Animal models of stroke: are they relevant to human disease?

Scientists have been modeling ischemic stroke in animals for over 150 years and have devised many ingenious methods to try to mimic the human condition. Nevertheless, the task remains formidable, even for the most talented scientist utilizing the most sophisticated technology. The challenge would be exceedingly difficult, even if we already understood all the pathophysiology underlying human stroke. While the use of these experimental models has provided much information about the methods of producing and potentially treating cerebral ischemia and infarction in specific animal species and experimental circumstances, the relevance of most of these data to the human condition remains dubious. Pulsinelli and Buchan recently pointed out that although animal models of cerebral ischemia have been used extensively to test new therapies in human stroke, their record for identifying clinically effective drugs has been disappointing. They noted 25 different compounds of "proven" efficacy for treating focal and global brain ischemia in animal models over the past 10 years based on published articles in refereed journals. Among those compounds subjected to clinical trials, none has proven efficacious, nor have any of these agents come into general clinical use. To explain the failure of the pharmacologic studies to translate from the animal model to the clinical setting, the authors point to "inexperience and scientific frailties" of the investigators and "weak experimental design and technique," including lack of historical controls, inadequate numbers of animals to rule out Type II errors, inadequate physiological monitoring, inappropriate histological analysis, and lack of control over reproducibility of the ischemic insult. In a similar commentary, Molinari eloquently summarized the rationale for using animal stroke models in stroke research, pointing out several advan-