Validation of the National Institutes of Health Stroke Scale in Intracerebral Hemorrhage

Abstract

We sought to determine if the National Institutes of Health Stroke Scale (NIHSS) has a greater discriminative power than Glasgow coma scale (GCS) to identify patients at risk of poor early functional outcomes and large hematoma volumes. We prospectively collected clinical assessments, imaging, and outcome data in consecutive patients with intracerebral hemorrhage, and determined the ability of GCS and NIHSS to predict poor functional outcome (modified Rankin scale 3–6) and hematoma volume >30 cm 3 using receiver operating characteristics analysis, C‐statistics, and the DeLong test. We studied 672 patients with intracerebral hemorrhage (mean age 62±14 years; 56% men; median intracerebral hemorrhage score=1, interquartile range (IQR) 0–2; median intracerebral hemorrhage volume 7 cm 3 , IQR 2–19) with median NIHSS of 8 (IQR 3–18) and GCS 15 (IQR 7–15). NIHSS correlated strongly to GCS (r=−0.773; P <0.001). Admission NIHSS (C‐statistic: 0.91; 95% CI, 0.89–0.93) predicted better than GCS (0.78; 95% CI, 0.75–0.81) discharge poor functional outcome (DeLong test P <0.001). NIHSS (0.82; 95% CI, 0.78–0.86) also discriminated better than GCS (0.78; 95% CI, 0.73–0.83) patients with large hematoma volume (DeLong test P =0.029). The NIHSS has a greater discriminative power than GCS to identify patients at risk of poor early functional outcomes and large hematoma volumes.