CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

Results: Between ApoE ε4 carriers (n=26) and non-carriers (n=73), there were no significant differences in age, education, sex, race, HIV disease, HAND diagnosis, and most individual NP test scores, except for memory performance measured by the Hopkins Verbal Learning Test–Revised (HVLT-R). Carriers had significantly lower delayed recall and retention score than non-carriers (p<0.05), but there was no interaction between age and ApoE ε4 status. For MRI, there was no difference in gray matter volume or cortical thickness between carriers and non-carriers. In contrast, the FC between the right caudate and right hippocampus was significantly lower in carriers (p=0.0002) and correlated with HVLT-R retention (p=0.015), along with a significant interaction between ApoE ε4 genotype and CD4 nadir (p=0.026). A similar marginal, but non-significant, effect was found in the FC between the left caudate and left hippocampus. Conclusion: In this sample of PWH (41-70 years old), ApoE ε4 was associated with reduced verbal memory performance and disrupted FC between the caudate and the hippocampus, suggesting that ApoE ε4 may be a genetic risk factor for memory impairment in PWH. In addition, the interaction between ApoE ε4 allele and CD4 nadir on FC suggests that the severity of HIV disease may exacerbate the effect of ApoE ε4 on brain health, resulting in an increased risk of dementia and Alzheimer's disease later in life. 388 NEUROCOGNITIVE AND VOLUMETRIC CHANGES AFTER 24 WEEKS OF DTG/3TC/ABC DISCONTINUATION Ignacio Pérez-Valero 1 , Alfonso Cabello 2 , Pablo Ryan 3 , Maria Luisa Montes 1 , Sara De La Fuente Moral 4 , María J. Vivancos-Gallego 5 , Guillermo Cuevas 3 , Alberto Díaz De Santiago 4 , Lucio Jesus Garcia-Fraile Fraile 6 , Mario Gil-Correa 7 , Norberto Malpica 8 , Guadalupe Rua 1 , María Yllescas 9 , Alicia Gonzalez 1 1 Hospital La Paz Institute for Health Research, Madrid, Spain, 2 Fundacion Jimenez Diaz, Madrid, Spain, 3 Hospital Universitario Infanta Leonor, Madrid, Spain, 4 Puerta de Hierro Research Institute and University Hospital, Madrid, Spain, 5 Instituto Ramón y Cajal de Investigación Sanitaria, Madrid, Spain, 6 Hospital Universitario de La Princesa, Madrid, Spain, 7 Universidad Rey Juan Carlos, Madrid, Spain, 8 Universidad Rey Juan Carlos Alcorcón, Alcorcón, Spain, 9 Fundación SEIMC-GeSIDA, Madrid, Spain Background: Dolutegravir/lamivudine/abacavir (DTG/3TC/ABC) fixed dose combination (FDC) discontinuation is associated with improvement of neuropsychiatric (NP) symptoms. However, limited data exist on the effect of DTG/3TC/ABC discontinuation on neurocognitive (NC) function and brain structure. We assessed NC function and neuroimaging in participants enrolled in the DREAM study, a multicenter clinical trial designed to evaluate the reversibility of NP symptoms in virologically suppressed patients on DTG/3TC/ ABC switching to Cobicistat-boosted elvitegravir/emtricitabine/tenofovir alafenamide FDC (E/C/F/TAF). Clinical results from this trial have demonstrated significant improvements in NP symptoms when switching from DTG/3TC/ABC to E/C/F/TAF. Methods: All participants performed a comprehensive NC assessment (7-domain) following Frascati criteria and a 3-Tesla brain MRI on day 1 and 24 weeks after switching therapy. Global NC performance was assessed using the global deficit score (GDS). Changes in NC function and brain volumes were determined using neuromorphometrics atlas and analyzed using t-test. Multiple comparisons testing was corrected using the false discovery rate (FDR) adjustment. Results: 38 participants, mostly Caucasian male of middle age with good immunological status, normal NC function that received DTG/ABC/3TC for a mean time of 1.45 years, were included. At week 24 after switching to E/CF/TAF, we observed significant improvements in the global NC function (mean±SD GDS change: 0.12±0.32; p=0.029) and in the speed of processing (-0.26±0.86; p<0.001), delayed recall (-0.33±0.8; p=0.019) and motor (-0.51±0.85; p=0.001) domains. Significant changes in several brain volumes were observed (table). After FDR adjustment, only the changes in the right frontal pole, a cerebral region involved in information processing, emotion and motivated behaviors, remained significant (p<0.03). We also observed a significant correlation between GDS changes and volume changes in the right superior occipital gyrus (r=0.53) Conclusion: Our study suggests that switching from DTG/3TC/ABC to E/C/F/TAF was associated with an improvement in NC functioning, especially in speed of processing, delayed recall and motor domains. Brain volumes changes observed in our study could be useful to delve into the pathological mechanisms of DTG/3TC/ABC-related NP toxicity.

