CROI 2018 Abstract eBook

Abstract eBook

Oral Abstracts

133 PREDICTORS OF KS-IRIS IN MILD/MODERATE AIDS-KS DURING ART IN LOW-RESOURCE SETTINGS Thomas Campbell 1 , Minhee Kang 2 , Trijn Umbleja 2 , Susan E. Krown 3 , Agrey Bukuru 4 , Jeffrey N. Martin 5 , Catherine Godfrey 6 , Mulinda Nyirenda 7 , Naftali Busakhala 8 , Margaret Borok 9 , Deborah Langat 10 , Brenda Hoagland 11 , Noluthando Mwelase 12 , Mina C. Hosseinipour 13 1 University of Colorado, Aurora, CO, USA, 2 Harvard University, Cambridge, MA, USA, 3 AIDS Malignancy Consortium, New York, NY, USA, 4 Joint Clinical Research Centre, Kampala, Uganda, 5 University of California San Francisco, San Francisco, CA, USA, 6 DAIDS, NIAID, Rockville, MD, USA, 7 Malawi College of Medicine-Johns Hopkins University Research Project, Blantyre, Malawi, 8 Moi University, Eldoret, Kenya, 9 University of Zimbabwe, Harare, Zimbabwe, 10 Kenya Medical Research Institute/ Walter Reed Project, Kericho, Kenya, 11 Institute de Pesquisa Clinica Evandro Chagas, Rio de Janeiro, Brazil, 12 Wits Reproductive Health and HIV Institute, Johannesburg, South Africa, 13 University of North Carolina Project–Malawi, Lilongwe, Malawi Background: Kaposi sarcoma-associated immune reconstitution inflammatory syndrome (KS-IRIS) is characterized by abrupt worsening of KS after antiretroviral therapy (ART) initiation. Incidence and risks for KS-IRIS are not well-established, particularly in resource-limited settings, and well-defined diagnostic criteria have rarely been applied. In a planned secondary analysis, we evaluated the incidence and baseline factors associated with KS-IRIS in a multinational clinical trial. Methods: 190 chemotherapy and ART-naïve, HIV-1-infected adults with mild-to-moderate KS in Africa and South America were randomized 1:1 to ART (TDF/FTC/EFV) alone or with delayed oral etoposide (ET; As-Needed) vs ART plus up to 8 cycles of ET (Immediate). KS-IRIS was prospectively defined as clinical progression of KS (PD) within 12 weeks of ART initiation associated with CD4+ increase ≥50 cells/uL and/or decreased plasma HIV RNA ≥0.5 log 10 c/mL. KS response criteria were strictly defined and PD confirmed by an Independent Endpoint Review Committee. Analysis was restricted to participants who had week 48 data potential. Cumulative incidence was calculated using Kaplan- Meier methods and Gray’s approach to compare between arms was used to account for competing risks, with censoring at wk 12. Wk 48 outcome was compared between participants with/without KS-IRIS in the As-Needed arm. Potential baseline characteristics associated with time to KS-IRIS were assessed using cause-specific Cox regression models and all subsets variable selection. Results: The cumulative incidence of KS-IRIS was 21/94 (23.1%; 95%CI 15.0, 32.3) in the As-Needed arm vs 7/96 in the Immediate arm (7.4%; 95%CI 3.2, 13.9; P=0.003); among KS-IRIS median baseline CD4+ 217 vs 268 cells/uL and plasma HIV RNA 5.09 vs 5.54 log 10 c/mL. Covariates associated with increased hazards of KS-IRIS in the final model are provided in the Table. Among evaluable participants in the As-Needed armwho did not initiate alternate KS treatment), there were trends toward more wk 48 PD (73% vs 28%) and less response (27% vs 45%) among 11 participants with KS-IRIS compared to 55 without KS-IRIS. Death by wk 48 among evaluable participants was 11%with KS-IRIS vs 7% without KS-IRIS. Conclusion: KS-IRIS was common among persons with AIDS-KS initiating ART in low-resource settings; early ET reduced KS-IRIS risk. The presence of raised skin lesions, low albumin and low performance status should be considered as markers for higher risk of KS-IRIS during ART.

