REDUCING COGNITIVE DECLINE WITH PROBUCOL IN ISCHEMIC STROKE PATIENTS WITH HIGH RISK OF CEREBRAL HEMORRHAGE: PICASSO-COG TRIAL

Abstract

Multiple cerebral microbleeds (CMBs) and prior intracerebral hemorrhage (ICH) are associated with higher risk of cognitive decline after stroke. We aimed to investigate the efficacy of probucol, a lipid-lowering and anti-oxidative agent, to prevent cognitive decline in stroke patients with multiple CMBs or prior ICH. PICASSO-Cog study is a predefined substudy of a randomized controlled clinical trial, PreventIon of CArdiovascular events in iSchemic Stroke patients with high risk of cerebral hemOrrhage (PICASSO). Patients with non-cardioembolic ischemic stroke or transient ischemic attack within 180 days and with previous ICH or multiple CMBs on gradient echo imaging were randomized to probucol versus no probucol groups from 61 institutes of South Korea. Mini-mental state examination (MMSE) and Montreal cognitive assessment (MoCA) was conducted at 4 (baseline), 13, 25, 37, and 49 months after index-stroke. Changes in MMSE and MoCA scores over time from baseline were analyzed using mixed effects model. A total 892 subjects were included in the analysis, with a median follow-up of 20.9 months. Mean age was 64.9 ± 10.8 years, median National Institute of Health Stroke Scale score 1 (IQR 0 - 3), and baseline MMSE score 25.0 ± 4.6. Mean changes of MMSE from baseline to each follow-up was 0.02±2.45 (1st, n=888), -0.15±2.65 (2nd, n=593), -0.24±3.11 (3rd, n=361), and -0.88±3.06 (4th, n=138). The MMSE scores over time showed a favorable trend for probucol treatment, but not significant. Among them, a total of 877 subjects underwent MoCA at least twice after randomization. Baseline MoCA score 19.3 ± 6.2, and mean changes of MoCA from baseline to each follow-up was 0.02 ± 2.83 (1st, n=871), -0.20 ± 3.17 (2nd, n=582), -0.09 ± 3.59 (3rd, n=354), and -0.52 ± 3.49 (4th, n=132). Probucol could significantly prevent cognitive decline in MoCA scores compared to placebo (p=0.01 for treatment effect). These effects were observed in the subgroups without diabetes mellitus, with concomitant lipid-lowering agent, baseline MMSE score > 24, and mild to moderate white matter hyperintensities.