CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

215 Infer and Characterize a Transmission Network in an Opioid-Driven HIV-1 Outbreak Ellsworth M. Campbell 1 ; Romeo R. Galang 2 ;Walid Heneine 2 ;William Switzer 2 ; Philip Peters 2 ; MichaelT. Spiller 2 ; Hongwei Jia 2 ; Silvina Masciotra 2 ; for the HIV Outbreak Investigation Team 1 Oak Ridge Inst of Sci and Educ, Decatur, GA, USA; 2 CDC, Atlanta, GA, USA Background: In January 2015, investigation of a sudden upsurge in new HIV-1 infections in a rural county in Indiana linked to injection drug use (IDU) identified a large outbreak (n=181). Here we describe the integration of epidemiologic and laboratory data to infer and characterize the transmission network to inform future prevention efforts. Methods: Serum specimens were used to determine recency of infection and to obtain HIV-1 polymerase ( pol ) sequences for genetic analysis. Putative undirected transmission links were drawn between sequences with genetic distance <1.5%. Standardized interviews were conducted with HIV-positive patients to collect high-risk behavior and contact data. High-risk contacts (sexual, IDU, or both) were considered as continuous risk factors, regardless of reporting direction. The reported contact network and inferred transmission network were compared. A decision tree was generated and logistic regression performed concerning contact type and occurrence with respect to infection status. Results: HIV-1 pol sequences were obtained from 157 persons epidemiologically linked to the outbreak. Phylogenetic analysis inferred a monophyletic pol clade with limited diversity. Of 123 specimens available for avidity testing, 113 (91.9%) were recent (<8 months prior to collection). Network analysis showed that each type of high-risk contact was correlated with HIV infection (p<0.01). However, 82.3% of potential transmission events corresponded to a reported IDU contact, as opposed to 11.0% for reported sexual contacts. The decision tree revealed that the likelihood of infection for persons with ≥4 recent injection partners was 96.2% (101/105). 77.4% (24/31) of persons with multiple sexual partners with whom they also share injection equipment, and who also had 2 or 3 additional recent injection partners, were infected. In contrast, 20% (6/30) of persons with the same number (2 or 3) of recent injection partners and who shared syringes with only one sexual partner were infected. Conclusions: A single HIV-1 strain was detected in this outbreak of HIV among persons who inject drugs, suggesting recent and rapid transmission. Comparison of reported contact and transmission networks reveal that IDU drove transmission and led to explosive growth of the outbreak. Integration of sequence and epidemiologic data facilitated a better understanding of transmission dynamics in a rural community of persons who inject prescription opioids.

216 HIV Phylodynamics in North Carolina: Detecting “Active”Clusters for Intervention Ann M. Dennis 1 ; Stephane Hue 2 ; Joseph Sebastian 3 ;Victoria Mobley 4 ;WilliamMiller; Joseph J. Eron 1 Univ of North Carolina at Chapel Hill, Chapel HIll, NC, USA; 2 London Sch of Hygiene & Trop Med, London, UK; 3 Lab Corporation of America, Research Triangle Park, NC, USA; 4 North Carolina DHHS, Raleigh, NC, USA Background: Phylogenetic analyses of HIV sequences can be used to monitor epidemic trends and help target prevention through the identification of transmission clusters. Characterizing such clusters in NC, where >1,500 new HIV diagnoses are reported annually despite widespread prevention, is needed. We evaluated temporal cluster expansion and epidemic growth to detect “active” clusters as potential intervention targets. Methods: We analyzed 15,247 HIV-1 pol sequences (each from an individual patient) derived from genotypes sampled in NC from 1997-2014 by the largest commercial laboratory. Putative transmission clusters were clades with high branch support (≥0.90) and maximum pairwise genetic distance <3.5% on a maximum-likelihood (ML) tree with the GTR model. Clusters were confirmed and transmission dynamics evaluated using BEAST with Bayesian skyline and relaxed molecular clock priors, inferring internal node ages (for each cluster ≥10 members). Reproductive numbers (R 0 ) for individual large clusters were estimated using birth-death models. Results: Most samples were frommen (71%) with median age 40 years (IQR 32-48), and HIV-1B (98%). 7,647 (50%) sequences were identified in 2,318 clusters (median size 3 members). 74 large clusters (≥10) involved 1,062 patients and were highly supported in BEAST. Compared to non-clustered patients, large clusters were associated with sex (81% vs. 70%men), more recent sampling year (65% vs. 46%≥2009), younger age (49% vs. 15%≤30 years), and sampling from Raleigh metropolitan area (54% vs. 42%) [all P <0.001]. However, 16 (22%) large clusters were ≥50%women and had earlier origin dates (1996 vs. 1999) and longer time spans (18 vs. 15 years) than male dominated clusters. The 8 largest clusters (n=22-36 members) originated between 1997 and 2004 and spanned a mean 12.5 years with 2.2 transmissions/year. The median estimates of R 0 ranged 1.2-4.1. Most (75%) had highly supported internal nodes 2010-2014, indicating cluster expansion. Most of these clusters were homogenous by sampling region, and all but one were >90% men. Conclusions: Phylodynamics revealed transmission cluster expansion in this densely sampled region and allowed estimates of R 0 to help identify active clusters contributing to recent transmission. While large clusters originated >10 years ago, most continue to expand, particularly among young men. Enhanced prevention interventions to active clusters are planned to halt further propagation.

Poster Abstracts

83

CROI 2016

Made with FlippingBook - Online catalogs