CROI 2015 Program and Abstracts

Abstract Listing

Poster Abstracts

240 Efforts to Characterize Community HIV Transmission Dynamics May Be Critically Dependent on Provision of Both Partner Services and Genetic Sequence Analysis Nella L. Green ; Christy Anderson; Sergei L. Kosakovsky Pond; Martin Hoenigl; David M. Smith; Sanjay Mehta; Susan Little University of California San Diego, San Diego, CA, US Background: Genetic analysis of HIV-1 sequences has become the gold standard for inferring HIV transmission networks. However, the added value of epidemiological links supplied via partner counseling and referral services (PCRS) is unclear. Methods: We examined bulk pol sequence and epidemiological data from the San Diego Primary Infection Resource Consortium (PIRC) collected between 1996 and 2013. PCRS services were routinely provided to recently HIV infected persons to characterize epidemiologically linked partnerships (named partners). Two individuals whose nucleotide HIV-1 sequences differed by at most 1.5% (Tamura-Nei 93 distance) were considered genetically linked. We compared two types of putative transmission pairs: Group 1: genetically and epidemiologically linked, and Group 2: genetically, but not epidemiologically linked. Results: 591 newly HIV infected persons were identified in the PIRC cohort. Provision of PCRS yielded 184 epidemiologically linked partnerships, of which 52 (28%) were also genetically linked (Group 1). Of the remaining 132 partnerships, 93 were linked to HIV-negative individuals and 39 (21% of total) were linked to HIV-infected persons whose genetic sequences differed. Sequence analysis alone identified 459 dyads (i.e. Group 2). 337 individuals made up the 511 pairs (Groups 1 + 2) included in this analysis. There were no significant differences comparing dyad concordance between Groups 1 and 2 with regard to race, ethnicity, income, CD4, viral load, and recent or current sexually transmitted infections. Partner age differences were greater in Group 2 (p = 0.012). The elapsed time between identification of the index and their linked partner was significantly different (p<0.001) between groups (Group 1 median 15 days [IQR 7-46 days]; Group 2 median 532.5 days [IQR 222-1194 days]). Conclusions: Sequence analysis and PCRS may identify a uniquely different population than that identified with sequence analysis alone. As expected, only a minority (28%) of PCRS pairs were corroborated by genetic data, yet these partnerships yielded a more rapid linkage identification (15 days vs 532 days) and also captured individuals more similar in age, as compared to partnerships identified solely by genetic analysis. We conclude that the primary and valuable benefit of PCRS is to yield recently connected putative transmission pairs that may represent more attractive targets for interventions and are more likely to represent direct transmission events. 2:30 pm– 4:00 pm Transmission Clusters 241 Growth and Geographic Spread of HIV Transmission Clusters, United States, 2007-2012 Alexandra M. Oster 1 ; JoelWertheim 2 ; Ellsworth Campbell 1 ; Angela L. Hernandez 1 ; Neeraja Saduvala 2 ;WilliamM. Switzer 1 ; M. Cheryl Ocfemia 1 ; Anupama Shankar 1 ; H. Irene Hall 1 1 US Centers for Disease Control and Prevention (CDC), Decatur, GA, US; 2 ICF International, Atlanta, GA, US Background: Molecular epidemiology can be used to identify clusters of persons with genetically related viruses. Clusters that continue to grow over time likely represent ongoing transmission and are potential points for intervention. Predicting which clusters are likely to grow could guide the appropriate allocation of limited prevention resources toward ensuring viral suppression and stemming transmission. Methods: We aligned HIV-1 sequences (1/person) reported to the U.S. National HIV Surveillance System to a reference sequence, conducted pairwise comparison of 70,669 sequences, and constructed an HIV transmission network of sequences with Tamura-Nei genetic distance ≤ 1.5%. We used HIV diagnosis year to construct the network over time and assess changes during 2007–2012. For clusters with ≥ 5 persons in 2007, we characterized growth through 2012 and used multivariable logistic regression to examine potential predictors of various levels of growth, including cluster size in 2007, cluster growth from 2006 to 2007, and demographic/risk characteristics. Finally, we described geographic characteristics of the largest clusters. Results: From 2007 to 2012, the number of persons with sequences and the number of clusters grew substantially (Table). In 2007, 177 clusters contained ≥ 5 persons. By 2012, these 177 clusters grew 235% overall, from 1,492 to 3,500 persons. Twenty-nine (16%) clusters grew ≥ 200% during 2007–2012, representing 54% of growth among clusters of ≥ 5 persons. Clusters that grew ≥ 200% did not differ from clusters that grew <200%with respect to size in 2007, growth from 2006 to 2007, racial/ethnic makeup, or the percentage of men who have sex with men (MSM). However, higher percentage of persons aged 13–19 years was associated with growth ≥ 200% (p=0.048). Growth ≥ 100% (n=76 clusters) was associated with rate of cluster growth from 2006 to 2007 (p=0.01) and higher percentage of MSM (p=0.04). Of the 8 largest clusters (size=79–155 in 2012), 5 consisted nearly exclusively (>90%) of persons living in the same state at diagnosis, one included persons from 7 Southern states, and two included persons from across the United States. WEDNESDAY, FEBRUARY 25, 2015 Session P-B3 Poster Session Poster Hall

Poster Abstracts

Characteristics of HIV transmission clusters, 2007 and 2012 Conclusions: We found substantial growth of clusters over a 5-year period. The best predictors of cluster growth ≥ 100%were growth in the previous year and the percentage of MSM. Rapid growth ( ≥ 200%) was best predicted by the percentage of adolescents. These data suggest that monitoring the growth and composition of clusters can help to identify prevention priorities.

218

CROI 2015