Vikas Pandey, MD

Abdul-Rahum A, Fulton R, Sucharew H, Kleindorfer D, Khatri P, Broderick J, et al. National Institutes of Health Stroke Scale Item Profiles as Predictor of Patient Outcome: External Validation on Independent Trial Data. Stroke. 2014

The NIH stroke scale is something repeated to the point of unintentional memorization for neurologists across the country, but the logical response to hearing the total NIHSS value that a patient has obtained is “What did the patient get points for?”. As inherent with any scale, the points may be achieved by an extremely high number of combinations. An NIHSS of 18 may be achieved by a patient with hepatic encephalopathy with no focal weakness or a patient with a full right MCA syndrome. Also now evident is that patients with low NIHSS scores (i.e. 0-5) may still be left with significant disability, nullifying the fact that significant disability is only for those with high NIHSS scores. To investigate the predictive value of different NIHSS profiles on stroke outcome, the group from the United Kingdom and Cincinnati retrospectively analyzed patients from the VISTA registry of collaborated acute stroke trails and took the patients that received either placebo or receiving a study drug with no proven efficacy on stroke outcome, however did include those receiving IV tPA as standard of care. 

The patient cohort included 10,271 patients with acute ischemic stroke and separated them according to NIHSS item profiles A to F, in decreasing order of stroke severity (median NIHSS 19 for profile A and NIHSS 4 for profile F). The profiles consisted of phenotypes essentially translated with profile A being a dominant hemisphere large stroke, profile B being a non-dominant hemisphere large stroke, profile C being a dominant stroke with language deficit, profile D being a dominant stroke without language deficit, profile E being a mild non-dominant hemisphere stroke and profile F being a very mild stroke. The more severe profiles were more likely to receive IV thrombolysis treatment. They analyzed the 90 day outcome in terms of mRS and mortality rates and found that across the different profiles, the less severe profiles had increasing odds of better outcome in a step-wise manner, even with adjustments made for age and thrombolysis treatment. Interestingly, those in profile C and D had similar baseline NIHSS (10 and 9, respectively) however had a significant difference in risk of 90 day mortality (43% and 81%, respectively relative to Profile A risk). These results were validated with secondary independent latent class analysis using different distinct NIHSS item profiles.

Looking deeper in to the difference in the profiles, one can see that the main difference between the profiles C and D mentioned were that profile C includes language deficit whereas profile D did not, and profile C had significantly less stroke mortality reduction compared to profile D. This is not surprising given the concept that aphasia has significant effects on rehabilitation prospects, leading to greater disability. Missed in the profiles, however, is the concept of neglect in non-dominant hemisphere strokes which is also a poor predictor of rehabilitation benefit. The group sheds light on the concept that comes to every neurologists’ mind when assessing the true symptoms a patient possesses when making decisions for thrombolysis and further intervention which is that a point given for aphasia is different than a point given for sensory deficit (all points are not equal). Perhaps this may be an early step in an NIHSS overhaul such as giving more points for things such as aphasia or finding a way to present patients as fitting a certain NIHSS profile instead of simply reporting a number, as we can see that the profile provides more valuable information regarding stroke outcome.