CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

population in the Netherlands. This can be used to estimate the proportion of introductions that actually resulted in onward transmission and assess whether this differs between subtypes. Methods: The ATHENA observational cohort includes anonymized patient data of virtually all patients with HIV in care in the Netherlands. Polymerase sequences were available for 39% (8978) of patients in November 2014. Maximum likelihood trees were built by subtype in MEGA with 500 bootstraps, under a General Time Reversible model and five gamma distributed categories. We considered clusters with bootstrap values ≥70% including ≥2 MSM (indicating onward transmission), and those including ≥3 MSM (as indicative of ongoing established transmission clusters). The remaining MSM sequences were regarded as separate introductions, although some sequences formed clusters that were not statistically supported. The mean outbreak size per subtype was calculated as the number of sequences fromMSM divided by the number of introductions amongst MSM. Results: In ATHENA, 2131 of patients diagnosed between 1981-2014 with an HIV-1 non-B infection have a sequence available. Of these, 1467 self-reported being infected through heterosexual and 337 through MSM contact. Of the heterosexuals, 71% (1037) originated from Sub Saharan Africa, and 18% (262) from the Netherlands; of the MSM 8% (26) originated from Sub Saharan Africa and 66% (222) from the Netherlands. We built separate trees for subtypes A, CRF01AE, CRF02AG,C,D,F,G. The main results are summarized in table 1. The largest tree of 636 subtype CRF02AG sequences also contained most sequences fromMSM. The proportion of successful introductions was highest for subtype F, the tree had the highest proportion of sequences fromMSM, and was found to have the largest average outbreak size. Interestingly the median viral load at first hospital visit was significantly higher for MSM infected with subtype F compared to the MSM infected with other non-B subtypes. Conclusions: Non-B subtypes are being introduced into the MSM population in the Netherlands and have gone on to form national sub-epidemics with different mean outbreak sizes.

222

Spatiotemporal Dynamics of HIV-1 Transmission in France, 1999-2014 Antoine Chaillon 1 ; Asma Essat 2 ; Pierre Frange 3 ; Francis Barin 4 ; Jade GHOSN 5 ; Davey M. Smith 1 ; Christine Rouzioux 6 ; Cécile Goujard 2 ; Laurence Meyer 7 ; Marie Laure Chaix 8 1 Univ of California San Diego, San Diego, CA, USA; 2 Univ Paris Sud, INSERM CESP U1018, Paris, France; 3 EA7327, Univ Paris Descartes, Paris, France; 4 François-Rabelais Univ, Tours, France; 5 Hosp Hotel Dieu de Paris, Paris, France; 6 Necker Hosp, AP-HP, Paris, France; 7 INSERM, CESP U1018, Le Kremlin-Bicêtre, France; 8 INSERM U941, Univ Paris Diderot, Paris, France Background: Identifying and monitoring HIV transmission networks can be important in understanding the evolutionary patterns and geospatial spread of the epidemic. Here, we reconstructed the broad molecular epidemiology of HIV in France from individuals enrolled at the time of primary HIV-1 infection (PHI) in the national ANRS PRIMO Cohort over 15 years. Methods: Sociodemographic, geographic, clinical, and pol sequence data from 1356 cohort participants were collected between 1999 and 2014. Viruses mostly belonged to subtype B (71.5%) and CRF02_AG (15.3%). Network analysis was performed to infer genetic relationships, i.e. clusters, between HIV sequences. Bayesian coalescent-based methods were used to examine the temporal and spatial dynamics of identified clusters from different regions in France and the geographical patterns of viral spread and growth rate of HIV-1 subtype B and CRF02_AG subtypes. Results: Participants were mostly Caucasian (85.9%) and men (86.7%) who reported sex with other men (MSM, 71.9%). Overall, 387 individuals (29%) were involved in 78 dyads (n=156) and 231 participants fell in the remaining 42 clusters with 3 or more people (median size: 4, range 3-41). Compared to those who did not cluster (n=969), those in clusters with ≥3 people were more frequently men (98.3% vs 83%, p<0.01), MSM (88.3% vs 65.6%, p<0.01) and infected with CRF02_AG (27.3% vs 13.4%, p<0.01). Participants entering the cohort after 2011 (21.3%) clustered significantly more than participants entering before 2011 (20.6% in 2006-2010 and 10.1% before 2006, p<0.01). Phylogeographic analyses found a higher density of CRF02_AG clustering among participants residing in Paris versus those living outside Paris (p<0.01). Evolutionary analyses revealed that HIV-1 CRF02_AG epidemic arose more recently (time from the most common ancestor [TMRCA]=17.3 years, mean logistic growth rate [LGR]=0.051 year -1 , mean evolutionary rate=4.91x10 -3 subs/ site/year) and spread more rapidly among individuals with PHI in France than subtype B lineages (TMRCA=26.4 years, mean LGR=0.048 year -1 , mean evolutionary rate=2.89x10 - 3 subs/site/year). Conclusions: These analyses support the hypothesis of a recent and rapid rise of CRF02_AG within the French HIV-1 epidemic among MSM. Since CRF02_AG is historically associated with various parts of Africa, such results may help us understand the relationships between migration of human populations and HIV-1 diffusion.

Poster Abstracts

86

CROI 2016