Determinants of Outcome After Endovascular Middle Cerebral Artery Occlusion in Rats in the SPAN Trial

Abstract

BACKGROUND: The SPAN (Stroke Preclinical Assessment Network) is a confirmatory trial platform to test the efficacy and safety of candidate cerebroprotective interventions in acute stroke. As the largest multicenter preclinical stroke trial to date, the SPAN1 trial (first SPAN) prospectively captured many biological and procedural variables, revealing a high degree of heterogeneity introduced by the multicenter approach that may impact stroke outcomes. Here, we examined the biological and procedural predictors of tissue and neurological outcomes after focal cerebral ischemic stroke in rats. METHODS: SPAN1 enrolled and randomized 698 rats to various active treatment arms or controls. Rats were subjected to transient middle cerebral artery occlusion for 60 (spontaneously hypertensive rats) or 120 minutes (young, healthy Sprague-Dawley rats) and followed for 1 month. Eight biological and procedural independent variables (sex, weight, strain, intervention arm, site, endovascular filament silicone tip coating characteristics, anesthesia duration, and intervention protocol) and 5 dependent outcome variables (weight loss, 4-point neuroscore, corner test, infarct volume, and mortality) were captured. Multivariable regression was used to identify independent predictors of each outcome readout and determine their effect sizes. RESULTS: Spontaneously hypertensive rats exhibited larger infarcts than Sprague-Dawley rats, particularly among females. Neuroscores were also worse in spontaneously hypertensive rats. Prolonged anesthesia exposure was associated with smaller cortical and hippocampal infarcts. Filament thickness and length showed a complex association with different regional infarct volumes, neuroscores, weight loss, and corner test outcomes. Mortality was worse among females. Bivariate analysis of dependent variables revealed moderate correlations among the tissue and neurological outcomes. CONCLUSIONS: Using the large and multicenter, prospective SPAN1 dataset, our multivariable analyses identified several predictors influencing rat middle cerebral artery occlusion outcomes and refuted others previously reported. Investigators should consider whether biological and procedural predictors identified herein should be standardized, accounted for, or stratified during subject allocation to decrease variability and avoid confounders in future multicenter preclinical trials.