CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

were genetically unrelated. In unadjusted analyses, membership in a cluster, was associated with younger age (Odds ratio (OR) 1.4 [95% Confidence Interval (CI) 1.2-1.7] per 10 year age decrease), female sex (OR 1.5 [95% CI 1.0-2.2]), HIV infection (OR 1.8 [95% CI 1.2-2.6]), and living in East (versus West Baltimore, OR 2.2 [95% CI 1.4-3.6]). In adjusted analyses, younger age (OR 1.4 [95% CI 1.0-1.8-] per 10 year age decrease), HIV infection (0R 1.6 [95% CI 1.1-2.4] and living in East (versus West Baltimore, OR 1.97 [95% CI 1.2-3.2]) remained independently associated with being in a cluster of related hepatitis C infections. Conclusion: In a cohort of PWID in Baltimore, 25% of participants were in large genetic clusters suggestive of HCV transmission networks. Targeted treatment of PWID in transmission networks should be explored as a means to enhance effectiveness of HCV TasP and elimination efforts. 582 UNRAVELLING HCV1A WHEREABOUTS IDENTIFIES A NEED FOR A PAN- EUROPEAN PREVENTION PROGRAM Lize Cuypers 1 , Bram Vrancken 1 , Lavinia Fabeni 2 , Velia C. Di Maio 3 , Milosz Parczewski 4 , Vladimir Chulanov 5 , Perpetua Gomes 6 , Anna Maria Geretti 7 , Christoph Sarrazin 8 , Cillian De Gascun 9 , Saleta Sierra 10 , Federico García 11 , Anne- Mieke Vandamme 1 , Francesca Ceccherini Silberstein 3 1 Katholieke University Leuven, Leuven, Belgium, 2 IRCCS Lazzaro Spallanzani, Rome, Italy, 3 University of Rome Tor Vergata, Rome, Italy, 4 Pomeranian Medical University, Szczecin, Poland, 5 Central Research Institute of Epidemiology, Moscow, Russian Federation, 6 Centro Hospitalar de Lisboa Ocidental, Lisbon, Portugal, 7 University of Liverpool, Liverpool, UK, 8 University Hospital Frankfurt, Frankfurt, Germany, 9 University College Dublin, Dublin, Ireland, 10 University of Cologne, Cologne, Germany, 11 Hospital Universitario San Cecilio, Granada, Spain Background: Viral factors can impair the efficacy of direct-acting antiviral (DAA) based therapies for the hepatitis C virus (HCV). Naturally occurring and treatment-emerging resistance-associated variants (RAVs) in the NS3 and NS5A genes have been observed to interfere with the action of DAAs, especially in case of HCV1a. Therefore, investigating the origins of new infections is of public health relevance. Methods: NS3 (n=2514) and NS5A (n=1957) sequences from HCV1a patients, including 1825 newly generated taxa from nine European countries (Belgium, Germany, Ireland, Italy, Poland, Portugal, Russia, Spain and the United Kingdom (UK)), were complemented with publicly available time and geo-referenced virus genetic data from around the world, among which sequences from four additional European countries (France, Netherlands, Sweden and Switzerland). A fast and scalable Bayesian phylogeographic approach was used to elucidate the population level transmission patterns through time and space using a model that allows for different migration rates depending on the direction of movement. To minimize the potentially biasing impact of the sampling process on the migration rate estimates, the latter were informed by both gene datasets simultaneously. Results: Mapping migration pathways for both genes shows extensive movements both across continents, with an almost exclusive role for the United States (US) in seeding HCV1a into Europe (74.6%), as well as within Europe. Within the European continent there are no clear source-sink relations, and the migration network becomes increasingly complex and diffuse over time. Particularly Germany seems to function as a hub for the introduction of HCV1a lineages in Europe, seeding mainly to France, Italy, the UK and Spain. There were no specific patterns observed for known NS5A RAVs. In line with previous findings, the NS3 variant Q80K was highly abundant in one of the two clades in which HCV1a segregates. Conclusion: In-depth phylogeography of the European HCV1a epidemic illustrates complex patterns of human migrations, without a clear differentiation of HCV1a patients at risk of acquiring RAVs. The existence of a pan-European migration network impairs the efficacy of national-based intervention programs. In turn, this indicates that a supra-national coordinated approach is needed to more quickly and more thoroughly avert the spread of HCV, including of strains that could restrict DAA treatment options. 583 NS5A RESISTANCE ASSOCIATES WITH ADVANCED LIVER FIBROSIS BUT NOT TRANSMISSION CLUSTERS Milosz Parczewski 1 , Ewa Janczewska 2 , Justyna Kordek 1 , Iwona Cielniak 3 , Arkadiusz Pisula 2 , Monika Bociąga-Jasik 4 , Władysław Łojewski 5 , Aleksandra Szymczak 6 , Łukasz Socha 1 , Magdalena Leszczyszyn-Pynka 1 , Anna Urbańska 1 , Ewa Siwak 3 , Nadine Lübke 7

