The failure of neuronal protective agents versus the success of thrombolysis in the treatment of ischemic stroke. The predictive value of animal models.

Agents claimed to be neuroprotective in animal stroke models have all failed in human trials. Thrombolysis has been reported as beneficial in animal and human stroke. We explore the reasons for this disparity, using a review of published results of agents tested both in animal stroke models and in human stroke trials. In animals the effect of neuroprotective agents and of thrombolytic agents on infarct size is time-dependent: early initiation of treatment works best; and benefit is progressively--and eventually totally--lost with increasing delay of time of first treatment. The animal data also show that, overall, the beneficial effects of the neuroprotective agents are weaker, and are totally lost sooner, than those of thrombolytics. The human data show that the failed trials of the neuroprotective agents had entry windows that went far beyond the windows of (any) success seen in tests of these agents in animals. By contrast, human thrombolysis trials uniformly restricted time of entry to windows in which these agents have shown beneficial effect in animals. In clinical stroke trials, neuroprotective agents failed to produce benefit because their effects at best are too weak, and they were used at times predictable from the animal models as too late. Thrombolytic therapy, which has a stronger effect than neuroprotective agents in animal models, was used clinically during the early window of optimal effectiveness, and produced beneficial results. "Too little/too late" is the recipe for failure in the treatment of ischemic stroke.