Vascular cognitive syndromes: relation to stroke etiology and topography

Abstract

Cognitive syndromes (CS) after stroke may be important to measure and monitor for management and emerging therapies. To describe the spectrum and frequency of CSs in the first month after stroke and to relate these to stroke etiology and topography. A validated cognitive examination was administered during the first month of stroke presentation and analyzed according to five large-scale networks for cognition and correlated with neuropsychological tests. A multivariate analysis was performed to determine association of CSs with etiology (TOAST classification), topography and neurological deficit by National Institute of Health Stroke Score (NIHSS). Of a total of 2105 patients, one or more patients with CS was present in 1569/1796 (87%) stroke patients vs 112/309 (36%, P < or = 0.001) transient ischemic attack (TIA) patients. The frequency of frontal network syndromes (FNS) was 908/1796 (51%), left hemisphere network (LH) syndromes 646/1796 (36%), right hemisphere (RH) network syndromes 275/1796 (15.3%), occipitotemporal network (OT) syndromes 107/1796 (6%), hippocampal limbic (HL) network syndromes 397/1796 (22%) and miscellaneous (M) syndromes 481/1796 (27%). Stroke etiology and their signature CS by multivariate analyses revealed significant associations for LH with cardioembolism (OR 1.61, P = 0.0029), FNS and 'other' etiology (OR 1.96, P < or = 0.0001) and HL also for 'other' etiology (OR 1.57, P = 0.0026). Coma (OR 2.95, P < or = 0.0001) and encephalopathy (OR 2.82, P < or = 0.0001) were both associated significantly with hemorrhage. A left hemisphere lesion was associated with LH CSs (OR 9.26, P < or = 0.0001). An FNS was associated with frontal lesions (OR 5.19, <0.0001) as well as subcortical lesions (OR 1.91, P < or = 0.0001). The M group of CS was associated with subtentorial (OR 1.86, P = 0.0283) and right hemisphere lesions (OR 2.47, P < or = 0.0001). The LH and RH syndromes had the highest NIHSS and differed significantly from all others. (1) CSs are present in the vast majority of stroke patients. (2) Particular stroke etiological subtypes are associated with specific CS. (3) Certain signature CS results from lesions that relate to the major anatomical cognitive networks.