CROI 2018 Abstract eBook

Abstract eBook

Oral Abstracts

125 IDENTIFICATION OF BRAIN INJURY USING A NOVEL DEFINITION OF COGNITIVE IMPAIRMENT Jonathan Underwood 1 , James H. Cole 1 , Davide De Francesco 2 , Caroline Sabin 2 , Alan Winston 1 1 Imperial College London, London, UK, 2 University College London, London, UK Background: Despite the reportedly high prevalence of HIV-associated cognitive impairment in the antiretroviral era, the relationships with neuroimaging measures of brain injury have been inconsistent. This lack of consistency may be explained by recent modelling data which has suggested that the two common methods of defining impairment, the HAND or ‘Frascati’ criteria and the global deficit score (GDS), have specificities below 80%. Here, we tested the hypothesis that a novel, multivariate method (NMM) of defining cognitive impairment would be more reliably associated with brain injury. Methods: Cognitive function was tested across seven domains and T1- weighted MRI data were obtained from virally suppressed people living with HIV (PLWH) from the CHARTER cohort (n=139, mean [SD] age 45.2 [7.45], 79.1% male, 46.0%white ethnicity). Cognitive impairment was defined using the Frascati criteria and GDS as well as a NMM based on the Mahalanobis distance, with a priori specificity of 85% based on modelling data. Associations with brain volumes were assessed using multiple linear regression adjusting for age, intracranial volume, scanner and comorbidity status. In addition, a multivariate, machine learning model of healthy brain ageing using volumetric neuroimaging data (training dataset n=2001, age range 18-90 years, model R²=0.88) was used to quantify the difference between apparent and chronological brain age and the associations with different definitions of cognitive impairment were determined. Results: The prevalence of cognitive impairment was 53.2% using the Frascati criteria, 37.4% using the GDS and 20.9% using the NMM, despite similar median global T-scores between the groups (table). There were no associations between Frascati or GDS defined cognitive impairment and grey matter volume (p>0.5 for both). However, PLWH with cognitive impairment defined using NMM had 21.0 (3.51-69.2) mL less grey matter than PLWH without impairment (p<0.01). White matter volume did not differ between PLWH with and without cognitive impairment, regardless of the definition used (p>0.2 for all). Compared to PLWH without NMM defined cognitive impairment, PLWH with NMM impairment had older appearing brains (3.43 [0.17-6.68] years, p=0.04, others p>0.4). Conclusion: The prevalence of cognitive impairment in virally suppressed PLWH varies depending on the definition used. The NMMmethod, where specificity can be set a priori, was associated with brain injury more consistently than the Frascati and GDS methods.

