Impact of Highly Active Antiretroviral Treatment on HIV Seroincidence Among Men Who Have Sex With Men: San Francisco

Unfortunately, availability of highly active retroviral therapy (HAART) paradoxically may increase sexual practices that lead to HIV transmission. Community surveys indicate that, as a result of the availability of HAART, HIV-negative men who have sex with men (MSM) are less concerned about contracting HIV; HIV-positive MSM are less concerned about transmitting HIV; and both groups are more likely to engage in unsafe sex. Also because HAART decreases mortality and improves the quality of life of persons with AIDS, it has increased the number of persons living with HIV/AIDS who are engaging in sexual relations.

Researchers conducted an ecological study to assess the potentially conflicting effects of HAART on seroincidence among MSM in San Francisco. They tracked changes in AIDS prevalence, sexual risk behavior, incidence rates of rectal gonorrhea, and seroincidence among MSM who sought HIV testing at an anonymous testing site or who sought STD services at the county STD clinic after HAART became available.

The incidence of AIDS among MSM was highest in 1992 and declined steadily thereafter. The number of deaths among MSM with AIDS increased somewhat from 1992 through 1995, followed by a dramatic decline through 1999, corresponding to the period of availability of HAART. The number of MSM living with AIDS who were using HAART increased from 278 (4 percent) in 1995 to 3,959 (54 percent) in 1999. Because the decline in AIDS deaths was greater than the decline in AIDS incidence, the number of MSM living with AIDS increased between 1994 (6,509) and 1999 (7,272).

Trends in sexual behavior among MSM were determined from cross-sectional surveys conducted by the San Francisco-based Stop AIDS Project. MSM were interviewed in a variety of settings (e.g., neighborhoods, clubs, bars, gay-oriented events, streets with large amounts of pedestrian traffic). During the study, 26,176 first-time interviews with MSM were conducted. Fewer interviews were conducted in later years (decreasing from 6,235 in 1994 to 3,303 in 1999) owing to fewer volunteers and less interest by the target audience in discussing HIV issues. Previously interviewed persons were excluded. Reanalyzing the data to include previously interviewed individuals did not change the results.

Cross-sectional interviews showed that the percentage of MSM who reported always using condoms in the preceding six months when engaging in anal intercourse decreased from 70 percent in 1994 to 54 percent in 1999 (P<.001). Over the same period, the percentage of MSM reporting both unprotected anal sex and multiple sexual partners increased from 24 percent to 45 percent (P<.001). The decreases in respondents' reports of always using condoms and the increases in their reports of unprotected anal intercourse and multiple partners seen between 1994 and 1999 were statistically significant (P<.05) within all age groups studied and all five ethnic groups of MSM, with the exception for Native American MSM condom use, which did not achieve statistical significance, owing to its small sample size.

Trends in HIV incidence were determined via an algorithm using two enzyme-linked immunoassays to estimate the length of time recently infected individuals have been infected (duration), which was then used in estimating incidence rates from cross-sectional serosurveys. Results were available for 11,347 MSM at the anonymous testing site from 1996 to 1999 (average 2,837 annually) and 4,669 identity-stripped MSM seeking STD services from 1995 to 1999 (average 778 annually).

Among MSM tested at the anonymous site between 1996 and 1999, the estimated HIV annual incidence rate (using the algorithm) was 2.2 percent. The annual number of serological HIV tests declined during this time period. Annual estimated HIV incidence rates were lowest in 1997 (1.3 percent) and highest in 1999 (4.2 percent [P<.015]). Among the STD clinic patients, annual estimated HIV incidence was 5.3 percent, and this rate remained stable between 1994 and 1999 (P<.26) despite an increased number of tests performed.

The number of male rectal gonorrhea cases reported in San Francisco increased from 72 in 1994 to 160 in 1999, with a concomitant increase in the rate of gonorrhea per 100,000 men 15 years or older. The positivity rate of gonorrhea tests performed at the STD clinic remained stable, despite an increased frequency of testing, from 1994 to 1998, but decreased in 1999.

The results of the ecological study are consistent with the hypotheses that (1) the availability and increased use of HAART have resulted in an increased number of unsafe sexual contacts and (2) the impact of the increased number of unsafe contacts on seroincidence has been equal to or greater than the impact of HAART use on per contact transmissibility of HIV.

Because this was an ecological study, researchers could not establish that the relationship between availability of HAART, increased risk behavior, and a flat or increasing seroincidence is causal. However, a linkage of the AIDS registry in San Francisco with the STD registry showed that MSM with AIDS who were treated with HAART were more likely to acquire an STD than were MSM not treated with HAART. Also, the increases in unsafe sex parallel the increases in HAART use, and other studies have established that perceptions and behaviors of MSM have been influenced by HAART. The results suggest the need for MSM-focused prevention campaigns that take into account changes in perceived risk of HIV owing to HAART, the authors concluded.

Back to other CDC news for April 16, 2002

Previous Updates