Neuroanatomic correlates of poststroke hyperglycemia

Objective A study was undertaken to determine associations between ischemic stroke sites and poststroke hyperglycemia (PSH). Methods Nondiabetic patients with first ever ischemic stroke confirmed by imaging were prospectively included. Blood glucose level (BGL), National Institute of Health Stroke Scale (NIHSS) score, and clinical parameters were assessed on admission. BGL was dichotomized for elevated versus normal levels using a cutoff value of >7.0 mmol/l. Clinical parameters were correlated with BGL and were compared between patient groups with elevated versus normal glucose values. A voxel‐based lesion symptom mapping (VLSM) analysis adjusted for confounding variables was performed correlating sites of ischemic lesions with PSH. Results Of 1,281 stroke patients screened, 229 (mean age = 66.3 ± 15.9 years) met the inclusion criteria. Patients with elevated BGL were older, had higher NIHSS scores, and had larger infarcts compared to those without elevated glucose levels. Spearman rank analysis showed correlations between BGL and age, infarct size, heart rate (HR), and NIHSS scores (p ≤ 0.05). The VLSM analysis adjusted for these confounding factors demonstrated associations between PSH and damaged voxels in right hemispheric insular and opercular areas. Interpretation The data indicate that damage in the right insulo‐opercular areas contributes to PSH. The association between sympathetically mediated increase of HR and BGL suggests disinhibition of sympathetic outflow as a possible mechanism for PSH. Ann Neurol 2015;77:262–268.