we analyzed a common HO-1 promoter region (GT)n dinucleotide repeat polymorphism implicated in regulating HO-1 promoter transcriptional activity (shorter repeats associate with greater transcription), in an expanded version of this HIV autopsy cohort. Methods: HO-1 (GT)n polymorphism allele lengths were determined by PCR and capillary electrophoresis, and brain prefrontal cortex RNA expression was analyzed by qPCR from an autopsy cohort of HIV-, HIV+, and HIV encephalitis (HIVE) subjects (n=554). Statistical analyses were performed by chi-square, Kruskal-Wallis, and multivariate linear regression. Results: The HO-1 (GT)n polymorphism allele repeat lengths ranged from 13 to 44 repeats with a trimodal distribution of peaks at 23, 30, and 39, as in previous reports. Based on this allele repeat length distribution, we assigned genotypes as short (< 27), medium (27-34) and long (> 34) HO-1 alleles. HIV+ subjects with short alleles had a significantly lower risk of HIVE (p=0.04, odds ratio=0.62). In HIV+ subjects without HIVE, the presence of a short allele correlated significantly with lower brain type I interferon responses (ISG15 p<0.001, MX1 p<0.001, IRF1 p=0.007) and T-lymphocyte activation markers (CD38 p=0.008, GZMB p=0.01, CD8A p=0.07). No correlations were found with macrophage markers (CD163, CD68), endothelial markers (PECAM, VWF), or the B-lymphocyte maker CD19. Also the presence of a short allele did not correlate with plasma or CNS viral loads or CD4 T-cell counts or nadirs. Conclusion: Our data suggest unique modifying risk effects for HIV- induced CNS neuroinflammation and associated neuropathogenesis that are driven by an individual’s HO-1 promoter (GT)n polymorphism allele repeat length. The presence of shorter HO-1 alleles might provide neuroprotection through decreased neuroimmune activation and neuroinflammation as a result of increased HO-1 promoter activity. Therapeutic strategies that induce HO-1 expression may further decrease HIV-associated CNS neuroinflammation and decrease the risk for development of HIV neurological disease. 127 RETREATMENT OF HEPATITIS C INFECTION IN PATIENTS WHO FAILED GLECAPREVIR/PIBRENTASVIR David L. Wyles 1 , Ola Weiland 2 , Betty Yao 3 , Robert Reindollar 4 , Frank Weilert 5 , Jean-Francois Dufour 6 , Stuart Gordon 7 , Fred Poordad 8 , Albrecht Stoehr 9 , Ashley Brown 10 , Stefan Mauss 11 , Suvajit Samanta 3 , Tami Pilot-Matias 3 , Lino Rodrigues 3 , Roger Trinh 3 1 Denver Health Medical Center, Denver, CO, USA, 2 Karolinska University Hospital, Stockholm, Sweden, 3 AbbVie, Inc, North Chicago, IL, USA, 4 Piedmont Healthcare, Statesville, NC, USA, 5 Waikato Hospital, Hamilton, New Zealand, 6 University Hospital of Bern, Bern, Switzerland, 7 Henry Ford Hospital, Detroit, MI, USA, 8 University of Texas at San Antonio, San Antonio, TX, USA, 9 Institute for Interdisciplinary Medicine, Hamburg, Germany, 10 Imperial College London, London, UK, 11 Center for HIV and Hepatogastroenterology, Düsseldorf, Germany Background: Glecaprevir*/pibrentasvir (G/P; *identified by AbbVie and Enanta) is a next-generation Hepatitis C virus (HCV) treatment regimen that has demonstrated high sustained virologic response (SVR) regardless of HCV genotype (GT) or baseline patient or viral characteristics. Approximately 1% of patients treated in the G/P clinical trial program to date had virologic failure (VF) and no data have been presented on their outcomes following retreatment. These patients were enrolled into a retreatment study, MAGELLAN-3 (NCT02939989). Methods: MAGELLAN-3 is an ongoing phase 3b, open-label trial, in which patients who had VF following G/P were retreated with the combination of G/P 300 mg/120 mg once daily (QD) + sofosbuvir (SOF) 400 mg QD + ribavirin (RBV) 1,000–1,200 mg (weight based, twice daily). Patients who were non-GT3- infected, non-cirrhotic, and naïve to protease inhibitor and/or NS5A inhibitor prior to VF with the G/P regimen received 12-week (Arm A) treatment with the combination regimen; all other enrolled patients who did not meet any of these criteria received the same regimen for 16 weeks (Arm B). Efficacy (primary outcome is SVR at post-treatment (PT) Week 12 [SVR12]), safety, and baseline resistance were assessed. Preliminary SVR at PT Week 4 (SVR4) results, safety, and baseline resistance are reported here. Results: As of 15 September 2017, 24 patients were enrolled (3 in Arm A; 21 in Arm B). Baseline characteristics are presented in the table. To date, 12 of 13 patients who completed PT Week 4 achieved SVR4. One patient in Arm B who had a GT1 infection and prior treatment experience with protease inhibitor and/or NS5A inhibitor before failing the G/P regiment experienced relapse at PT Week 4. Adverse events (AEs) reported in ≥10% of patients overall were headache (25.0%), pruritus (25.0%), dizziness (16.7%), and irritability

Oral Abstracts

126 HEME OXYGENASE-1 POLYMORPHISM ASSOCIATES WITH NEUROIMMUNE ACTIVATION IN HIV SUBJECTS Alexander J. Gill 1 , Rolando Garza 1 , Suren A. Ambegaokar 2 , Patricia J. Vance 1 , Benjamin B. Gelman 3 , Dennis L. Kolson 1 1 University of Pennsylvania, Philadelphia, PA, USA, 2 Ohio Wesleyan University, Delaware, OH, USA, 3 University of Texas at Galveston, Galveston, TX, USA Background: Heme oxygenase-1 (HO-1) is a critical cytoprotective enzyme that limits oxidative stress, inflammation, and cellular injury within the central nervous system (CNS) and other tissues. We previously demonstrated that HO-1 protein expression is decreased within the brains of HIV+ subjects and that this HO-1 reduction correlates with CNS immune activation and neurocognitive dysfunction. To define a potential CNS protective role for HO-1 against HIV,

44

CROI 2018