CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

435 BRAIN WHITE MATTER HYPERINTENSITIES, HIV DISEASE, COGNITION AND DIABETES Minjie Wu 1 , Omalara Fatukasi 1 , Shaolin Yang 2 , Jeffry Alger 3 , Peter B. Barker 4 , Tae Kim 1 , Andrew Levine 3 , Eileen Martin 5 , Cynthia Munro 4 , Todd B. Parrish 6 , Ann B. Ragin 6 , Ned Sacktor 4 , Eric C. Seaberg 4 , James T. Becker 1 1 University of Pittsburgh, Pittsburgh, PA, USA, 2 University of Illinois at Chicago, Chicago, IL, USA, 3 University of California Los Angeles, Los Angeles, CA, USA, 4 Johns Hopkins Hospital, Baltimore, MD, USA, 5 Rush University Medical Center, Chicago, IL, USA, 6 Northwestern University, Chicago, IL, USA Background: HIV encephalitis includes the presence of white matter pallor and multinucleated giant cells at neuropathology. Since the onset of CART use, the incidence of HIV-Associated dementia and of encephalitis have fallen dramatically. The present study investigates the extent of white matter hyperintensities (WMHs) among individuals with HIV Disease, their impact on psychomotor speed, and factors that predict their presence (see left-hand panel of figure for example of peri-ventricular WMHs). As HIV Disease is associated with increased risk of cerebrovascular disease and stroke, we focused on factors such as diabetes and hypertension as risk modifiers. Methods: 322 men participating in the Multicenter AIDS Cohort Study (MACS) (185 HIV-infected, age: 57.5 (+6.0)) underwent MRI scans of the brain on Siemens 3T systems. T1-weighted MP-RAGE and Fluid Attenuated Inversion Recovery (FLAIR) images were obtained and processed using a semi-automated method for identifying and measuring WMHs. To segment WMHs, the algorithm uses the image intensity histogram to automatically select “seeds” of possible WMH and then iteratively clusters voxels based on their adjacency and affinity (i.e., fuzzy connectedness) to the seeds. WMH burden was expressed as the percentage of total white matter that was abnormal. Results: There were no statistically significant associations between WMHs and HIV Disease. However, in an adjusted model the extent of WMH was greater in men age > 60 (β=.21), of Non-Caucasian Race (β=.14), with a lower glomerular filtration rate (β= -.12) and with Diabetes (β=.14)(See right-hand panel of figure). There was no interaction between HIV status and age, although the interaction between age and diabetes was significant (β=.14). Race had a direct impact on WMH volume, and GFR was also predicted by Race (β=.15) and Hypertension (β= -.19). The extent of WMHs was significantly associated with performance on neuropsychological measures of psychomotor speed (β=.24), even after adjusting for age, education, and race. Conclusion: WMHs were a hallmark of HIV-Associated Dementia and Encephalitis prior to the use of CART. In today’s therapeutic environment, factors that affect the vasculature (e.g., hypertension and vascular disease) are the best predictors of WMHs and brain vascular disease. Early treatment and prevention could result in better brain health and better cognitive functions among infected individuals. Extra attention should be paid in African-Americans who have an increased risk of CVD.

Increased lesion loads were associated with decreased neurocognition in the HIV+ groups. Ventricular enlargement, a measure of brain atrophy, was associated with HIV infection within the two well-matched NIH cohorts and may highlight the role HIV plays in brain volume loss. Use of illicit drugs seems to compound effects of HIV in producing both WM lesions and brain atrophy. Ongoing analyses will further explore these trends.

434 ANALYSIS OF THE EFFECTS OF HIV DISEASE ON FUNCTIONAL NEUROMAGNETIC CONNECTIVITY

Mark D. Kelley 1 , Melissa F. Murray 1 , Donna Martineck 1 , Ricardo Bruna 2 , Anto Bagic 1 , James T. Becker 1 1 University of Pittsburgh, Pittsburgh, PA, USA, 2 Technical University of Madrid, Madrid, Spain Background: There is no generally accepted biomarker linking expression of cognitive impairment to HIV disease. Magnetoencephalography (MEG) is a noninvasive test that measures magnetic fields induced by synchronous activity of neuron groups. This study investigated the merits of MEG to identify alterations in functional connectivity (FC) apparent before onset of HIV- associated dementia. Methods: 54 individuals (31 HIV+) participated in this study. Subjects were 37-64 years old, right handed, and native English speakers; 36 were men, and 28 had more than 14 years of education. None had histories of ADD/ADHD, active drug abuse, current Axis I psychiatric disorder, or neurological disease. An Elekta NeuroMag scanner with 102 triplet sensors (2 gradiometers; 1 magnetometer) was used to examine all participants. Five minutes each of “eyes open” and “eyes closed” task-free data were collected and divided into approximately 100 three-second epochs per condition. The eyes open gradiometer MEG data underwent power and sensor-space connectivity analyses using HERMES software (hermes.ctb.upm.es). Brain structural integrity was analyzed using Voxel-Based Morphometry and SPM Anatomy. Phase-Locking Values (PLV) were calculated between all pairs of sensors (20,706) and quantitative volumetric data were measured for 84 anatomically defined regions. Results: Relative power in the left temporal lobe was lower in HIV+ participants (delta frequency) and higher in the left and right parietal ROIs (gamma frequency). The theta:gamma ratio in the right parietal region was higher in HIV+ participants. No measures of PLV were significantly associated with HIV infection (0.5 PLV threshold). Among infected persons only, longer duration of infection was associated with higher PLVs within bilateral temporal, right parietal, and right frontal ROIs. Strength of connection between the posterior cingulate cortex and both temporal ROIs (gamma frequency) was higher in seropositive individuals, and had a negative correlation with local gray matter volume. Conclusion: There is increasing evidence that MEG may be a useful imaging method to detect alterations in brain function prior to onset of clinical symptoms. Relative power decreases in HIV+ individuals is consistent across studies; we show here that changes in FC in a circuit linked to cognition is affected both by HIV infection and brain structural integrity suggesting the possibility of both an ongoing process due to infection, and a legacy effect from prior structural damage.

Poster Abstracts

436 MULTIVARIATE PATTERN ANALYSIS OF VOLUMETRIC NEUROIMAGING DATA IN TREATED HIV-DISEASE Jonathan Underwood , James H. Cole, Robert Leech, Alan Winston Imperial College London, London, UK Background: Accurate prediction of longitudinal changes in cognitive function in treated HIV-disease would potentially allow targeted intervention in those at greatest risk of cognitive decline. High resolution neuroimaging and machine learning techniques have shown promise in other disease areas but there are

CROI 2018 154