The impact of imbalances in baseline stroke severity on outcome in the National Institute of Neurological Disorders and Stroke recombinant tissue plasminogen activator stroke study

Study objective: The National Institute of Neurological Disorders and Stroke (NINDS) Recombinant Tissue Plasminogen Activator (rtPA) Stroke Study demonstrated a clinically meaningful and statistically significant benefit of tissue plasminogen activator (tPA). Adjusting for the baseline National Institutes of Health (NIH) Stroke Scale, the benefit of tPA remained. However, other authors suggest that an imbalance in baseline stroke severity between the tPA and placebo groups confounded the results. Another issue that has been raised concerns a possible increase in early mortality for individuals given tPA. In post hoc subgroup analysis, we describe the effect of tPA across a spectrum of time from stroke onset to treatment and stroke severity subgroups. Stroke severity was based on the NIH Stroke Scale. We also compare early mortality (2-week and 30-day) in the tPA and placebo groups. Methods: Using combined data from the 2 NINDS rtPA Stroke Study trials, we performed post hoc subgroup analyses of 3-month favorable outcome (defined by the NIH Stroke Scale, Barthel, Rankin, and Glasgow Outcome Scales). We categorized patients from the trials into onset to treatment (0 to 90 minutes, 91 to 180 minutes) by NIH Stroke Scale (<= 5, 6 to 20, > 20) subgroups. Analyses were adjusted for all variables previously shown to be associated with favorable outcome at 3 months. We also compared early mortality within onset-to-treatment subgroups. Results: For all the 12 specified onset-to-treatment-NIH Stroke Scale subgroups, the adjusted odds ratio for a favorable 3-month outcome was greater than 1.0 and favored tPA. We detected no difference in mortality between patients treated with rtPA and those treated with placebo by 2 weeks posttreatment (rtPA=9%, placebo=13%; P=.49) or by 30 days (rtPA=11%, placebo=16%; P=.30). Conclusion: These are descriptive post hoc subgroup analyses. Using cut points defined in previous critiques of the NINDS trials, these analyses give results consistent with previous NINDS Study Group reports. Baseline NIH Stroke Scale imbalance does not account for the better outcome of rtPA-treated patients.