Introduction

Catalonia is an autonomous region in Spain with six million inhabitants. By December 2000, 13 275 cases of AIDS had been reported and the annual AIDS incidence rate was 81.8 per million in 2000. As a transmission group, homosexual men represented 20% of the total number of AIDS cases in 2000, showing a slight increase since 1998 (17.6%) (1). As part of the Sistema integrat de vigilància epidemiològica de l’HIV/sida a Catalunya (Catalan integrated surveillance system of HIV/AIDS), monitoring of HIV and sexual risk behaviour among men who have sex with men (MSMs) was introduced in 1993 (2).

Recently, different studies among MSMs have reported an increase in HIV incidence (3-5) and sexually transmitted infections (6-8), and an increase in sexual risk behaviour (9,10). This change to unsafe sexual behaviour has been explained in part by the optimism surrounding HIV treatment and the absence of the threat of death that existed a decade ago. Some authors found a significant relationship between unprotected anal intercourse and certain aspects of optimism in the context of new HIV treatments (11-12).

The objectives of this paper are to describe the trends in prevalence of HIV infection and in risk behaviour in 1995, 1998, and 2000 among MSMs recruited in different venues in Barcelona; to describe trends in knowledge and perception of antiretroviral treatment since 1998; and to examine relationships between the practice of unprotected anal intercourse and the perception of antiretroviral treatment.

Methods

Four cross-sectional surveys (1,2) have been carried out since 1993, with the participation of a community-based gay organisation (Stop sida). A convenient sample of MSMs was recruited each year in three saunas, two sex shops, a pick up site in a public park and through a mailing list of a community-based gay organisation. All venues were located in Barcelona.

The questionnaire, used since 1995, was validated and adapted from one developed by the Institut Universitaire de Médecine Sociale et Préventive, Lausanne (13). It collects information about demographic data, social network, sexual practices with steady and casual partners, drug use, HIV testing behaviour, and, since 1998, knowledge and perception of antiretroviral treatment and post-exposure prophylaxis. The behaviour described refers to the 12 months before the survey.

Around 2000 anonymous self-administered questionnaires were distributed each year by four health educators of a gay association. Questionnaires were returned by mail. In addition, to estimate the prevalence of HIV infection, around 300 saliva samples were requested each year to determine HIV antibodies using standard enzymatic immunoassay techniques (14). Furthermore, between 1000 and 1400 questionnaires were also sent to all male members of a gay association in Catalonia.

A descriptive statistical analysis was performed using χ 2 for comparisons of proportions and logistic regression to adjust estimates. The SPSS-PC V10 statistical software package was used. When differences were not statistically significant between the three surveys, only the results from the 2000 survey are presented. Unprotected anal intercourse (UAI) with a steady partner and with casual partners was calculated as the proportion of those who never or occasionally used a condom during the previous 12 months when practising anal intercourse with steady or casual partners. Statements about knowledge and perception of antiretroviral treatment (ARV) had five response options: strongly agree (1), agree (2), disagree (3), strongly disagree (4), do not know (5). For analysis, options have been collapsed (1+2 = agree; 3+4 = disagree) and option 5 (do not know) has been removed.

Table 1

Age group, HIV testing and sexual behaviours of men who have sex with men

Table 2

Knowledge and perception (agrees versus disagrees) regarding combined antiretroviral therapies (ARV). Comparison between 1998 and 2000. Logistic regression

Table 3

Association of unprotected anal intercourse (UAI) with casual partners versus with casual partners with statements about antiretroviral therapies ARV). Logistic regression

Results

The questionnaire return rate ranged from 29% (741/2543) in 1995 to 27% (713/2775) in 1998, and 24% (828/3445) in 2000. The mean age increased over the years (33.7, 34.8, and 36.1 years old in 1995, 1998, and 2000, respectively, p<0.001). With the exception of age, the selected samples were similar for demographic characteristics. Most of them had a high level of education (49% were university graduates) and were living in Barcelona (70%).

The vast majority of the homosexual men studied had been tested for HIV antibodies previously, and this proportion increased over the years surveyed (table 1). We also observed an increase in the proportion of those who had been tested more than once (40.9%, 50.9%, and 53.4% in 1995, 1998, and 2000, respectively). Self-reported prevalence of HIV was 16.4% [95% confidence interval: 13.1-19.8] in 1995; 16.6% [13.4-20.5] in 1998; and 18.8% [15.7-21.9] in 2000. Prevalence of HIV estimated in the saliva samples was 14.2% [95% confidence interval: 10.3-18.1] in 1995, 15.5% [11.3-19.8] in 1998, and 17.9% [13.6-22.2] in 2000. Although both self-reported and saliva prevalence of HIV show an increasing trend, neither are statistically significant.

