Abstract Body: Introduction: Many post-cardiac arrest (CA) therapies shown to be effective in animal models have failed to translate in human clinical trials. The predominance of single site pre-clinical studies, which lacks the heterogeneity and variability inherent in multicenter human clinical trials, could contribute to this failure. Multicenter pre-clinical studies might yield variability that more closely approximates human clinical trials, and thus result in reduction of bias and better initial screening of interventions. Here we describe the variability of CA characteristics among participating centers of an international multicenter preclinical CA research network.
Methods: This is a retrospective analysis of control group subject data from swine CA studies submitted to the TRANSatlantic Cardiac arrEst Neuroprotection aDvancement (TRANSCEND) Network pre-clinical registry. Data for 101 animals from ten different models at seven institutions across the U.S.A, Europe, and Asia were included in this analysis. Data related to CA characteristics (intervals), post-CA characteristics (hemodynamics and blood gas at 1 hour post-ROSC), and neurological outcome were compiled and compared between the ten models. One-way ANOVA and Bartlett Test were used to estimate differences in means and standard deviations (SD). Chi Square test was used to compare proportions.
Results: There were significant differences in CA characteristics between all ten models. The mean (SD) interval from CA to start of cardiopulmonary resuscitation was 8 (4)min (range of 2 to 14min) and from onset of CA to ROSC was 19 (7)min (range 13 to 38min) (p<0.0001). There were also significant differences between models when comparing sex, weight, post-ROSC hemodynamics, arterial blood gas results, and neurological outcome (p<0.0001). Data is detailed in the table.
Conclusion: This analysis illustrates significant variability in CA and post-ROSC characteristics and neurologic outcomes when comparing different swine CA models in an international pre-clinical registry. Such variability could make interventional CA studies performed in a multicenter network more predictive of human clinical trials.