Defining a Target Population to Effectively Test a Neuroprotective Drug

Background and Purpose: We aim to identify the subgroup of acute ischemic stroke patients with higher probabilities of benefiting from a potential neuroprotective drug using baseline outcome predictors and test whether different selection criteria strategies can improve detected treatment effect. Methods: We analyzed the association between final infarct volume (FIV), measured on 24- to 72-hour computed tomography, and National Institutes of Health Stroke Scale at discharge/day 5 of acute stroke patients who underwent endovascular treatment. Models were adjusted for age, sex, and affected hemisphere. We analyzed the impact of absolute (5–15 mL) and relative (33%) FIV reductions in the National Institutes of Health Stroke Scale in the whole population and in different subsets of patients selected according to baseline imaging criteria using computed tomography perfusion. Results: We analyzed 627 patients; association between FIV and 5-day National Institutes of Health Stroke Scale was best described with a quadratic function, with a regression coefficient β=1.56 ([95% CI, 1.45–1.67]P <0.001) in the adjusted analysis. In the models considering a fixed absolute (5/15 mL) FIV reduction, treatment effect was highest when patients with predicted larger FIV were excluded, whereas in a 33% FIV reduction model, treatment effect increased with the exclusion of patients with expected excellent outcomes.Conclusions: Patients either with excellent outcomes after endovascular thrombectomy or with large infarcts may dilute the treatment effect in stroke neuroprotective drug trials. Computed tomography perfusion on admission may help selecting adequate patients according to expected drug effect profile.