Oral Abstracts

132 STATIN EXPOSURE IS ASSOCIATED WITH DECREASED RISK OF CANCER Roger Bedimo 1 , Fatma Shebl 2 , Keith M. Sigel 3 , Sheldon T. Brown 4 , Kristina Crothers 5 , Matthew B. Goetz 6 , Amy C. Justice 7 , Janet Tate 7 1 VA North Texas Health Care Center, Dallas, TX, USA, 2 Yale University, New Haven, CT, USA, 3 Icahn School of Medicine at Mt Sinai, New York, NY, USA, 4 James J. Peters VA Medical Center, Bronx, NY, USA, 5 University of Washington, Seattle, WA, USA, 6 VA Greater Los Angeles Health Care System, Los Angeles, CA, USA, 7 VA Connecticut Healthcare System, West Haven, CT, USA Background: Beyond inhibition of cholesterol biosynthesis, statins appear to have pleiotropic effects, including modulation of cell growth, apoptosis, and inflammation. Statins may also reduce cancer risk, particularly among HIV- infected (HIV+) subjects who experience chronic inflammation and immune activation. Small observational studies have suggested an association between statin use and lower cancer risk in HIV+ but small sample sizes limited cancer type-specific analyses. Comparison of the association of statin exposure with cancer in HIV+ and HIV-uninfected comparators (HIV-) is also lacking. We used the Veterans Aging Cohort Study (VACS), a large observational cohort with cancer registry linkage and detailed pharmacy data to address these questions Methods: We followed statin users identified between 2000-2012, beginning 180 days after an index date defined as first statin prescription for users and a random visit date in the same year for non-users. To account for known and potential confounders we fit a propensity score (PS) model for statin use including age, calendar year, smoking, chronic diseases (e.g., diabetes, hypertension, HCV co-infection, alcohol use disorder), and laboratory values (e.g., LDL, albumin). We matched each statin user to up to four non-users by PS. We used Cox proportional hazards regression models to estimate hazard ratios (HRs) and 95% confidence intervals (CI) associated with statin use for all cancers, individual cancers, infection-related cancers (anal, colorectal, head and neck, liver, lymphoma, and stomach) and notinfection-related cancers. Results: The PS-matched sample included 48,214 participants, of whom 23,512 (48.8%) were incident statin users. Incident cancers were diagnosed in 940 (9.7%) of 9,649 HIV+ and 3,079 (8.0%) in 38,565 uninfected. Overall, statin use was associated with ~20% reduced risk of any cancer [HR 0.82 (95% CI 0.77 – 0.88)] and ~40% lower risk for infection-related cancers [HR 0.62 (95% CI 0.55 – 0.70)]. In general, the association was stronger in HIV+, but the interaction did not reach statistical significance except for non-Hodgkin lymphoma Conclusion: Statin exposure is associated with lower risk of cancer independent of HIV status. This protective effect appears to be stronger for infection-related cancers.

134 ACTG 5282: HPV TEST & TREAT VS CYTOLOGY-BASED CERVICAL CANCER PREVENTION IN HIV+WOMEN Timothy Wilkin 1 , Roy Matining 2 , Vikrant Sahasrabuddhe 3 , Huichao Chen 2 , Reena Masih 4 , Rosie Mngqibisa 5 , Mohammed Rassool 6 , Yamikani Mbilizi 7 , Tsitsi M. Magure 8 , Cynthia Riviere 9 , Aida Asmelash 10 , Ramesh Bhosale 11 , Pamela Michelow 12 , Catherine Godfrey 13 , Cindy Firnhaber 14 1 Weill Cornell Medicine, New York, NY, USA, 2 Harvard University, Boston, MA, USA, 3 National Cancer Institute, Frederick, MD, USA, 4 Social &Scientific Systems, Silver Spring, MD, USA, 5 Epicentre, Mbarara, Uganda, 6 Clinical HIV Research Unit,

47

CROI 2018