Incidence of HIV infection in homosexual men in London, 1988-94

BMJ 1995; 311 doi: (Published 26 August 1995) Cite this as: BMJ 1995;311:545
  1. E Miller, head of immunisation divisiona,
  2. P A Waight, principal scientista,
  3. R S Tedder, headb,
  4. S Sutherland, consultant medical virologistc,
  5. P P Mortimer, director of hepatitis and retrovirus laboratoryd,
  6. M S Shafi, directore
  1. aPHLS Communicable Disease Surveillance Centre, London NW9 5EQ
  2. bDepartment of Medical Microbiology, University College and Middlesex School of Medicine, Windeyer Building, London W1P 6DB
  3. cPublic Health Laboratory and Medical Microbiology, King's College School of Medicine and Dentistry, Dulwich Hospital, London SE22 8QF
  4. dVirus Reference Division, Central Public Health Laboratory, London NW9 5HT
  5. ePublic Health Laboratory, Central Middlesex Hospital NHS Trust, Park Royal, London NW10 7NS
  1. Correspondence to: Dr Miller.
  • Accepted 12 June 1995

Measures of the incidence of HIV infection in different groups are essential for targeting and evaluating health education. Incidence measures obtained by the follow up of closed cohorts become biased over time as early infection in those at highest risk results in a subsequent downward trend in incidence.1 2 We describe changes in the age specific incidence of HIV infection in open cohorts of homosexual men continuously recruited from among patients attending genitourinary medicine clinics in London from 1988 to 1994.

Incidence of HIV infection in HIV negative homosexual men by age and over time

View this table:

Patients, methods, and results

Homosexual men negative for HIV antibody who had attended genitourinary medicine clinics in London from January 1988 to April 1994 and been voluntarily retested for HIV antibody were identified from computerised records held by four laboratories participating in a Public Health Laboratory Service collaborative survey.3 Person time denominators were obtained by summating the time between the first and last test giving negative results for those remaining antibody negative and the time between the first negative and first positive result for those who became positive for HIV antibody. Incidence estimates over time and between age groups were compared with the Poisson likelihood ratio test.

A total of 6753 homosexual men were identified as being negative for HIV antibody, of whom 1759 (26%) had one or more repeat tests during the study period. The proportion retested was similar in each age group and did not vary significantly over time. The overall mean time between the first and last test was 676 days. A total of 124 men became positive for HIV antibody, 84 (68%) of whom were aged under 30. The overall incidence was 3.8 per 100 person years (95% confidence interval 3.2 to 4.5). The rate was highest in men aged 20-24 (6.2 per 100 person years) and declined significantly with age to 1.6 in men aged 40 and over (P<0.001).

Sixty nine of the 1418 men who had repeat tests within two consecutive calendar years became positive for HIV antibody. Incidence remained at around 9 per 100 person years among the younger men (under 30), with no evidence of a downward trend over time (P>0.5) (table). Incidence was significantly lower in older men (P=0.03), averaging 3.3 per 100 person years; the decline over the period in this group was not significant (P=0.10).


Our study illustrates the value of the open cohort method applied to data obtained from voluntary HIV antibody testing for monitoring incidence in groups at high behavioural risk.1 The difference in incidence with age found in homosexual men is unlikely to result from bias in selection for retesting as the proportion with repeat samples was similar across and within age groups over time.

The continuing high incidence of HIV infection in young homosexual men in London is worrying. Behavioural studies that will identify the factors leading to unsafe sexual behaviour in this group are urgently needed. Although the high incidence measured in our study cohort probably does not apply to all homosexual men, the identification of a subgroup at high risk of infection is important because targeting interventions at this subgroup could have a substantial effect on reducing the rate of transmission overall. If candidate HIV vaccines become available, young homosexual men attending genitourinary medicine clinics in London could be a suitable population in which to measure efficacy.

We thank Joan Richmond, Peter Luton, Pat Holland, Bina Patel, and Pradip Pandaya for their help in data collection, and Paddy Farrington for his help with the statistical analysis.


  • Funding None.

  • Conflict of interest None.


View Abstract