Background
The gap between expectations and outcomes in clinical stroke trials of neuroprotection has brought into question the validity and sufficiency of preclinical evidence.
Objective
A review was undertaken in order to identify therapies tested experimentally in animal models of neuroprotection and to assess the quality of experimental evidence supporting the use of such agents in acute stroke patients. Furthermore, to better understand the currently preferred experimental model of stroke - the focal ischemia model (STAIR, 1999) - a meta-analysis was conducted using data from the focal ischemia experiments.
Data Sources
Experimental acute stoke treatments and studies were identified from PubMed, from clinical trials databases and by cross-reference. For each drug, a PubMed search was conducted using the terms (drug name) AND (neuroprotection OR cerebral ischemia OR stroke).
Study Selection
Studies measuring treatment efficacy in controlled culture and in vivo models of neuroprotection were selected for analysis.
Results
Over 1000 experimental treatments were identified, of which, about 100 have been administered to acute stroke patients. Using a ten-point quality-of-evidence scale based on the Stroke Therapy Academic Industry Roundtable recommendations (1999), the general level of preclinical experimentation fell short of these drug development standards. Whilst the quality of evidence from animal experiments was found to be higher in those drugs that have gone to clinical trial, extensive testing of these drugs in animals was frequently undertaken only after treatments had already been given to patients. Further, the average protection afforded by clinically tested treatments did not differ significantly from those still in the experimental phase (30% vs. 25% respectively).
Conclusion
With hindsight, it is not evident that the best drugs have been selected for clinical trial. More rigorous preclinical testing and a greater understanding of experimental models will assist in the evaluation of promising therapies for clinical trial.