Why do neuroprotective drugs work in animals but not humans?

Animal models of stroke are not a perfect representation of human stroke, but they provide critical chemical and physiologic information for the development of technology and medications used in human studies. Animal models provide a large degree of assurance of reproducible injury from subject to subject, thus allowing for smaller numbers to compare treatment and placebo arms. One does not control the size of the obstructing clot or the vessel being occluded in human trials, so one should work to improve our ability to measure the injury in each patient to determine the initial insult. MR diffusion/perfusion scans could be used to measure tissue at risk so as to be able to assess more accurately the efficacy of an agent. Animal studies are subject to greater control of the variables that affect outcome and, therefore, have been able to use single outcome measures (usually infarct size) in small numbers of animals to show benefit of a neuroprotective agent. Human trials are subject to a wider variety of factors that are not easily controlled. The logical course would then be to refine our capabilities of determining and analyzing outcome measures. This refinement can include using imaging end points with MR imaging, and using statistical methodology such as nonparametric procedures and binary outcome measures instead of single, continuous outcome parameters. Not all drugs that work in animal models will work in humans; therefore, carefully designed pilot studies are very useful to help determine which drugs are likely to be translatable to humans before large and expensive phase III trials are undertaken. Information from the basic work should be translated into human studies, and questions raised and identified in clinical trials should lead to further hypothesis testing in the animal models. This testing can be done by evaluating questions generated from the primary end point conclusions and by developing new hypotheses regarding mechanisms based on data obtained from the measurement of secondary surrogate markers and the expected and unknown responses of the subject. It is no surprise given the organ at risk that cerebrovascular injury is a complex issue and one that will not lend itself to an easy solution. It is incumbent on the neuroscience community to nurture the obvious symbiotic relationship between the basic and the clinical research fields in order to bring effective neuroprotective agents to clinical use as expediently possible.