CROI 2019 Abstract eBook

Abstract eBook

Oral Abstracts

infection. To study whether treatment with TAF is associated with an increased risk of neuronal harm compared to TDF, we compared plasma NFL levels in patients switching from coformulated elvitegravir/cobicistat/emtricitabine/ tenofovir disoproxil fumarate (E/C/F/TDF) to E/C/F/TAF with those who continued E/C/F/TDF. Methods: Plasma NFL was analysed at baseline, week 24, and week 84, in stored plasma samples from 414 participants (272 switching to E/C/F/TAF and 142 continuing E/C/F/TDF) enrolled in the randomized, active-controlled, multicenter, open-label, noninferiority Gilead GS-109 trial. For quality control (QC) plasma samples with NFL concentrations of 12.1 pg/mL and 188 pg/mL, intra-assay coefficients of variation (CVs) were 7.8% and 6.7%, respectively. Results: We found a small but statistical significant decrease in plasma NFL in the E/C/F/TAF arm after 84 weeks from 10.3 to 9.6 pg/mL, p<0.01 (Figure). The change was significantly different (p<0.01) from the E/C/F/TDF arm, in which plasma NFL increased from 11.1 to 11.7 pg/mL (ns). As expected, eGFR increased in the E/C/F/TAF arm but not in the E/C/F/TDF arm. Plasma NFL was significantly correlated with age and eGFR. Delta eGFR and treatment group were both found as independent predictors of plasma NFL changes from baseline to week 84 in a multiple linear regression analysis. Conclusion: We found no evidence of increased risk of CNS injury when switching from TDF to TAF. It should be noted that the NFL levels in both arms were within the limits normally found in HIV-negative controls; it is unclear whether the small decrease in plasma NFL found after switch to TAF is of any clinical significance. This study indicates that switching from TDF to TAF appears safe with regard to neuronal injury.

summary z-score (NPZ-4). Participants undergo repeat neurocognitive testing every 48 weeks. The 10-year ASCVD risk score and FRS were calculated at entry. We first assessed how well the baseline ASCVD risk score and FRS predicted NPZ-4 at Year 4 in unadjusted linear regression models. We then performed stepwise linear regression (Table) to determine the covariate-adjusted association between baseline 10-year CV risk and NPZ-4 at Year 4. Results: Of 988 participants, mean age was 52 years, 20%were women, and 90% had an undetectable viral load. Mean ASCVD risk score and FRS were 6.8% and 13.1%, respectively. Both risk scores were lower in women than men (ASCVD 4.1% vs. 7.5%, p<0.001; FRS 8.1% vs. 14.3%, p<0.001). For every 1% higher baseline ASCVD risk, NPZ-4 at Year 4 was lower by 1.4 SD (p=0.003). Baseline ASCVD risk predicted NPZ-4 at Year 4 overall and in both women and men (Table). In adjusted models, for every 1% higher baseline ASCVD risk, NPZ-4 at Year 4 was 1.1 SD lower, though this did not reach statistical significance (p=0.085). Baseline ASCVD risk significantly predicted NPZ-4 at Year 4 for women (-3.1 SD per 1% higher risk, p=0.010) but not for men (-0.4 SD per 1% higher risk, p=0.55), even after adjustment for NPZ-4 at entry. The associations between baseline FRS and NPZ-4 were comparable, although higher ASCVD risk had a greater effect on NPZ-4 than higher FRS (Table). Conclusion: Higher baseline 10-year CV risk predicted worse cognitive function at Year 4 in PLHIV, though this association was attenuated in men after adjusting for covariates. A higher CV risk score may help to identify PLHIV who are at risk for worse cognitive function over time.

Oral Abstracts

129 OBESITY IS INDEPENDENTLY ASSOCIATED WITH NEUROCOGNITIVE DECLINE IN HIV Jeremiah Perez 1 , Adriana Andrade 2 , Ronald J. Ellis 3 , Mary Clare Masters 4 , Karl Goodkin 5 , Susan L. Koletar 6 , Frank J. Palella 4 , Kevin Robertson 7 , Ned Sacktor 8 , Katherine Tassiopoulos 1 , Kristine M. Erlandson 9 1 Harvard University, Boston, MA, USA, 2 DAIDS, NIAID, Rockville, MD, USA, 3 University of California San Diego, San Diego, CA, USA, 4 Northwestern University, Chicago, IL, USA, 5 East Tennessee State University, Johnson City, TN, USA, 6 The Ohio State University, Columbus, OH, USA, 7 University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, 8 Johns Hopkins Hospital, Baltimore, MD, USA, 9 University of Colorado, Aurora, CO, USA Background: Neurocognition may decline more with age among people living with HIV (PLWH) compared to uninfected persons. The factors related to this decline are not well understood in the current antiretroviral therapy (ART) era. Methods: AIDS Clinical Trials Group (ACTG) A5322 (HAILO) is an observational cohort study of PLWH ≥ 40 years old, on ART. Participants undergo annual assessments for neurocognitive impairment (NCI), with NCI defined by ≥1 z-score ≥2 SD below 0 or ≥2 z-scores ≥1 SD below 0 on Trailmaking A and B and the Wechsler Adult Intelligence Scale-Revised Digit Symbol tests. Obesity was defined as body mass index (BMI) >30 kg/m², overweight as 25-30 kg/m², normal weight as 18.5-25 kg/m², and underweight < 18.5 kg/m². Participants who developed NCI during the first 3 years were compared to persons who maintained normal neurocognition. We used logistic regression to assess the age-adjusted associations between NCI and baseline covariates including sex, race, alcohol use, BMI, waist circumference, nadir CD4, history of AIDS defining illness, hemoglobin A1C. Only covariates with a p-value < 0.1 from age-adjusted analysis were included in the multivariable models. Results: Of 929 participants, 81%were male, 31% Black, and 20% Hispanic. Median age was 51 years (IQR 46-56). Most individuals (92%) had undetectable

128 CARDIOVASCULAR RISK SCORES PREDICT LONGITUDINAL COGNITIVE FUNCTION IN OLDER PLHIV Felicia C. Chow 1 , Asya Lyass 2 , Joseph Massaro 2 , Virginia Triant 3 , Kunling Wu 4 , Baiba Berzins 5 , Kevin Robertson 6 , Ronald J. Ellis 7 , Katherine Tassiopoulos 4 , Babafemi Taiwo 5 , Ralph D’Agostino 2 1 University of California San Francisco, San Francisco, CA, USA, 2 Boston University, Boston, MA, USA, 3 Massachusetts General Hospital, Boston, MA, USA, 4 Harvard University, Boston, MA, USA, 5 Northwestern University, Chicago, IL, USA, 6 University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, 7 University of California San Diego, San Diego, CA, USA Background: Cardiovascular (CV) disease (CVD) and associated risk factors have been linked with neurocognitive impairment (NCI) in cross-sectional studies of persons living with HIV (PLHIV), although the specific CV risk factors that correlate with NCI have varied. We examined the utility of two commonly used 10-year CV risk scores--the Atherosclerotic CVD (ASCVD) and Framingham Heart Study Global CVD risk score (FRS), which combine multiple CV risk factors--to predict longitudinal cognitive function in an observational cohort of older PLHIV. Methods: Participants from the ongoing AIDS Clinical Trials Group A5322 study who underwent neurocognitive testing (Trailmaking A and B, Hopkins Verbal Learning Test-Revised, Digit Symbol) at entry were eligible. Raw scores are standardized using demographics-adjusted norms and combined into a

48

CROI 2019