Improving patient selection for clinical acute stroke trials

Objective: To optimize patient inclusion criteria for clinical acute stroke trials. Methods: We stratified probabilities of death and complete recovery based on two validated prognostic models using age and the National Institutes of Health Stroke Scale (NIH-SS) at admission. In an independent data set of 1,725 consecutively admitted patients with acute ischemic stroke, computer simulation with various inclusion thresholds was used to calculate the number and percentage of potential treatment responders, i.e. who had not died or spontaneously recovered. Results: Using defined thresholds for recovery and mortality, inclusion and exclusion criteria could be designed to considerably decrease both trial time and study size compared to a fixed inclusion criterion based on the NIH-SS alone. Other thresholds may allow optimization of either trial time or study size. Conclusions: The resulting models provide a validated approach for an efficient study inclusion of potential treatment responders based on the Barthel Index 100 days after ischemic stroke. These techniques provide the opportunity for improved stroke trials in terms of enrollment speed, treatment effect size or both.