A epidemia de HIV-1 está mudando em São Paulo, com uma crescente preocupação no número de jovens de homens que fazem sexo com homens (HSH). A epidemia é dominada pelos subtipos B, F1, porém tem aumentado a prevalência do subtipo C. O objetivo do estudo foi identificar e caracterizar as redes de transmissão em São Paulo, comparando entre os subtipos epidêmicos, fatores de risco e exposição à terapia antirretroviral. Nós compilamos dados clínicos, epidemiológicos e as sequencias virais dos indivíduos acompanhados no Instituto Adolfo Lutz de Jan 2004 a fev de 2015. Após o controle de qualidade, 2.260 sequencias parciais do gene pol foram subtipadas previamente como B, C e F1 e então incluídas no estudo. 2.107 sequencias únicas controle foram selecionadas pelo blast contra as nossas sequencias no banco de dados de Los Alamos e banco de dados de Portugal. Todas as sequencias foram investigadas para a identificação de mutação de resistência (TDRM). Árvores filogenéticas foram construídas em RaxML para cada subtipo. Nós usamos Cluster Picker para analise da dinâmica de transmissão de acordo com os parâmetros de distância genética e bootstrap. Análises estatísticas foram utilizadas para identificar possíveis correlações dos clusters. 414 (18,3%) da nossa população foi incluída em cluster (2-12 indivíduos). A taxa de cluster não diferiu entre B e F1, no entanto as sequencias do subtipo C se agruparam duas vezes mais (p<0.001). Mais clusters foram identificados entre a população HSH quando comparada a de heterossexuais independente do subtipo (p<0.001). A taxa de TDRM foi maior em cluster que fora de cluster nos subtipos B e F1 (p<0.001 e p=0.009), respectivamente. Apesar do alto número de clusters no subtipo C verificamos baixa prevalência de TDRM. Os indivíduos que estavam em clusters eram 5 anos mais jovens que os de fora de cluster nos subtipos B e C (p<0.001 e 0.02, respectivamente). Nossos resultados indicam que, independente do subtipo, a epidemia de HIV-1 em São Paulo é auto-sustentada pelos pacientes HSH virgens de tratamento. O subtipo C apresentou maior proporção de clusters. No entanto, mais análises são necessárias para clarificar se isto implica em alta taxa de transmissão do subtipo em São Paulo.

The HIV-1 epidemic in Sao Paulo is changing, with a worrying increase in the number of new infections in young men and in the Men who have Sex with Men (MSM). The epidemic is dominated by subtypes B and F1, but theres been a recent increase in the prevalence of subtype C. The aim of this study was to identify and characterize HIV-1 transmission networks in Sao Paulo, comparing across subtype epidemics, risk factors, and HAART exposure therapy. We gathered epidemiological, clinical and viral sequence data from HIV-1 infected individuals followed in Adolfo Lutz Institute/SP-BR from Jan 2004 through Feb 2015. After quality control, 2,260 sequences of the partial pol gene were subtyped as previously reported as B, C and F1 were included on this study. 2,107 unique background control sequences were selected by blasting our sequences against Los Alamos database and the Portuguese HIV-1 database. The dataset was screened for Drug Resistance Mutations to identify transmitted drug resistance mutations (TDRM). Maximum-Likelihood phylogenetic trees were built in RaxML for each subtype separately. We used Cluster Picker to analyze transmission dynamics according the thresholds: genetic distance (0.06) and bootstrap over 90%. Statistical analyzes were performed to identify possible correlates of clustering. Categorical variables were compared using Chi-square or Fishers exact test and quantitative variables were analyzed using Mann-Whitney test (SPSS). 414 (18,3%) of our population were included in clusters (range: 2-12 individuals). The rate of clustering did not differ between subtypes B and F, however more subtype C sequences (40%) significantly clustered together (p<0.001). Furthermore, more clusters were significantly found in the MSM group when compared to heterosexuals for all subtypes, (p<0.001). Also, drug naïve patients were more likely to be in clusters when compared to treated patients in all subtypes (p <0.001). TDRM were more prevalent in clustering than in non-clustering in subtypes B (p <0.001) and F1 (p =0.009), and more related to MSM group. Despite the higher number of clusters, subtype C presented a lower prevalence of TDRM, although without significant difference between cluster and non-cluster. Clustering individuals were also 5 years younger than non-clustering individuals for subtypes B and C (p<0.001 and 0.02, respectively). Our results indicate that, regardless of subtype, the HIV-1 epidemic in Sao Paulo is self-sustained by treatment-naive MSM patients, who transmit the infection before starting treatment. Subtype C presents a higher proportion of patients that cluster together. However, further analyses are necessary to clarify whether this implies a higher transmission rate of this subtype in Sao Paulo.