Background: The National Institutes of Health Stroke Scale (NIHSS) is widely used to measure neurological deficits, evaluate the effectiveness of treatment and predict outcome in acute ischemic stroke. It has also been used to measure the residual neurological deficit at the chronic stage after ischemic events. However, the value of NIHSS in ischemic cerebral small vessel disease has not been specifically evaluated. The purpose of this study was to investigate the link between the NIHSS score and clinical severity in a large population of subjects with CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy), a unique model to investigate the pathophysiology and natural history of ischemic small vessel disease. Methods: Demographic and clinical data of 220 patients with one or more lacunar infarcts confirmed by MRI examination and enrolled from a prospective cohort study were analyzed. Detailed neurological examinations, including evaluation of the NIHSS and modified Rankin Scale score (mRS) for evaluating the clinical severity, were performed in all subjects. The sensitivity, specificity, positive and negative predictive values of various NIHSS thresholds to capture the absence of significant disability (mRS < 3) were calculated. General linear models, controlling for age, educational level and different clinical manifestations frequently observed in CADASIL, were used to evaluate the relationships between NIHSS and clinical severity. Results: In the whole cohort, 45 (20.5%) subjects presented with mRS >= 3, but only 16 (7.3%) had NIHSS >5. All but 1 subject with NIHSS >5 showed mRS >= 3. NIHSS <= 5 had an 85.3% positive predictive value for no or slight disability with only 33.3% specificity. The NIHSS, MMSE score and presence or absence of gait disturbances were found to be strongly and independently correlated with disability (all p < 0.001). Altogether, they accounted for 73% of the variance of mRS in contrast with the NIHSS alone accounting for only 50% of this variance. Among patients with NIHSS <= 5, subjects with mRS >= 3 showed a lower MMSE score than those with mRS <3 (p < 0.001). All patients with NIHSS <= 5 but with mRS >= 3 presented either with gait disturbances or MMSE score <25. Conclusions: The present results suggest that the NIHSS cannot reflect the extent of neurological deficit and clinical severity in subjects with lacunar infarctions in the context of a chronic and diffuse small vessel disease. A specific and global neurological scale, including the assessment of cognitive and gait performances, should be developed for ischemic cerebral microangiopathy. Copyright (C) 2012 S. Karger AG, Basel