In a recent CMAJ article,1 David Gladstone and Sandra Black stated that the National Institute of Neurological Disorders and Stroke (NINDS) study2 provided valid evidence that patients treated with tissue plasminogen activator (tPA) within 3 hours of symptom onset achieved greater neurologic recovery and experienced less disability than patients who received placebo. Additional data published by the NINDS investigators3 and the US Food and Drug Administration medical officer's review of data submitted in support of a new drug application,4 show a significant in baseline stroke severity between the tPA-treated and placebo groups in the NINDS trial. Statistical correction for this baseline imbalance has not been provided in published reports and commentaries concerning this trial. Because baseline stroke severity has a significant effect on stroke outcome, I believe that accurate interpretation the results of the NINDS trial, or any similar trial, is not possible without using a statistically appropriate analytic equation to account for the differences in stroke severity between the trial groups.
The TOAST stroke trial5 demonstrated that very small differences in baseline stroke severity have large effects on stroke outcome. I applied stroke outcome information derived from this trial to the NINDS data.6 My analysis indicates that the difference in stroke outcome between the treatment and placebo groups in the NINDS trial may be accounted for solely by the baseline imbalance in stroke severity between the groups.
Jeffrey Mann Emergency Physician Salt Lake City, Utah
References
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.