1 Pomeranian Medical University, Szczecin, Poland, 2 ID Clinic, Mysłowice, Poland, 3 Hospital for Infectious Diseases, Warsaw, Poland, 4 Jagiellonian University, Kraków, Poland, 5 Regional Hospital in Zielona Góra, Zielona Góra, Poland, 6 WrocławMedical University, Wrocław, Poland, 7 Heinrich Heine University Hospital, Düsseldorf, Germany Background: In the era of HCV treatment with directly acting antiviral (DAA) regimens presence of drug resistant variants (RAVs) negatively affecting treatment efficacy remains a concern. This study aimed to characterize pretreatment NS5A RAV frequency among Polish genotype 1 HCV-monoinfected and HIV/HCV-coinfected patients including transmission patterns and association with liver fibrosis. Methods: NS5A sequences from 388 DAA-naive G1 infected individuals [54, 13.92% genotype 1a (G1a) and 334 (86.08%) 1b (G1b)] were obtained by population sequencing. Within this dataset 122 (31.44%) samples were obtained from HIV/HCV coinfected [including 55 (14.17%) HIV patients with the documented history of acute hepatitis C (AHC)] and 388 (68.55%) from HCV monoinfected cases. RAVs were called using the geno2pheno algorithm. A maximum likelihood methodology was used to identify the clustering, separately for G1a and G1b sequences, with clusters defined by posterior values >0.85 for both G1a and G1b. Liver fibrosis was assessed based on histopathology or ultrasound elastography (available for 190 cases, METAVIR scale). For statistics Chi2 test or two-sided Fisher’s exact test were used, as appropriate. Results: NS5A RAVs were found among 35/388 (9.02%) sequences, being similarly distributed between G1a (3/54, 5.56%) and G1b (32/334, 9.58%) (p=n.s). Variation in 31 and 93 NA5A codon positions was only found in G1b cases (14/334 (4.19%) for L31I/F/M and 18/334 (5.39%) for Y93H). Among HIV/ HCV coinfected RAVs were present in 2/122 (5.71%) cases compared to the 33/233 (12.41%) HCV monoinfected individuals (p<0.001), being absent among AHC patients. Increased frequency of NS5A RAVs was found among cases with advanced fibrosis (6/40, 15.0% for F3-F4, vs. 7/150, 4.67% for F0-F2, p=0.02) and liver cirrhosis (6/28, 21,43% for F4 vs. 7/162, 4.32% for F0-F3, p<0.001). Higher prevalence of Y93H variant among cirrhotic patients was also noted (3/28, 10.71% vs. 3/162, 1.85%, p=0.04). No clustering of NS5A drug resistance variants was observed neither for G1a nor G1b sequences. Conclusion: Selection of NS5A RAVs including Y93H associated with more advanced liver fibrosis and represents de novo selection of variants rather than transmission of drug resistant strains. This is reflected by lack of clustering for sequences containing RAVs and absence of this resistance among cases with AHC. Higher frequency of NS5A RAVs in cirrhotic patients may associate with lower virologic response rates to DAA treatment. 584 GENETIC CORROBORATION OF HCV TRANSMISSION AMONG YOUNG ADULT INJECTING PARTNERSHIPS Damien C. Tully 1 , Judith Hahn 2 , David J. Bean 1 , Jennifer Evans 2 , Kimberly Page 3 , Todd M. Allen 1 1 Ragon Institute of MGH, MIT and Harvard, Cambridge, MA, USA, 2 University of California San Francisco, San Francisco, CA, USA, 3 University of NewMexico, Albuquerque, NM, USA Background: The current opioid epidemic across the US has fueled a dramatic surge in the rate of HCV infections among young persons who inject drugs (PWID). Central to designing and monitoring effective HCV prevention measures is the identification of partners who transmitted HCV. Viral genetic sequence data can then be used to infer the transmission network by reconstructing the path that links individuals who are infected with genetically similar viruses. Here, using a cohort of sero-discordant injecting partners we evaluated the degree of overlap between self-reported injecting partnerships and inferred genetic networks. Methods: We conducted a prospective study of injecting partnerships, utilizing a large community-based epidemiologic study of HCV infection in young PWIDs in San Francisco (“UFO” study), in which index cases (chronic HCV positive) were invited to bring their current regular injecting partner(s) who were HCV RNA sero-discordant or had evidence of recent HCV infection. Partnerships in which new HCV infection occurred were subjected to Illumina deep sequencing of 2 distinct genomic regions (Core-NS2 and NS5B). Transmission clusters were identified using maximum-likelihood methods with stringent genetic distance and bootstrap support thresholds. Results: Deep sequencing of 51 specimens from 25 partnerships in which new HCV infections occurred identified 15 phylogenetic clusters at a conservative genetic threshold of 0.01 substitutions/site. The majority of clusters, 71%

Poster Abstracts

CROI 2018 214