CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

926 GEOGRAPHIC CHARACTERISTICS OF HIV GENETIC CLUSTERS AMONG NEWLY DIAGNOSED CASES IN NC Andrew E. Cressman 1 , Erika Samoff 2 , Victoria L.Mobley 2 , Simon Frost 3 , Kimberly Enders 1 , Shuntai Zhou 1 , Ronald Swanstrom 1 , Joseph J. Eron 1 , William C. Miller 4 , Myron S. Cohen 1 , Ann M. Dennis 1 1 University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, 2 North Carolina Division of Public Health, Raleigh, NC, USA, 3 Cambridge University, Cambridge, UK, 4 The Ohio State University, Columbus, OH, USA Background: Identifying both geographic clusters and genetic clusters are routine parts of HIV surveillance aiming to help focus prevention efforts. Integrating geographic and genetic analyses, especially beyond traditional surveillance borderlines, will inform the ability of geographic clustering to identify linked HIV transmission networks and help allocate prevention efforts. Methods: We assessed genetic clusters among those >=13 years old newly diagnosed with HIV in North Carolina (NC) between 2016 and 2019. Data of those with complete residential address information at the time of HIV diagnosis and either a pol sequence reported to NC or from sequence analysis of the diagnostic specimen received from the NC State Lab of Public Health were assessed (n=2,679 persons, approximately 69% of new diagnoses reported in NC). Clusters were constructed with <1.5% pairwise genetic distance (TN-93) between two members and restricted to >=5 total members for this analysis. Addresses were geocoded, and planar distances between those with genetically linked infections were calculated using address coordinates. Results: In total, we identified 67 genetic clusters involving 565 persons. Cluster members were mostly male (93%), African American (67%), and men who have sex with men (78%). The median cluster size was 7 members (range: 5-28), and most clusters were composed of a majority of members who lived in the same NC Field Services Unit Region (87%), of which there are seven, or county (58%), of which there are 100 with a median area of 436 square miles, at the time of diagnosis. The median geographic distance among linked members across all clusters was 25 miles (range: 0, 234), and 40 genetic clusters (60%) had a median geographic distance <25 miles among their linked members. Most clusters had maximum distances >100 miles (54%) and minimum distances <10 miles (97%) among linked members. Genetic clusters with median geographic distances >=25 miles among linked members were more likely to have members who were African American (71% vs. 63%), younger at HIV diagnosis (53% vs. 46% 18-24 years old), and in non-metropolitan (micropolitan, small town, or rural) areas (16% vs. 6%) compared to clusters with median geographic distances <25 miles among linked members. Conclusion: While most genetic clusters had a majority of members located within traditional surveillance borderlines of regions and counties, most also included greater geographic distances between genetically-linked infections.

