CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

Conclusion: HCV DAA treatment for HIV/HCV coinfected individuals is likely cost-effective in Myanmar. A simplified treatment protocol and/or lower drug costs could enhance cost-effectiveness. 591 TEMPORAL PATTERNS IN HCV PHYLOGENETIC CLUSTERING AMONG PWID IN BALTIMORE, MD Jada Hackman 1 , Oluwaseun Falade-Nwulia 2 , Eshan U. Patel 2 , Shruti H. Mehta 2 , Zach Downing 2 , Gregory D. Kirk 2 , Jacquie Astemborski 2 , Stuart C. Ray 2 , David L. Thomas 2 , Oliver Laeyendecker 1 1 NIAID, Bethesda, MD, USA, 2 Johns Hopkins University School of Medicine, Baltimore, MD, USA Background: HCV infection occurs in 30-90% of people who inject drugs (PWID). Phylogenetic analysis can be used to inform strategies to interdict transmission. This study examines patterns in HCV phylogenetic clustering overtime among PWID in Baltimore city. Methods: Community-based PWID were prospectively recruited for The AIDS Linked to the IntraVenous Experience (ALIVE) cohort in Baltimore, MD. Viral RNA underwent Polymerase Chain Reaction with primers targeting the 5’ end of the envelope-1 region and sequenced using Sanger Sequencing methods. There were 820 HCV RNA+ participants from 1988-1989 and an additional 512 unique HCV RNA+ participants from 2005-2016. Networks were rendered at a 4% genetic distance threshold using HIV-TRACE and participants were geographically mapped using Microreact. Prevalence ratios (PR) and 95% CIs of being in a cluster (≥2 participants) were calculated using Poisson regression with robust variance. Results: There were 15 clusters found among the participants in 1988-89 and 22 clusters identified in 2005-16. In both time periods, two large genotype 1a clusters were observed with 586/716 (82%) in 1988-89 and 113/302 (37%) in 2005-2016. When combining data from 1988-89 with 2005-16, the two large genotype 1a clusters were maintained (Figure). Participants from 2005-16 (59% [303/512]) were less likely to be in a cluster compared to the participants from 1988-89 (87% [716/820]) independent of HIV status, age, sex, race, zip code (adjusted PR, 0.71 [95% CI, 0.64-0.79]). The percentage of individuals in a cluster was consistently lower across all two-year intervals in the 2005-16 period in comparison to the 1988-89 two-year interval. Similar findings were observed when stratifying the analysis by genotype 1a and 1b. Among the clusters, there was a greater number of linkages among the 1988-89 individuals (median, 28 [IQR, 9-78]) compared to 2005-16 individuals (median, 5 [IQR, 1-16.5]; (p<0.001)). Conclusion: We observed greater cluster diversity in the participants recruited in 2005-16 indicative of a less connected network of individuals sharing transmission risk, though major viral strains did persist over time in this cohort.

592 COMPLEX PHYLODYNAMICS OF HCV AMONG PWID IMPACT INFERENCE OF TRANSMISSION NETWORKS Rebecca Rose 1 , Christopher W. Rodriguez 1 , James J. Dollar 1 , Susanna Lamers 1 , Guido Massaccesi 2 , William Osburn 2 , Stuart C. Ray 2 , David L. Thomas 2 , Andrea Cox 2 , Oliver Laeyendecker 2 1 Bioinfoexperts, LLC, Thibodaux, LA, USA, 2 Johns Hopkins University, Baltimore, MD, USA Background: The vast majority (up to 80%) of new HCV infections in high- resource countries are among people who inject drugs (PWID). Uncovering the common sources of infection and transmission networks is critical for effective intervention. Here, we study transmission dynamics in a highly-networked Baltimore cohort of PWID using longitudinally-sample HCV sequences. Methods: Subjects were enrolled in the “Baltimore Before and After Acute Study of Hepatitis” (BBAASH) cohort, composed of PWID followed prospectively from 1996-2016. All subjects were seronegative at enrollment. We used a subset of 89 subjects who acquired HCV during the study with sequences from>1 visit over 1-14 years. Viremia status (RNA+/-) was available for an average of 14 visits per individual. Bulk HCV sequences from the E1 region (H77 nt 943-1288) were studied, with a final alignment comprising 743 sequences. We used HIV-TRACE to assign clusters. Maximum likelihood (ML) trees for each cluster were inferred using RaxML. Results: Thirty clusters contained >1 subject whose sequences had <3% genetic distance. Trees showed that three clusters each contained a pair of subjects (A-C) who shared identical sequences, although detected 7-10 years apart. A longitudinal plot of viremia and cluster assignments (i.e. “variants”) showed that in all three pairs, the subject with later detection of the shared variant previously had an RNA- visit and/or detection of a different viral variant, while in two pairs, the subject with earlier detection of the shared variant subsequently had both an RNA- visit and detection of a different viral variant (Figure). Conclusion: HCV infection among PWID is dynamic and the resulting viral sequences reveal complex patterns of shared infection and superinfection from multiple sources. These data suggest that HCV transmission events are likely underestimated in most phylodynamic models.

Poster Abstracts

CROI 2019 223