Criteria for valid preclinical trials using animal stroke models.

T he possibility that stroke may be preventable and once developed may be amenable to therapeu- tic intervention is a major impetus in stroke research.Advances in our understanding of the patho- physiology of cerebral ischemia and the biochemical mechanism of ischemic neuronal death have raised hopes of preventing or reducing brain injury after ischemic stroke.A recent surge in claims of successful therapeutic interventions in preclinical trials using ani- mal stroke models is encouraging.However, the pros- pects for an imminent breakthrough in treating stroke patients based on animal studies was appropriately dampened in recent editorials in Stroke." 2 Innovative interventions that were successful in animal stroke models in the early 1980s failed to materialize in stroke patients in subsequent multicenter clinical trials.3-6To date, not a single therapeutic agent has been approved by the Food and Drug Administration (FDA) for treat- ment of acute ischemic strokes in this country.Difficul- ties in extrapolating therapeutic success in animal mod- els to clinical trials challenge the legitimacy of preclinical trials using animal stroke models.12A sub- sequent editorial, however, eloquently defended the future use of animal models in studying the biochemical mechanism of ischemic brain injury and new drug development.7Zivin and Grotta7 and later Milliken8 also pointed out a serious "mismatch" between animal experiments and therapeutic interventions in stroke patients.Can we avoid recurrent failure to replicate preclinical therapeu- tic success in future clinical trials?Experienced re- searchers are less likely to be optimistic.First, one must recognize that many shortcomings inherent to clinical trials are virtually absent in animal experiments.These, including the lack of more objective outcome measures, diversity in stroke pathology, heterogeneity of demo- graphic factors, concurrent medical problems, and vari- able delay in starting treatment, greatly magnify the difficulty in discerning the neuroprotective effect of a promising drug in multicenter clinical trials.It is safe to state that to demonstrate a favorable drug effect will continue to be much easier in animal models than in stroke patients in the Decade of the Brain.