Explore individuals’ changes in modified Rankin Scale (mRS) between 90 to 365 days after ischemic stroke onset and if they alter the results of pivotal stroke trials.

The most common 1° outcome for randomized controlled trials in ischemic stroke is a dichotomized 90-day mRS. 

In a secondary analysis of patients in the NINDS tPA (Parts 1 and 2) and IMS III trials with mRS at 90 and 365 days, we used the same 1° outcome as those trials: mRS 0-1 (good) vs. 2-6 (poor) for NINDS tPA and 0-2 (good) vs. 3-6 (poor) for IMS III. We compared the 1° outcome by randomization group at 90 and 365 days and evaluated intrapatient mRS shift (?mRS=day 365–day 90 mRS). 

We included 624 patients from NINDS tPA and 606 from IMS III. In NINDS tPA, the randomization of tPA vs. placebo had good outcome in 42.6% vs. 26.6% at 90 days (p<0.001) and 40.7% vs. 27.6% at 365 days (p=0.001). Despite 1° outcome consistency, 59 (9.5%) patients shifted 1° outcome category (p<0.001 for patient change in 2x2 table by chi-squared test). The patient-level ?mRS was: ≤-2=3.0%, -1=13.9%, 0=63.5%, 1=9.8%, ≥2=9.8%.  In IMS III, the randomization of tPA+endovascular thrombectomy vs. tPA had good outcome in 30.0% vs. 28.2% at 90 days (p=0.654) and 34.8% vs. 32.5% at 365 days (p=0.583). Again, despite 1° outcome consistency, 58 (9.6%) patients shifted 1° outcome category (p<0.001 for change). The patient-level ?mRS was: ≤-2=3.8%, -1=19.6%, 0=64.4%, 1=8.4%, ≥2=3.8%. Our findings were maintained in sensitivity analyses where we excluded patients who died or had recurrent stroke.

Although 1° outcome shifts from 90 to 365 days after ischemic stroke were balanced enough to maintain the conclusions of NINDS tPA and IMS III, patient-level shifts were significant and suggest continued dynamism to post-stroke recovery up to a year from onset.