Increased HIV incidence among men who have sex with men in Rome

Among 976 men who have sex with men (MSM) who had undergone repeat HIV testing between 1984 and 2003 at a sexually transmitted infection clinic in Rome, Italy, we observed a dramatic increase in HIV incidence in 2002 and 2003, with the cumulative incidence for 2000-2003 being twice as high as that for 1984-1995, and significantly higher than that for 1996-1999. This trend suggests the need for interventions aimed at encouraging behavioural changes among MSM.

Several studies have reported substantial increases in the incidence of early syphilis and gonorrhoea among men who have sex with men (MSM) in Europe and north America ^[1,2], indicating that there has been an increase in at-risk behaviour for sexually transmitted infections (STI) in this population group. This has been confirmed by reports that MSM have been increasingly adopting behaviours at risk for HIV infection, leading to an increase in the incidence of HIV ^[3,4]. At the STI clinic of the San Gallicano Institute in Rome, Italy, the city's largest STI clinic, we observed that the number of syphilis diagnoses among MSM has dramatically increased since 2001 ^[5], leading us to evaluate the recent trends in the incidence of HIV infection among non drug-using MSM attending this clinic.

The incidence survey was conducted as a retrospective longitudinal study among individuals who in the past 20 years have undergone HIV testing more than once at the STI clinic. Since 1984, all MSM who attended the clinic and were tested within the preceding 6 months have been routinely offered HIV counseling and testing. Of over 1200 HIV tests performed annually, approximately 25.0% are performed for MSM. To identify all MSM repeatedly tested for HIV antibodies between June 1984 and December 2003, computerized medical records were systematically reviewed, and all individuals whose first test result was negative and who were re-tested at least once during the study period were selected.

Seroconverters were defined as those individuals with a negative HIV test followed by an HIV-positive test within the study period. The diagnosis of HIV infection was performed by an enzyme-linked immunosorbent assay and confirmed using a Western blot.

According to the methods applied in a previous similar study ^[6], HIV incidence was estimated by dividing the number of seroconversions by the number of person-years (py) of follow-up, and was expressed as the number of newly acquired infections per 100 py. A Poisson distribution was used to estimate 95% confidence intervals (CI) of incidence rates.

The person-years of follow-up were calculated as the sum of the intervals of the time between the first recorded HIV-negative test and the last test, whether positive or negative.

To determine HIV incidence by calendar year and to describe temporal trends, we assumed the year of the first HIV-positive test as the year of seroconvertion, and annual incidence rates were calculated by dividing the number of seroconvertions observed in a calendar year by the person-years calculated for the same year.

Overall, 976 MSM were included in the study and followed for a total of 4621 py. During the study period, 125 individuals seroconverted. The median of follow-up time was similar when comparing seroconverters with individuals who were persistently HIV negative (4.8 versus 4.6 years, respectively). The mean of HIV repeated tests was 5.2 (± 3.6) for seroconverters and 4.2 (± 4.0) for persistently HIV-negative individuals. The median age at first test was 29 years (range 16-62) for seroconverters and 32 years (range 13-79) years for non-seroconverters. The cumulative incidence was 2.70 per 100 py (95% CI 2.47-3.54). The median of the seroconversion interval was 1.6 years (interquartile interval 25%, 0.7, 75%, 4.9); non-significant changes were observed among the seroconversion intervals over time, particularly during the last years of the study.

Between 1985 and 1996, the annual incidence peaked and then decreased greatly every 5 years (Fig. 1). After 1996, however, the annual incidence showed a slightly different pattern. After a progressive decrease in 1997 and 1998, a new peak was observed in 1999, followed by only a slight decrease in 2000 and 2001. After this, the incidence rapidly increased, reaching 5.48 in 2002 and 11.90 in 2003 (Fig. 1). Poisson regression analysis showed a statistically significant increase in HIV cumulative incidence in the period 2000-2003, compared with the period 1984-1995 (incidence rate ratio 2.20, P < 0.001; Table 1). The observed increase was also confirmed after adjusting for age (data not shown).

Fig. 1

Table 1

Our findings are consistent with data from other longitudinal studies among MSM in north America and Europe ^[7-9] and with recent data provided by surveillance systems in the United Kingdom ^[10] and the United States ^[11]. The increase in HIV incidence suggests that since 1996 there have been dramatic changes in at-risk sexual behaviour among MSM in Rome. This is also confirmed by the significant increase in the occurrence of infectious syphilis observed among MSM attending our centre between 2001 and 2003 ^[5]. However, in interpreting the results of this study, it should be considered that they cannot necessarily be generalized to all MSM living in Rome, especially those who do not undergo HIV testing, and that repeat testers may differ from those who do not repeat the test.

The results of the study stress the need for continued surveillance of HIV infection among MSM, and for additional analyses for evaluating the determinants of the observed increase in incidence. Moreover, the recent increase in HIV incidence emphasizes the need to maximize efforts aimed at encouraging MSM to make behavioural changes.