Confirmation of tPA Treatment Effect by Baseline Severity-Adjusted End Point Reanalysis of the NINDS-tPA Stroke Trials

Abstract

Baseline severity-adjusted end point analysis, an emerging approach to the evaluation of primary end points in acute stroke trials, offers a novel means of adjusting trial analysis for baseline imbalances in presenting stroke severity among treatment groups, a factor that has complicated interpretation and reception of the results of the pivotal National Institute of Neurological Disorders and Stroke tissue plasminogen activator (NINDS-tPA) trials. The sliding scale dichotomy end point responder analysis applied in recent acute ischemic stroke clinical trials was used to analyze NINDS-tPA stroke trials 1 and 2. Good outcomes were: 3-month Rankin scale=0 if pretreatment NIHSS scores were 1 to 7; 3-month Rankin scale=0 to 1 if pretreatment NIHSS scores were 8 to 14; 3-month Rankin scale=0 to 2 if pretreatment NIHSS scores were >14. Both of the NINDS-tPA stroke trials showed a statistically significant beneficial treatment effect of tPA. In unadjusted analyses, in trial 1, good outcomes in tPA versus placebo patients were 39.6% versus 28.6% (odds ratio 1.64, P=0.049); in trial 2, 35.7% versus 24.2% (odds ratio 1.74, P=0.024). Among all 624 patients in trials 1 and 2 combined, good outcomes occurred in 37.5% versus 26.3% patients (odds ratio 1.68, P=0.0034). In the 91- to 180-minute onset to treatment time subgroup of patients among whom baseline severity imbalance was particularly severe, good outcomes were noted in 36.1% versus 24.0% (odds ratio 1.80, P=0.021). Odds ratios favoring tPA generally further increased after adjustment for 12 additional covariates known to predict acute stroke outcome. Baseline-adjusted severity end point reanalysis of the NINDS Stroke tPA trials confirms a beneficial treatment effect of intravenous tPA.