The bisexual bridge revisited

From 1998 to 2003 in San Francisco, unprotected anal intercourse (UAI) with one or more partner of unknown HIV serostatus (potentially serodiscordant UAI) increased among bisexual men who have sex with men (MSM) in parallel and in magnitude compared with other MSM. Potentially serodiscordant UAI and unprotected vaginal intercourse increased from 1998 to 2001 then decreased from 2001 to 2003. Although elements of a ‘bisexual bridge’ are present, we note an encouraging decrease in risk from 2001 to 2003. Over the past few years, unprotected anal sex, sexually transmitted infections (STI), and HIV incidence has increased among men who have sex with men (MSM) in San Francisco [1,2] and urban MSM communities elsewhere in north America, Europe, and Australia [3–7]. In our previous studies, trends in sexual risk behavior were analysed combining both MSM exclusively and men who have sex with men and women (MSMW) [1,2]. Although without conclusive evidence, a ‘bisexual bridge’ has been a cause for concern in the past because of the possible spread of HIV into the general population [8]. In the present report, we update our sexual risk behavior data and compare trends in MSM with MSMW in San Francisco. We used the serial, cross-sectional surveys collected by the STOP AIDS Project in the course of administering health education outreach to MSM and MSMW in San Francisco. Details of the survey methods and questionnaire have been described previously [1,2]. Briefly, the study population is a convenience sample recruited at a wide variety of gay-identified venues and streets in gay neighborhoods of San Francisco. The questionnaire includes demographics and sexual behaviors in the preceding 6 months, including anal and oral sex, condom use, number of partners, unprotected vaginal sex with women, self-reported HIV serostatus, and knowledge of partners’ HIV status. In the present report, we focus on trends in unprotected anal intercourse (UAI), UAI with one or more partners of unknown HIV serostatus (potentially serodiscordant UAI), and unprotected vaginal intercourse (UVI) among MSMW. As in a previous report [9], our definition of potentially serodiscordant UAI does not include situations in which respondents knowingly had unprotected sex with individuals of opposite HIV serostatus. Because knowingly exposing someone to HIV can be a felony in California, partner serostatus was not ascertained to avoid self-incrimination. The exclusion of knowingly having UAI with someone of opposite HIV serostatus would result in the probable underestimation of overall serodiscordant UAI [10]. We examined data from the second half of 1998 to the first half of 2003 for consistency in the wording of questions. During the study period, 17 479 individuals completed the survey. Of these, we defined MSMW as men who either called themselves bisexual or who had had anal or oral sex with men and vaginal sex with women in the past 6 months (n = 2461). UAI among MSMW tracked but lagged behind MSM (Fig. 1). The proportion of MSMW reporting any UAI with men increased from 25% in 1998 to 28% in 2003 compared with an increase from 32 to 40% for MSM. The proportion of MSMW reporting potentially serodiscordant UAI followed very closely and at times exceeded that reported by MSM. The Cochran–Armitage test determined that trends in UAI (P = 0.047) and potentially serodiscordant UAI (P < 0.001) for MSMW significantly increased from 1998 to 2003. Of note is the fact that potentially serodiscordant UAI peaked in 2001 and declined to 2003 for both MSM and MSMW, the first indication of any recent decrease in sexual risk behavior among MSM in San Francisco. The downward trend in potentially serodiscordant UAI from 2001 to 2003 was significant for MSM (P = 0.022), but not for MSMW (P = 0.075), most probably because of the lower sample size in the MSMW group. Although there was no significant overall change from 1998 to 2003, UVI among MSMW significantly increased from 1998 to 2001 (P < 0.001) and significantly decreased from 2001 to 2003 (P = 0.039).Fig. 1. Trends in unprotected anal intercourse, unprotected anal intercourse with one or more partner of unknown HIV serostatus and unprotected vaginal intercourse among men who have sex with men and women and men who have sex with men, San Francisco, CA, 1998–2002.: MSM, Men who have sex with men; MSMW, men who have sex with men and women; UAI, unprotected anal intercourse; potentially serodiscordant UAI, unprotected anal intercourse with one or more partner of unknown HIV serostatus; UVI, unprotected vaginal intercourse.We conducted further analyses comparing MSMW with other MSM using multivariate logistic regression. Significant predictors of being MSMW included African-American ethnicity [adjusted odds ratio (AOR) 1.41; 95% confidence interval (CI) 1.20–1.65], younger age (AOR 0.98 per year; 95% CI 0.97–0.99), oral sex to ejaculation (AOR 1.34; 95% CI 1.17–1.53), and living in the Western Addition (a historically African-American neighborhood) (AOR 1.31; 95% CI 1.11–1.55). Men who were white (AOR 0.62; 95% CI 0.54–0.71), HIV-positive (AOR 0.73; 95% CI 0.62–0.86), had unprotected anal sex (AOR 0.68; 95% CI 0.61–0.75), lived in the Castro (a strongly gay-identified neighborhood) (AOR 0.68; 95% CI 0.59–0.79), and who had been previously interviewed by the STOP AIDS project team (AOR 0.63; 95% CI 0.57–0.70) were less likely to be MSMW than exclusively MSM. There were significant interactions between being white and engaging in unprotected oral sex (AOR 1.20; 95% CI 1.00–1.43), being HIV positive and having unprotected anal sex (AOR 1.44; 95% CI 1.05–1.97), and being of a younger age and living in the Castro (AOR 0.97; 95% CI 0.95–0.99). In a separate analysis, we compared MSMW reporting UVI with those who did not. Having UVI was significantly and positively correlated with older age, being white, having UAI, oral sex to ejaculation, and living in the Western Addition. African-American MSMW were significantly more likely to call themselves ‘bisexual’ and significantly less likely to call themselves ‘gay’ and to live in the Castro, compared with MSMW of other races. In the past 5 years, sexual risk behavior among MSMW has increased in parallel with other MSM. Clearly, elements of a ‘bisexual bridge’ are present. However, past dire predictions of the widespread transmission to the heterosexual population in San Francisco have failed to come true [11]. Of note is the fact that unprotected vaginal sex for MSMW peaked in 2001, with rates almost returning to 1998 levels by 2003. Moreover, we found for the first time an encouraging decrease in sexual risk behavior for both MSM and MSMW from 2001 to 2003. Of note is the fact that African-American MSMW were more likely to identify as bisexual and less likely to identify with and live within the mainstream gay community of San Francisco. Our data may support other reports in the scientific literature and the popular press that among MSM, African-American men tend to engage in more bisexual activity [12,13] and African-American MSM and MSMW are more closeted about their sexuality (sometimes called the ‘down-low’ phenomenon) [14–17]. These factors may partly explain the relatively high prevalence of HIV among African-American women in San Francisco [11] and the United States [18].