterns of serosorting and strategic positioning were evident in HIV-positive MSM in San Francisco and New York, even though the samples excluded the most successful seroadapters—those who reported no unprotected intercourse with negative or unknown-status partners [22]. In 1992, Hoff et al. found UAI much more likely to occur in seropositive concordant rather than discordant relationships (54% vs. 17%) among MSM in Portland, Oregon, and Tucson [23]. These examples are not an exhaustive list.

However, the present study is the first in which the data collection methodologies and analytic frameworks were chosen for the specific purpose of learning the degree to which seroadaptive tactics arose among sexual partnerships defined as at high-risk for HIV-transmission (involving HIV-positive MSM who practice UAI) that eluded both public health-sanctioned messages and empirical detection. It is also a unique effort to assess the potential impact of such tactics on decreasing exposures that might lead to transmission of new HIV infections. Whether this HIV prevention strategy arose from “grassroots” origins and, especially if they are effective, a theoretical question is raised: Do such tactics add up to a community-based prevention strategy? And under what conditions can they appear? Answers to these questions have bearing on the development of HIV prevention policy and programs.

As early as the mid-1980s, epidemiologists were using sexual mixing–based models of the epidemic that drew heavily on the theory and methods of social network research [24][29]. Social network studies have three fundamental characteristics: relationships rather than individuals are studied, relationships between individuals can be drawn or graphed in “sociograms” that illustrate social structure, and social structure can therefore be subjected to mathematical analysis and modeling. An advantage of these methods in studies of HIV seroincidence or superinfection is that exposure data are collected partner by partner, and the characteristics of each partner are documented separately. Thus, one can determine if partnership characteristics predict sexual risk and compare different analytic approaches to see how well they represent sexual risk in the sample overall.

Methods

Objectives

We had two primary objectives in this analysis. The first was to examine sexual partnership data for evidence of seroadaptive tactics among HIV-positive MSM in San Francisco before the recognition and discussion of serosorting in the peer-reviewed scientific literature, among public health officials, and in HIV-prevention policy. We tested the hypothesis that HIV-positive MSM reduced the risk of infecting partners both by serosorting and by strategic positioning—the restriction of UAI, especially unprotected insertive UAI, mainly to seroconcordant partnerships.

Our second objective was to provide an empirically based estimate of the impact of seroadaptation, especially serosorting and strategic positioning, on containing the HIV epidemic. In addition, we examined whether different methodological and analytic approaches to the study of HIV transmission risk could distort the impact of behavior changes like seroadaptive tactics on transmission incidence. Abstinence, avoiding anal intercourse altogether, always using condoms for intercourse, and other seroadaptive strategies were not examined empirically.

Participants

Positive Partners was a prospective study of HIV-1 superinfection among HIV-positive men and women in San Francisco who reported frequent unprotected intercourse (>10 episodes) with at least one HIV-infected partner over the previous year. This eligibility criterion was optimized to look for evidence of HIV superinfection; hence, individuals with only negative or unknown-status partners were excluded. Prospective participants were screened for eligibility when they called the study in response to professional or personal referrals, brochures and fliers placed in gay and AIDS services venues, or advertisements printed bi-weekly in a gay-oriented newspaper. For the majority of participants (85.1%), a current HIV-positive sexual partner was also screened and enrolled. The enrollment of current seroconcordant partnerships was key to the superinfection study aims so that sexual exposure to a genetically divergent virus in an enrolled partner could be distinguished from exposure to unknown strains from other HIV-positive partners. Exposure to HIV-negative partners that could not lead to superinfection—or to unknown-status partners where the risk of exposure could only be estimated by HIV prevalence in the population—could be measured and controlled in estimating the risk of superinfection.

The data analyzed in this study were obtained from a subsample of participants in the Positive Partners Study consisting of all 168 MSM enrolled from January 2002 to December 2004 in San Francisco. We focused on this timeframe because subjects were observed during a period when seroadaptation had not been recognized or discussed in the scientific, public health, or popular literatures.

Procedures

Behavioral interviews were conducted and biological specimens (blood and semen) were collected at enrollment and at the 1-year follow-up. The sexual partnerships and contacts of each participant during the 3 months before the intake and exit interviews (past 3 months) were documented. Sexual contacts were documented with a novel instrument based on egocentric social network methods [30][43], which were adapted to study sexual networks [44][56]. Egocentric social/sexual network designs depend on informants (egos) to characterize their partners (alters) and the relationships between them. To sample partnerships, we first selected all sexual partners in the past 3 months. During the self-administered interview module, participants were asked to fill out a “partner journal” in which each partner was described in a separate section, including a distinguishing identifier (first name, nickname, or other descriptor). Information collected on each partner in the past 3 months included basic demographics, HIV status, specific types and numbers of sexual contacts, and an indicator of partnership concurrency (during which month or months had sexual contact occurred). During the interviewer-administered module, we selected as many as four of the most recent of those partners (if four or more had been reported) plus the enrollment partner, if one existed, into a subsample for extensive characterization. An additional 29 questions solicited information on the characteristics of each subsample partner, their relationship, and the timing of sexual contacts. Our analysis indicated that the subsample partnerships were representative of the overall sample of partnerships in the past 3 months.

Ethics

All subjects gave written informed consent to participate in the study. The protocol and consent forms were approved by the Committee on Human Research at the University of California, San Francisco.

Analytic and Statistical Methods

Seroadaptation in Partnerships: A network “sexiogram” differentiated sexual linkages that may or may not have caused new infections. In contrast to individual-based analysis, inclusion of partnership-specific information provides evidence of seroadaptation that may decrease exposures that could result in new HIV infections in partnerships. Partnership information also provides better resolution for evaluating the burden of exposure in the population. Figure 1 shows a reconstruction of partial sexual networks of two couples during the 3 months before enrollment in the Positive Partner’s study, including all partners of each individual and diagrams of the connections with shared partners. This sexual network diagram illustrates variation in the risk of HIV transmission even among couples who practice UAI and have multiple serodiscordant partnerships. Individuals A and D were both HIV-infected, had multiple partnerships, frequently practiced UAI, and had HIV-negative partners. Many analyses would characterize them as equally likely to spread the epidemic. However, unprotected intercourse between seropositive partners does not pose a threat of new HIV infections. Therefore, D (a “complete” seroadapter) has not had UAI with any of his negative partners, and so we do not count any of his partnerships as likely transmission linkages. In contrast, while A is a “partial” seroadapter, we can count potential transmission linkages separately from low-risk partnerships. For this reason, sexual network data are particularly well suited to estimating the epidemiological impact of seroadaptation.

See full information

Be the first to comment

Leave a Reply

Your email address will not be published.


*