CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

not been well studied for other preventive medications, including other lipid- lowering (LL) drugs. Methods: We compared ongoing and past use of statins and other preventive drug classes for their association with death, cancer, severe infection (excluding bronchitis, cellulitis, and urinary infections), and cardio- or cerebrovascular (ASCVD) events identified by ICD-9 code. We included all HIV infected US Veterans from 1995-2011 after their 1st undetectable HIV viral load on HAART and used Cox models with inverse probability weighting (IPW) for treatment and censoring. We built time-updated drug exposure models from pharmacy outpatient refill and inpatient prescription data and categorized drug exposure on a weekly basis. We defined ≥75% drug use in the past month as current use, ≥98% use in the past year as consistent use, and last drug exposure >1 year ago as remote use. We calculated propensity scores (PS) for each exposure category using multivariable generalized linear models of main effects and 2-way interactions of the following parameters: calendar year, VA station size, follow- up frequency, demographics, comorbidities including smoking and drug use, HAART-type and -adherence rate, degree of virologic suppression, CD4 counts, liver and kidney function, hemoglobin, body mass index, systolic blood pressure as well as total and HDL cholesterol. Results: We followed 23,267 patients for a median of 5.2 years (IQR: 2.5-9.2). Median age at inclusion was 53 years (IQR 46-60). 97% of patients were male, 46% black, 37%white, and 56% ever smoked. 36% had an exposure to statins, but only 16% of follow-up years were classified as ongoing statin exposure. Hazard ratios with 95% confidence intervals for death and NADC are shown in the table. Conclusion: We show a protective effect of statins and other LL drugs on death and NADC which had not been described for other LL drugs. The statin mortality benefit may be reflective of the reduced rates for cancer and infections and was seen despite their positive association with ASCVD events that remained after IPW. This association was weaker for other LL drugs, possibly explaining their greater mortality benefit. Of note, use of cardiac aspirin was not only associated with an increased risk of death but also cancer and infection.

body mass index (rho=-0.33, p=0.002) and HDL-C (rho=-0.41, p<0.001), whereas no association was found with LDL-C and hsCRP. Initiation of ART was associated with a significant increase in TC, LDL-C, HDL-C and lipoprotein(a) levels and a significant decrease in PCSK9 and hsCRP levels. These changes were consistent for different ART regimens. TC and HDL-C but not LDL-C variations were associated with PCSK9 variation (Table 1). Conclusion: Baseline PCSK9 levels are related to immuno-virological parameters but appear uncoupled from LDL-C levels. A complex lipid profile perturbation, including also a PCSK9 reduction, follows ART initiation

Poster Abstracts

651 LIPID CHANGES ASSOCIATED WITH TAF ARE REVERSIBLE BY SWITCHING BACK TO TDF Ana Milinkovic 1 , Florian Berger 2 , Alejandro Arenas-Pinto 3 , Stefan Mauss 2 1 Chelsea and Westminster NHS Foundation Trust, London, UK, 2 Center for HIV and Hepatogastroenterology, Düsseldorf, Germany, 3 University College London, London, UK Background: Switching from tenofovir disoproxil fumarate (TDF) to tenofovir alafenamide (TAF) has shown worsening of lipid profile in people living with HIV (PLWH), but there is little data exploring changes in lipid profile in PLWH switching back from TAF to TDF. Methods: This analysis consists of a retrospective data collection on effectively suppressed HIV-positive patients who were initially switched from TDF to TAF-based antiretroviral treatment (ARVT) due to medical or economic reasons or as a result of optimization of therapy in a single site(Center for HIV and Hepatogastroenterology). After generics of TDF were introduced a substantial proportion of patients were switched back from TAF to TDF. This analysis includes patients switched back from TAF to TDF. All components of ARVT for all patients analysed were maintained the same with the single initial substitution of TDF to TAF, and subsequent substitution of TAF to TDF. Only patients on stable lipid lowering therapy were included. We previously reported increase in total and LDL-cholesterol in patients switched from TDF to TAF. This analysis includes patients switched back to TDF after at least of 60 weeks of exposure to TAF in regular clinical care. Lipid profile was measured at 12 weeks intervals. Results: 385 virologically suppressed PLWH were initially included. Duration of TDF exposure before switching to TAF was 350(SD±201) weeks. 72 were switched back from TAF to TDF after mean duration of 87 weeks (SD±22) on TAF. Median age of 50 (SD ±12) years, 88%were male, 93% Caucasian, with a median BMI of 23.85 sqm/kg (SD ±3.9), all patients had well controlled HIV, with mean CD4 cell count of 714/µl (SD±272). After the initial switch from TDF to TAF after 12 and 24 weeks total cholesterol (TC) had increased by + 17 (SD±24) and + 14 (SD±27)mg/dl (p<0.001) Mean triglycerides had increased by + 39 (SD±94) and + 25 (SD±94) mg/dl (p<0.001). At switching, median TC was 187 mg/dl (SD ±33). Switching back from TAF to TDF led to TC decrease at week 12 by -24 mg/dl (SD ±23) (p<0.001). TC improved in 35% of patients after switching back to TDF. Triglycerides improved by -23 mg/dl (SD ±101) (p=0.009). Conclusion: The results of our study confirm a reversible, pharmacological effect on lipid profile of a switch from TDF to TAF and back. This effect is not a universal phenomenon, but observed in about a third of the cohort. Clinical relevance of these findings requires further investigation in particular identifying patients at risk. 652 CHANGES IN LIPIDS AFTER A DIRECT SWITCH FROM TDF TO TAF Patrick W. Mallon 1 , Laurence Brunet 2 , Jennifer S. Fusco 2 , Girish Prajapati 3 , Andrew Beyer 3 , Gregory Fusco 2 , Michael Wohlfeiler 4

650 PCSK9 AND HIV INFECTION: CORRELATION WITH DYSLIPIDEMIA, INFLAMMATION, AND HAART Elisabetta Schiaroli , Salvatore Cardaci, Vanessa Bianconi, Massimo R. Mannarino, Matteo Pirro, Daniela Francisci, Franco Baldelli University of Perugia, Perugia, Italy

Background: Lipid profile is generally deranged in antiretroviral (ART)-naïve HIV+ patients due to HIV infection severity and immunodeficiency state. Proprotein convertase subtilisin/kexin type 9 (PCSK9), a major regulator of cholesterol metabolism, is induced in some inflammatory states with a trend toward an increase in plasma levels in HIV untreated/treated patients compared to healthy controls. Whether plasma PCSK9 levels may decrease after ART initiation is not established. We measured plasma lipid and PCSK9 levels in ART-naïve HIV+ patients and investigated the impact of ART initiation on these parameters Methods: This is a longitudinal study of 82 HIV+ ART-naïve patients not receiving any lipid-lowering treatment. At baseline and after three and six months of ART plasma total cholesterol (TC), low density lipoprotein (LDL)-C, high density lipoprotein (HDL)-C, triglyceride, lipoprotein(a), PCSK9 and high- sensitivity C-reactive protein (hsCRP) levels were evaluated Results: At baseline plasma PCSK9 levels were significantly associated with CD4 T cell count (rho=-0.52, p=0.001), HIV-1 RNA viral load (rho=0.44, p<0.001),

CROI 2019 247