Phylogenetic and demographic characterization of HIV-1 transmission in Madrid, Spain

The combination of phylogenetic analyses of HIV sequences with patients' demographic data allows us to understand local HIV transmission, a necessary knowledge for designing prevention strategies. The Community of Madrid represents a challenge for the control of HIV epidemic in Spain given its high HIV prevalence and increasing proportion of immigrant people among HIV-infected population.

We applied maximum likelihood methods and Bayesian Markov chain Monte Carlo (MCMC) inference using the program BEAST to a set of HIV-1 pol sequences from 1293 patients diagnosed in 1995-2010 in Madrid, Spain.

Two-hundred and thirty six patients (18.2% of the cohort) were included in 100 transmission chains using phylogenetic criteria, 67 (67%) belonging to HIV-1 subtype B and 33 (33%) to 11 different non-B strains, especially BG and BF recombinants. Most networks involved transmission between MSM (48/100). Half of non-B clusters (15/33) included at least one Spaniard. Sub-Saharan African patients presented a low linkage rate (9%) in contrast to Spanish (21%) and Latin American (25%) patients. Three clusters involving treatment-independent transmission of drug-resistance mutations were found.

One out of five HIV-infected patients in our cohort in Madrid was epidemically linked, mainly by transmission pairs. The inclusion in transmission networks was more likely for MSM, Spaniards and patients from Latin America. We found no evidence of self-sustained non-B epidemics due to the absence of large transmission chains with the exception of Cuban BG recombinants and CRF47_BF. However, the differences in transmission across variants are probably determined by the patient profile, especially the infection route.