389 HIV IS INDEPENDENTLY ASSOCIATED WITH BRAIN MRI WHITE MATTER HYPERINTENSITIES Yair Mina 1 , Tianxia Wu 1 , Hsieh Hsing-Chuan 2 , Brian K. Agan 2 , Dima A. Hammoud 3 , Daniel S. Reich 1 , Govind Nair 1 , Lillian Ham 3 , Joseph Snow 3 , Elizabeth F. Horne 1 , Avindra Nath 1 , Bryan Smith 1 , for the NIH-DOD NeuroHIV Consortium 1 National Institute of Neurological Disorders and Stroke, Bethesda, MD, USA, 2 Uniformed Services University of the Health Sciences, Bethesda, MD, USA, 3 NIH, Bethesda, MD, USA Background: Brain white matter hyperintensities (WMH) are nonspecific in etiology but occur commonly in the setting of HIV infection. Their relevance to disease remains uncertain as previous studies have used various methods of estimation in heterogenous groups of people with HIV (PWH). We evaluated the frequency, severity and clinical correlations of WMH in the setting of well controlled HIV infection, using the Fazekas visual rating scale. Methods: Research protocol images from brain 3D fluid attenuated inversion recovery (FLAIR) on a 3T MRI were reviewed to determine the Fazekas score (total score of 0-6, 0-3 for periventricular hyperintensity and 0-3 for deep WMH) in people with HIV (PWH) with well controlled infection (antiretroviral therapy for at least one year and plasma viral load <200 copies/ml) and in controls. Simple linear regression was used for covariate selection, and forward stepwise regression was performed to evaluate the effect of HIV on Fazekas scores. Results: Fazekas scores were determined in 203 PWH and 58 controls. The PWH group had a higher mean total Fazekas score compared to controls (2.2±1 vs. 1.8±1.3, p=0.008). Age, history of cocaine abuse, prior smoking history, acetylsalicylic acid use, hepatitis C virus antibodies, higher hemoglobin A1C level (A1C), and lower estimated glomerular filtration rate were all associated with a higher Fazekas score in PWH and/or control group. The final forward stepwise regression model included age and history of cocaine abuse (n=200, R2=0.24), or age and A1C (n=242, R2=0.27). In both models, the adjusted mean total Fazekas score remained higher in the PWH group compared to the control group [2.3 (95%CI 2.1-2.5) vs. 2.0 (95%CI 1.7-2.3) with p=0.046 adjusted for age and cocaine abuse, 2.2 (95%CI 2.0-2.3) vs. 1.8 (95%CI 1.5-2.0) with p=0.0039 adjusted for age and A1C]. No specific cognitive measure correlated with the Fazekas score. In the PWH group, CD4 nadir <200 cells/ml was associated with a higher Fazekas score [age-adjusted mean 2.4 (95%CI 2.2-2.6) vs. 2.0 (95%CI 1.9-2.2), p=0.01]. Conclusion: HIV infection contributes to the extent of brain WMH, even in the setting of well controlled infection. Prior immunosuppression evidenced by lower nadir CD4 partially explains this association. The cohort is undergoing serial MRI scans and neuropsychological testing to evaluate the long-term clinical effects of WMH. 390 BRAIN AGE BASED ON SLEEP ENCEPHALOGRAPHY IS ELEVATED IN HIV+ ADULTS ON ART Michael Leone 1 , Haoqi Sun 1 , Christine Boutros 1 , Robert Thomas 2 , Gregory K. Robbins 1 , Shibani S. Mukerji 1 , M. Brandon Westover 1 1 Massachusetts General Hospital, Boston, MA, USA, 2 Beth Israel Deaconess Medical Center, Boston, MA, USA Background: Co-morbidities and increased inflammation associated with HIV have raised concern for excess brain aging, yet diagnostic biomarkers for brain aging are lacking. Our lab developed a machine learning model that estimates age from sleep EEG (brain age, BA), which reliably predicts chronological age

Poster Abstracts

CROI 2020 134