Most respondents had sexual relationships with both steady and casual partners (53%), more than a third with casual partners only (36%), and 11% with steady partners only. Nearly a third had sexual encounters several times per week, and 14% irregularly, but with intensive periods. Although the frequency of sexual activity seems to be stable, the number of sexual partners shows an increasing trend. In 2000, 58% had more than 10 sexual partners during the previous year (45% and 54%, respectively, in 1995 and 1998, p<0.001). The association remains significant after adjusting for age (table 1).

Anal intercourse was practised by 87% with a steady partner and 83% with casual partners, and oro-genital sex by 98% and 96%, respectively. Unprotected anal intercourse with a steady partner was reported by 59%, and with casual partners by 25% (table 1). In both cases the proportion of UAI showed a slight increasing trend, but was not statistically significant. Similarly, we observed an increasing proportion of ejaculation in the partner’s mouth that only remained significant for casual partners after adjusting for age (table 1).

During the previous 12 months, 31% of the participants had experienced at least one condom breakage and 22% condom slippage. The participants who always used lubricants during anal intercourse had fewer accidents with condoms than those who did not. However, less than half of the men who used condoms during anal intercourse always used lubricants (40%) and of these, 18% chose liposoluble lubricants (oil, petroleum jelly, cream, etc) that can reduce the elasticity of the condom and thus increase the risk of breakage.

Prevalence of drug use before or during sexual encounters in the 12 months before the survey was high and, with the exception of alcohol, the use of drugs shows a clearly increasing trend over the years and remained significant after adjusting for age (table 1).

Responses to statements about ARV are presented in table 2. Although participants in 2000 seemed to be better informed, 5.6% still thought that "ARV therapies can cure HIV infection". More respondents than in the previous survey reported "side effects of ARV therapies are very uncomfortable" (70.2% and 83.5% in 1998 and 2000, respectively). In 2000, more MSMs had heard about post-exposure prophylaxis (PEP) than in 1998: 21.5% believed that "with ARV it is possible to avoid HIV infection after a potential sexual risk exposure" versus 7.4% in 1998. In general in 2000, there was a more optimistic perception regarding ARV therapy : a higher proportion than in 1998 believed that "with ARV, MSMs are less afraid of becoming HIV positive", and "pay less attention to prevention", and are "less worried if they take sexual risks".

We found a significant association between the statements that suggest optimistic perception of ARV therapy and unsafe sex, and this remained significant after adjusting for age, year of survey, and known serological status (table 3). Participants who agree that "HIV positive persons taking ARV therapy are unlikely to transmit HIV" are 1.9 times more likely to have UAI with casual partners, and those who agree that "with ARV it is likely to avoid HIV infection after a potential sexual risk exposure" are 1.7 times more likely. Similarly, those who report "being less afraid of becoming HIV positive", "paying less attention to prevention" or "being less worried if they take sexual risks" are also more likely to have UAI (odds ratio 1.9, 6.2, and 5.5, respectively).

Comment

In general, the results of this fourth survey did not show significant changes in the prevalence of HIV and in sexual risk behaviour of MSMs surveyed in Barcelona. Only three indicators showed a statistically significant increase: a higher proportion of participants than previous surveys reported having had more than 10 male sexual partners in the 12 months before the survey; a higher proportion of ejaculation in the mouth with casual partners; and a higher prevalence of drug use during or before sexual encounters. This is not consistent with the recently reported increase in both the HIV incidence rate (3-5) and in HIV risk behaviour (9,10) among MSMs from other countries.

Although cross-sectional studies suffer from methodological limitations, their repeated use over time with a standardised approach provides a useful way to describe the evolution of the HIV epidemic among MSMs. The return rate of questionnaires was not high, but was similar in the three surveys, and higher than in ‰ other studies (13). We cannot, however, generalise findings from this study to all MSMs in Barcelona. Self-reported sexual behaviour may also be subject to measurement error, but we attempt to reduce this by using an anonymous self-completed questionnaire.

The results of this fourth survey, however, still show a high prevalence of HIV among MSMs in Barcelona, a high level of unprotected sex with both steady and casual partners, and a high prevalence of drug use during or before sexual intercourse. In addition, among those who used condoms, accidents in the handling of condoms continued to be considerable and despite educational campaigns, the use of liposoluble lubricants is high. We also observed an increased trend of optimism regarding the efficacy of HIV combination treatment, a decreased worry about HIV infection because of PEP, and a significant association between optimism and UAI with casual partners, as reported in other studies (11,12).

We conclude that there is no clear evidence of an increase in sexual risk behaviour among MSMs, but the optimism regarding ARV may lead to a greater relaxation in protective measures in the near future. In addition, the high prevalence of HIV and risk behaviour, the difficulties to maintain safer sex practices ("AIDS burnout"), the longer survival of people infected with HIV, and the ineffectiveness of treatments due to drug resistant HIV strains may make the control of the epidemic difficult. Further and more creative interventions that take into account the new scenario are needed.

Acknowledgements

The authors thank Stop sida, the association that carried out the survey fieldwork, Kati Zaragoza, Rafa Muñoz, and all the men who voluntarily gave their time to answer the questionnaire.