Genome-Wide Association Studies of 3 Distinct Recovery Phenotypes in Mild Ischemic Stroke.

Background and Objectives: Stroke genetic research has made substantial progress in the past decade. Its recovery application, however, remains behind, in part due to its reliance on the modified Rankin Scale (mRS) score as a measure of poststroke outcome. The mRS does not map well to biological processes because numerous psychosocial factors drive much of what the mRS captures. Second, the mRS contains multiple disparate biological events into a single measure further limiting its use for biological discovery. This led us to investigate the effect of distinct stroke recovery phenotypes on genetic variation associations with Genome-Wide Association Studies (GWASs) by repurposing the NIH Stroke Scale (NIHSS) and its subscores.
Methods: In the Vitamin Intervention for Stroke Prevention cohort, we estimated changes in cognition, motor, and global impairments over 2 years using specific measures. We included genotyped participants with a total NIHSS score greater than zero at randomization and excluded those with recurrent stroke during the trial. A GWAS linear mixed-effects model predicted score changes, with participant as a random effect, and included initial score, age, sex, treatment group, and the first 5 ancestry principal components.
Results: In total, 1,270 participants (64% male) were included with a median NIHSS score of 2 (interquartile range [IQR] 1-3) and median age 68 (IQR 59-75) years. At randomization, 20% had cognitive deficits (NIHSS Cog-4 score >0) and 70% had ≥1 motor deficits (impairment score >1). At 2 years, these percentages improved to 7.2% with cognitive deficits and 30% with motor deficits. GWAS identified novel suggestive gene-impairment associations ( p < 5e ^-6 ) for cognition ( CAMK2D , EVX2 , LINC0143 , PTPRM , SGMS1 , and SMAD2 ), motor ( ACBD6 , KDM4B , MARK4 , PTPRS , ROBO1 , and ROBO2 ), and global ( MSR1 and ROBO2 ) impairments.
Discussion: Defining domain-specific stroke recovery phenotypes and using longitudinal clinical trial designs can help detect novel genes associated with chronic recovery. These data support the use of granular endpoints to identify genetic associations related to stroke recovery.