Sigma Research

HIV infection in a cohort of homosexual and bisexual men

BMJ (British Medical Journal), 1992, 305: 561-2.

Authors: AJ Hunt, PM Davies, TJ McManus, P Weatherburn, FCI Hickson, G Christofinis, APM Coxon, S Sutherland

Short article - full text

Epidemiological studies from the United States report rises in unprotected anal intercourse among gay and bisexual men, 1 after dramatic reductions in the mid 1980s there 2 and in the United Kingdom. 3 In 1991 rising rates of rectal gonorrhoea among homosexual men attending genitourinary medicine clinics 4 led to speculation about an increase in HIV risk behaviour in the United Kingdom.

This speculation is complicated by the bias in samples of clinic attenders, towards more sexually active men but may be clarified by data from samples of non-attenders.

Subjects, methods, and results

In 1987-8, we interviewed a cohort of 930 homosexual and bisexual men in 10 cities in England and Wales who have been recruited from sources other than genitourinary medicine clinics. Subsequently 77%, 65%, and 50% were re-interviewed annually. All men were interviewed about current and past sexual behaviour. Respondents in London and South Wales were asked, though not required, to provide a blood sample for testing for HIV-1 antibody; for each year 344, 290, 240 and 296, about two thirds of those eligible, did so. The results, with counselling, were available to the men at their specific request. They were linked with behavioural data only for analysis. Ethical committee approval was obtained for all localities at each stage of the study.

The median age of the original cohort was (29 range 15 - 81) years. Full details are given elsewhere. 5 The proportion of regular clinic attenders was 31% in 1987/8 and did not vary significantly. The proportion tested before the study began was 41% rising in 1991 to 54% of the men interviewed. The proportion of samples positive for HIV-1 antibody was significantly higher in the clinic attenders than non-attenders (15.6% v 3.8%, (02 = 11.35, p = 0.0018).

Unprotected anal intercourse was more common in men in regular than in casual relationships (19% v. 4%). The mean annual number of partners with whom anal intercourse occurred has increased from 2.0 in 1987-8 to 2.7 in 1990-1, and the number of partners increased from 12.3 to 17.9 (t = 2.53, p = .011). Meanwhile, the proportion reporting oro-anal contact in the month before interview rose from 31% to 41% and reporting digital-anal contact from 42% to 57%.

Seventy three men in the cohort were antibody positive, some of whom subsequently died; 13 do not know their HIV status. Eleven (15% of all positive) men who were antibody negative in 1987/8 have subsequently tested positive. The mean period between the positive result and the last negative result was 11(range 2 to 24) months . In ten men we could identify the year of seroconversion. Our best estimates are 1987, one man; 1988, two; 1989, two; 1990 three and 1991, two.

All 11 men reported unprotected anal intercourse and a range of other sexual acts before seroconversion. In five, unprotected receptive anal intercourse was the probable mode of transmission. In two men, unprotected insertive intercourse was the likely mode but both had also practised receptive fellatio with orgasm, one with a partner known to be antibody positive and one had also engaged in insertive fisting. In the four other men both unprotected receptive and insertive anal intercourse had occurred.

In eight men, the source of infection was clearly identified as a regular partner, four of whom were known to be antibody positive at the time. In only one man was the source clearly trace to a casual partner. Three of the seroconversions were not in regular clinic attenders, nor had the men attended in the year before interview. Four men had not previously been tested.


Although the small numbers make conclusive claims hazardous, these data do not substantiate a large increase in unprotected anal intercourse. One reason may be rises in oro-anal and anal-digital contact, which impugn the validity of rectal gonorrhoea as a surrogate marker for HIV risk behaviour.

Almost a fifth of men who tested antibody positive in this study were not clinic attenders. Just under a sixth of positive men became infected after the government campaigns began in 1986-7. Health promotion for gay and bisexual men remains a priority, and initiatives which concentrate on HIV transmission within relationships should be encouraged.

This research was funded by the Medical Research Council and the Department of Health.

1. Adib SM, Joseph JG, Ostrow DG, Tai M, Schwartz SA. Relapse in Sexual behaviour among homosexual men: a 2 year follow up from the Chicago MACS/CCS. AIDS 1991; 5:757-60.

2. Carne CA, Weller IVD, Johnson AM, Loveday C, Pearce F, Hawkins A, et al. Prevalence of antibodies to HIV, gonorrhoea rates and changed sexual behaviour in homosexual men in London. Lancet 1987, i:656-658.

3. Stall RD, Coates TJ, Hoff C. Behavioural Risk Reduction for HIV Infection among Gay and Bisexual Men. American Psychologist 1988; 43:878-885.

4. Waugh MA: Resurgent Gonorrhoea in Homosexual Men. Lancet 1991; 337: 375.

5. Davies PM, Hunt AJ, Macourt M, Weatherburn P: A Longitudinal study of the sexual behaviour of homosexual males under the impact of AIDS: A Final Report submitted to the Department of Health. London: Project Sigma, 1990.

Available FREE online