6 Hospital Universitario La Fe, Valencia, Spain, 7 Hospital La Paz Institute for Health Research, Madrid, Spain, 8 Hospital San Pedro, La Rioja, Spain, 9 Clínica Universidad de Navarra, Pamplona, Spain, 10 Hospital Universitario de La Princesa, Madrid, Spain Background: The HIV-1 TRACE (TRAnsmission Cluster Engine) is a new computational tool to identify molecular transmission clusters in large databases. This approach is based on viral genetic relatedness to a reference sequence in order to construct and visualize the connections among clusters. Our objective was to identify transmission clusters in CoRIS cohort (2018 update) by using HIV-1 TRACE computational tool focusing on subtype B patients and to compare TRACE identified clusters with phylogenetic approaches. Methods: We used the RT available regions from newly HIV diagnoses in 2018 in CoRIS. HIV-1 TRACE (http://hivtrace.datamonkey.org/hivtrace) was used to estimate transmission clusters in 484 subtype B antiretroviral-naïve patients enrolled in the CoRIS cohort. Phylogenetic analysis was conducted by maximum likelihood method (ML) with bootstrap using the GTR+G as nucleotide substitution model. Sequences were phylogenetically analysed along with all the most similar sequences as identified by a BLAST search. Local transmission networks (LTNs) were defined as phylogenetic clusters including sequences from Spain at proportions >70%, receiving bootstrap value >70%. Results: HIV-1 TRACE results showed that 354 patients (73.1%, n=354/484) were not involved in any cluster and 130 patients (26.9%, n=130/484) were grouped in 54 clusters: 39 clusters with 2 nodes, 11 clusters with 3 nodes, 2 clusters with 4 nodes, 1 cluster with 5 nodes and 1 cluster with 6 nodes (range 2-6). Phylogenetic analysis revealed that 330 (68.2%, n=330/484) and 154 patients (31.8%, n=154/484) were involved in 63 clusters: 48 clusters with 2 nodes, 7 clusters with 3 nodes, 4 clusters with 4 nodes and 4 clusters with 5 nodes (range 2-5). Overall, the concordance between phylogenetic approaches and HIV-1 TRACE tool was 84.4%. The discrepancies were not observed only in the number of clusters, as previously described, but also in the distribution, since phylogenetic tools identified 8 clusters with more than 3 nodes and HIV-1 TRACE identified only 4 of these clusters. Conclusion: The implementation of HIV-1 TRACE is an easy to use tool and it allows identification of transmission clusters. Our results revealed that HIV-1 TRACE identified fewer clusters among B-subtype patients than traditional phylogenetic approaches. Those discrepancies were due to the non-use of a threshold in the patristic distances in phylogenetic analysis. 928 RECONSTRUCTION AND ESTIMATION OF DIRECTED HIV-1 TRANSMISSION USING DEEP SEQUENCES Nicholas Bbosa 1 , Deogratius Ssemwanga 2 , Alfred Ssekagiri 3 , Yunia Mayanja 2 , Ubaldo M. Bahemuka 2 , Janet Seeley 2 , Deenan Pillay 4 , Christophe Fraser 5 , Pontiano Kaleebu 2 , Oliver Ratmann 6 , for the PANGEA Consortium 1 MRC/UVRI and LSHTM Uganda Research Unit, Entebbe, Uganda, 2 MRC/UVRI and LSHTM Uganda Research Unit, Entebbe, Uganda, 3 Uganda Virus Research Institute, Entebbe, Uganda, 4 Africa Health Research Institute, Mtubatuba, South Africa, 5 University of Oxford, Oxford, UK, 6 Imperial College London, London, UK Background: Key population in Uganda are disproportionately affected by HIV-1 relative to the general population (GP). Serial crossectional surveys were carried out in several HIV-1 high-risk and general population cohorts of the MRC/UVRI & LSHTM Uganda research unit to generate near full length (NFL) deep sequences. The aim of this study was to perform a source attribution analysis in the sampled populations to assess the extent to which high-risk groups contribute to the HIV epidemic in the general population and to further inform location focused interventions in key populations. Methods: We used the phyloscanner program developed to phylogenetically infer transmission fromwith-in and between-host HIV genetic diversity to reconstruct directed HIV-1 transmission networks from NFL deep sequences (n=2,531) from communities of women at high risk to HIV (WHR), the fisherfolk (FF) and the general population (GP). We used the phyloflows package implemented in the R software to correct for sampling heterogeneity and estimate HIV-1 transmission flows between the three populations. Results: Of the 2,531 HIV-1 NFL deep sequences analyzed in phyloscanner,105 highly supported HIV transmission pairs were identified with phylogenetic evidence for the direction of transmission (criteria for linkage: >60% and >60% for one direction). Our observed transmission counts showed majority of HIV-1 transmissions to be intra-population [GP-GP (34%)>FF-FF (31%)>WHR-WHR (10%)]. Between populations, transmission counts were more prevalent from the GP to FF (11%) followed by those from the FF to the GP (10%) (Figure 1). An estimation of HIV transmission flows showed results that were comparable

Poster Abstracts

927 HIV TRACE VS PHYLOGENETIC ANALYSIS: UNRAVELING TRANSMISSION CLUSTERS IN SPAIN Carlos Guerrero Beltrán 1 , Evangelia G. Kostaki 2 , Luca Carioti 3 , Marta Alvarez 1 , Julian Olalla 4 , María Carmen Vidal Ampurdanes 5 , Marta Montero 6 , Silvia García- Bujalance 7 , Jose-Ramón Blanco 8 , Maria Rivero 9 , Lucio Jesus Garcia-Fraile Fraile 10 , Maria M. Santoro 3 , Dimitrios Paraskevis 2 , Federico García 1 1 Hospital Universitario San Cecilio, Granada, Spain, 2 University of Athens, Athens, Greece, 3 University of Rome Tor Vergata, Rome, Italy, 4 Hospital Costa del Sol, Marbella, Spain, 5 Hospital Universitario de Son Espases, Palma de Mallorca, Spain,

CROI 2020 348