Developing New Stroke Treatments Using Preclinical Randomized Controlled Trials

Abstract

Listen

Translate article

Figure 2. Results are pooled from multiple centers performing 2 different stroke models. Anti-CD49d did not show a benefit in the fMCAO model while it did show a benefit in the combined data of the cMCAO model. Preclinical stroke research has suffered from a lack of translatability into human clinical trials. With the exception of tissue plasminogen activator, none of the countless therapies that have shown promise in animal models have demonstrated a beneficial effect in actual patients. As a result of multiple high-profile failures, pharmaceutical companies have eliminated investment in this area of research, forcing a critical evaluation of all stages of the drug discovery pathway. In 1999, the Stroke Therapy Academic Industry Roundtable (STAIR) was established to identify problems in stroke research methodology and make recommendations to guide researchers and funders in the evaluation of experimental quality. A large body of literature has since been produced on improving and standardizing animal testing for stroke therapies. In 2006, O’Collins et al. reviewed the history of experimental stroke research, identifying 1026 experimental stroke therapies published between 1955 and 2003. They reviewed the methodology and results of 912 therapies that were tested in animal models, 17 in only human trials and 97 in both animal and human trials. Each study was given a grade ranging from 0 to 10 on the basis of how many of the STAIR criteria were fulfilled, called the “scope of testing.” They observed that poor scope of testing correlated with a higher reported percentage of neuroprotection (Figure 1). They also noted that whether or not a drug advanced to clinical trials did not correlate with efficacy or scope of testing. With the STAIR preclinical testing recommendations and the poor historical record summarized by critical papers like O’Collins et al., funding for clinical trials has required a higher standard of preclinical evidence. Preclinical studies, however, still lacked the homogeneity and regulation to minimize bias and maximize accuracy. Several international research consortia including MultiPART (Multicentre Preclinical Animal Research Team), funded by the European Union for Stroke Research, and CAESAR (Consortium for PreclinicAl assESsment of cARdioprotective therapies), funded by the National Institutes of Health, have called for preclinical randomized controlled trials (pRCTs) and revealed inadequate data reporting, statistical flaws, and missing crossvalidation of positive results. In August, Llovera et al. published the first pRCT in any disease process. Six centers with demonstrated proficiency in 2 murine models of middle cerebral artery occlusion (MCAO) tested natalizumab, an anti-CD49d antibody and possible neuroprotective agent (Figure 2). The drug had passed all of the old hurdles of an optimal drug candidate to advance to human trials including minimal side-effect profile,multiple years already on themarket for a different disease process, and demonstrated efficacy in an animal model. The preclinical studies, however, were performed at individual institutions and produced conflicting resultswith 4 positive studies and