CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

model improved discrimination for CKD (c-statistic=0.72 to 0.76, p=0.0029), but only modestly for mortality (c=0.79 to 0.80, p=0.099). Clusters derived with all 8 markers were no better for discrimination than the 3-biomarker clusters. Conclusions: For predicting incident CKD in HIV-infected women, clusters developed from three urine-based kidney disease biomarkers were as effective as a panel of eight biomarkers in improving CKD risk discrimination.

692 Differences in Urine Metabolomes With HIV Infection and Antiretroviral Drug Exposure Andrew J. Chetwynd 1 ; Amanda Samarawickrama 2 ; Alaa Abdul-Sada 1 ; Jaime H. Vera 2 ; Stephen G. Holt 3 ; Martin Fisher 2 ;Yvonne Gilleece 2 ; Elizabeth M. Hill 1 1 Univ of Sussex, Brighton, UK; 2 Brighton and Sussex Hosps NHS Trust, Brighton, UK; 3 Univ of Melbourne, Melbourne, Australia

Background: Metabolomics is the analysis of small molecular weight compounds (<1000 Daltons) in biological samples. Fewmetabolomic studies have analysed the effect of combination antiretroviral therapy (cART) using urine in HIV-infected patients. We investigated the effect of cART on the urinary metabolome of HIV-infected patients. Methods: Fasted urine samples from randomly chosen HIV-infected patients and negative controls were analysed by non-targeted metabolomics using liquid chromatography mass spectrometry (LC-MS). Participants were grouped into 3 groups: HIV-infected on cART, cART-naïve and HIV-negative. Principal components analysis (PCA) was used to plot metabolomic differences associated with HIV and cART status. A supervised modelling approach identified metabolites associated with cART. Analysis of variance was used to compare differences in metabolomes between groups. Results: 89 participants (cART-naïve: 26; on cART: 50; negative: 13) were included (mean age 42.1 [SD 9.9] years, 100%male, 97.8%white). All those on cART (median CD4 525 [IQR 403,690] cells/µL, 100% VL <40 copies/mL) were on tenofovir/emtrictabine, ritonavir and atazanavir (ATZ, n=20) or darunavir (DRV, n=30). In cART-naïve participants, median CD4 was 490 (IQR 395,695) cells/µL. Serum renal and liver profiles were similar between all groups, except for higher bilirubin in those on ATZ. The PCA plot showed no difference in metabolomes between HIV-negative and HIV-infected groups, but discriminated between those on cART (ATZ versus DRV) and naïve groups (Figure). Bile acids (cholic acid glucuronide, cholic acid, glycocholic acid and norcholestanehexol glucuronide), which can be markers of liver toxicity, were significantly reduced in those on cART compared to cART-naïve (p<0.05) indicating potential disruption of bile acid transporter mechanisms in the kidney/liver. The nucleoside 5-deoxy-5(methylthio)adenosine, involved in cell apoptosis, was significantly reduced in both cART groups (p<0.05), suggesting that cART may be effective in reducing HIV-induced CD4 cell apoptosis. Conclusions: This to our knowledge is the first urinary trace metabolomic study in HIV infection. There were significant differences between the urine metabolome of those on protease inhibitor-based cART compared with cART-naïve and negative groups. Changes in urine metabolomes associated with cART may be useful in monitoring toxicity or adherence. Further work is needed to assess the clinical utility of metabolomics in HIV-infected patients.

Poster Abstracts

693 Early Markers of Renal Dysfunction Among Cocaine Users With HIV and HCV Infection Edana Cassol 1 ;Vikas Misra 2 ; Shibani S. Mukerji 3 ; Susan Morgello 4 ; Gregory D. Kirk 5 ; Shruti H. Mehta 5 ; Dana Gabuzda 2 1 Carleton Univ, Ottawa, ON, Canada; 2 Dana-Farber Cancer Inst, Boston, MA, USA; 3 Massachusetts General Hosp, Boston, MA, USA; 4 Mount Sinai Sch of Med, New York, NY, USA; 5 Johns Hopkins Univ, Baltimore, MD, USA Background: Cocaine use is associated with increased kidney and cardiovascular disease in the general population, but its effects on these comorbidities in people with HIV and HCV infection are not well characterized. Methods: Case-control study to examine effects of cocaine use on plasma metabolite profiles in relation to comorbidities in 104 subjects in 3 cohorts; cocaine users (mainly crack cocaine) matched to controls for demographics, HIV serostatus, ART use, CD4 counts, plasma HIV viral load (VL) (n=38 HIV-negative from ALIVE and Bioreclamation, n=36 HIV+ IVDU from ALIVE [77% on ART], n=30 HIV+ from NNTC and CHARTER [100% on ART]; 52% cocaine users, 60% HCV+ (ALIVE subjects were HCV aviremic), 76%with VL <400 cps/ml.

287

CROI 2016