The HIV epidemic among gay men in the Netherlands isn't going to decline as long as large, persistent, self-sustaining, and, in many cases, growing sub-epidemics shifting towards new generations of gay men, according to a new paper in PLOS Medicine by Daniela Bezemer from HIV Monitoring Foundation, Amsterdam, the Netherlands, and Anne Cori from Imperial College London, UK, and colleagues.

Despite decades of awareness, billions of dollars in outreach and prevention of HIV infection, the HIV epidemic among gay men is resurgent in many Western countries. For a new study, the researchers performed phylogenetic analysis, which uses genetic differences among viruses isolated from different individuals to derive patterns of HIV transmission, and mathematical modeling to determine whether the introduction of new strains or the spread of already circulating strains is responsible for the ongoing HIV epidemic occurring among gay men in the Netherlands.

The researchers took advantage of a large national observational HIV cohort, the ATHENA cohort, which includes anonymized data from HIV-infected patients followed longitudinally since 1996 in the 27 HIV treatment centers in the Netherlands. The study included 5,852 participants among whom the researchers were able to identify 106 large HIV transmission clusters, 91 of which were primarily among gay men. The researchers found that at least 54 (59%) of these 91 transmission clusters were already circulating before 1996, when combination antiretroviral therapy was introduced, and that they have persisted to the present. Moreover, about a third of new HIV infections diagnosed among gay men since 1996 have involved viruses included in these long-lived clusters. 

Using mathematical modeling to estimate the effective reproduction number (the number of secondary infections per primary infection) for all the transmission clusters among gay men, the researchers found that reproductive numbers were around one for the whole study period. Thus, these clusters were self-sustaining and not contracting. Notably, HIV transmission clusters (particularly the newer clusters) tended to have higher reproduction numbers in recent years. Moreover, although the average age at diagnosis within each of the clusters increased over the study period at a rate of 0.45 years/year, the average age at diagnosis was lower at initiation of new clusters and only increased by 0.28 years/ year.

As with any modeling study, the validity of the results depends on the accuracy of assumptions used in the model, for example, that the probability of obtaining viral genetic information and the probability of surviving until 1996 were similar across clusters.