Abstract WP144: Variables Associated With Recovery of Communication After Stroke

Abstract

Background: Communication deficits are common and disabling after both left hemisphere (LH) and right hemisphere (RH) stroke, and can be quantified by the number of Content Units (CU; concepts mentioned by healthy speakers) in describing the cookie theft picture from the NIHSS. In LH and RH stroke, CU correlates with lesion volume. Hypothesis: Communication recovery in LH and RH stroke is influenced by % damage to specific areas or vascular territories, independently of lesion volume and initial severity. Methods: 25 acute ischemic RH (n=9) and LH (n=16) stroke patients were evaluated with DWI and described the NIHSS picture <24 from onset and again at 6 months. Communication recovery was measured by change in CU divided by baseline CU. A radiologist identified % damage to each cortical lobe, thalamus, basal ganglia, and each vascular territory on DWI, after co-registration to an atlas. We identified variables (age, infarct volume, initial CU, % damage to each region and vascular territory) associated with achieving the highest quartile of recovery (“best recovery”) using logistic regression. Results: Together, age, lesion volume, initial CU, and % damage to each cortical/subcortical area were strongly associated with best recovery (F(9, 15) = 12.55; p< 0.00001; r 2 = 0.88), but only initial severity (CU; t=-4.95; p<0.001) and % damage to basal ganglia (BG) (t=4.77; p<0.0001) independently predicted best recovery after adjustment for other variables. Likewise, age, lesion volume, initial CU, and % damage to each vascular territory were strongly associated with best recovery (F(10, 14) = 10.13; p = 0.0001; r 2 = 0.88), but only initial CU (t=-4.56; p<0.00001) and % damage to lenticulostriate territory (t=6.44; p<0.00001) independently predicted best recovery. To explore why higher % damage to BG predicted best recovery, we examined Day 1 PWI scans. Patients with BG strokes who showed best recovery had >50% damage to BG and had cortical hypoperfusion at Day 1. We speculate that good recovery was due to cortical reperfusion, but repeat PWI scans were not obtained. Conclusions: Patients who had the most room to improve on our measure of communication showed the most recovery. Independently of initial severity, more damage to left or right BG was associated with best recovery.