From the mids to the mids there was a decrease in the percentage of gay men in Australia practising unprotected anal intercourse . Concomitantly, there was a decrease in the estimated number of new HIV infections among gay men from over in to a plateau of — infections per annum from to now [1,2].
The introduction of improved antiretroviral therapies in Australia and other western industrialized countries from the mid-to-late s has occurred at the same time as an increase in gay men's sexual risk practices [4—8] , with an increased incidence of new HIV infections reported in North America [9,10]. Examinations of sexual risk practices have found an association with optimism that antiretroviral therapy results in reduced infectivity [11—14] , although comparative studies show the heterogeneity across different countries of gay men's responses to HIV therapies .
We examined findings from cross-sectional surveys February to February assessing sexual risk behaviour among gay men in Sydney. Using an anonymous, self-completed questionnaire, surveys were conducted 6 monthly at four consistent sites: Men were recruited if they had had sex with another man in the past 5 years.
In addition, they had to live in Sydney or have participated regularly in the Sydney gay community. At the Fair Day, recruiters were positioned at strategic stalls and alleyways and they approached as many men as possible.
A six-item scale of optimism—scepticism in the context of HIV treatments Cronbach's alpha 0. Total scores on the scale could range from 6 sceptical to 24 optimistic. For the analysis of behavioural trends, data from the four consistent sites and the Fair Days, and for HIV-negative and HIV-positive men, were analysed separately. Retaining each of the whole samples as the base, anal intercourse with casual partners was collapsed into a binary variable: In all, 10 men ranging in age from 16 to 81 years median 34 years were recruited, at the four consistent sites overall response rate From the four consistent sites, Self-reported HIV status was However, for the reduced base of HIV-positive men who had anal intercourse with casual partners, there was a significant upward trend in UAI-C.
In February , mean scores on the HIV optimism scale tended towards scepticism. The significant increase in sexual risk behaviour among gay men in Sydney parallels that documented elsewhere [5—8].
The trends have been most evident among men recruited from gay community-based venues and clinics. However, by February , trends emerged among men recruited at the Fair Day. This highlights the importance of undertaking behavioural surveillance work across a range of settings, not just Pride Festivals, because different patterns of sexual risk taking may emerge from different sites.
Corroborating reports from elsewhere [11—14] , there is a significant association between sexual risk behaviour and HIV optimism. The current data, however, do not permit us to imply causality. Contrary to recent North American data [9,10] , there has been no increase in new HIV infections among gay men in Australia  , despite a national surveillance system for newly acquired HIV infection .
One possible explanation is Australia's relatively low prevalence of HIV and high-level access to and uptake of antiretroviral treatment [1,18] , with resultant reduced transmission. Another potential explanation is that the monitoring system for newly acquired HIV infection is insufficiently sensitive to detect small increases.
This is an important juncture in the epidemic, requiring close scrutiny of behavioural trends and HIV incidence rates, and an effective educational response, to prevent a resurgence in HIV infection related to a breakdown in safe sex norms that have served gay communities so well. Paul Van de Vena.