To evaluate agreement between National Institutes of Health Stroke Scale (NIHSS) assessed within 7-days follow-up and 90-day modified Rankin Scale (mRS) outcomes in randomized-controlled trials (RCTs).
Since factors independent of acute ischemic stroke (AIS) can confound 90-day outcomes in RCTs of acute treatments, early NIHSS may be a suitable alternative.
We conducted a meta-analysis including RCTs of acute therapies for AIS published until 01-April-2023, with available data for both 90-day mRS and NIHSS within 7-days. Primary outcome was agreement between trial results (classified as positive versus negative/neutral) based on 24-hour NIHSS and 90-day mRS. We additionally assessed agreement for 2-hours, 48-hours, 72-hours, 4-days, 5-days, 6-days and 7-days NIHSS. We aimed to validate our study-level findings using individual patient data from the ESCAPE and ESCAPE-NA1 RCTs, comparing NIHSS at 2-hours, 24-hours, 48-hours, and 5-7 days with 90-day mRS (both dichotomized at 0-2).
We included 116 trials (44,387 patients, median age 68 [IQR 66-72] years, 25,218 [56.8%] men), contributing 170 NIHSS assessments performed across the 8 timepoints. For 24-hour NIHSS, there was agreement with 90-day mRS assessments in 63/74 (85%) trials. Agreement varied numerically by timing of NIHSS assessments, ranging from 80%–100%, but differences were not statistically significant (p=0.76). Whereas ESCAPE and ESCAPE-NA1 data showed agreement in overall trial results for all timepoints except 2-hour NIHSS, agreement with 90-day mRS for individual patient outcomes was lower (3523/5195, 67.8%) but improved with later assessments (2-hours: 56.6%; 24-hours: 66.6%; 48-hours: 72.7%; 5-7 days: 76.5%,p<0.01).
NIHSS assessed between 24-hours and 7-days follow-up results in similar overall conclusions as 90-day mRS in identifying positive versus neutral/negative RCTs, indicating its value as an alternative primary outcome measure for AIS treatments. However, individual patient level data showed that NIHSS within 7-days risks misclassifying 1 in 3 patients with respect to 90-day mRS, although agreement improves with later assessments.