Abstract W P42: Ischemic Lesions Grow, but NIHSS Scores Improve Over 48 Hours After Stroke: Implications for Clinical Trials

Abstract

Background: DWI lesion growth is often used as a surrogate outcome measure in early-phase stroke clinical trials. Several studies have shown that admission DWI volumes and 3-mth clinical outcomes are strongly correlated, and that DWI volumes and NIHSS scores correlate at various times after stroke. However, there are few quantitative data comparing the dynamic changes of NIHSS scores to DWI lesion growth in acute ischemic stroke. Methods: We analyzed data from a phase II clinical trial of normobaric oxygen therapy (NBO) in acute ischemic stroke. Patients ineligible for tPA with imaging-confirmed ischemic stroke <9h and NIHSS>4 were randomized to NBO or Medical Air, delivered for 8h. There was no upper limit for age, NIHSS or stroke volume. NIHSS and DWI scans were obtained at 0 h, 4h (during therapy), 24h, and 48h. In this trial NBO had no significant effect on NIHSS change or DWI growth, hence data from NBO and Air groups were combined for this analysis. Results: Serial DWI and NIHSS were obtained in 60 subjects (mean age 72 yrs, 53% male). Mean DWI lesion volumes increased over time (baseline, 44 cc; 24 hrs, 68 cc and 48 hrs, 76 cc, p=0.1 ) but median NIHSS scores decreased over time (baseline, 12; 24 hrs, 10; 48 hrs, 9; p=0.1 ). There was no significant correlation between 0-24 hour change in NIHSS scores and 0-24 hour % DWI growth (r square=0.01, p=0.38), or absolute DWI growth (r square=0.033, p=0.21). There was no significant correlation between 0-48 hour change in NIHSS scores and 0-48 hour % DWI growth (r square 0.005, p=0.63), and only a mild correlation with absolute DWI growth (r square=0.11, p= 0.03). Stroke etiology affected these correlations: there was a significant correlation between NIHSS change and DWI growth at 48 hrs for cardio-embolic strokes, but not for large-artery strokes. Linear regression analysis adjusting for age, gender, mismatch and arterial recanalization status, showed that change in NIHSS was not associated with DWI growth neither at 24 ( p=0.41) nor at 48 hours (p=0.16). Conclusion: While the dynamic relationship between clinical deficits and lesion growth differs according to stroke mechanism, on average, infarcts grow but neuro deficits improve in the first few days after ischemic stroke. The mechanisms underlying this paradox require further study.