Functional measures are recommended as primary outcome within clinical stroke trials but there is substantial heterogeneity in the assessments used. There may be a similar lack of consistency to outcome assessment in preclinical studies. Our aim was to collate studies using animal models of stroke and vascular cognitive impairment (VCI) and describe the functional assessment measures used.
A focused literature search was conducted across 14 chosen journals. Inclusion was limited to original articles describing preclinical research in stroke or VCI, published between January 2005 and December 2015 inclusive. Complete articles were reviewed for functional assessments used in each trial. Primary analyses were numbers of functional tests used; frequency of use of each test and temporal trends.
Of 91,956 stroke articles screened, 1,303 studies were analyzed, of these 609 did not report any functional measures and 56 had no access to full article, giving 638 articles for final analysis. For VCI, 56 papers were assessed in full, of these 16 reported no functional measures, giving 37 papers for final analysis.
In stroke trials, 559 (89%) used an ischemic and 94 (14%) a hemorrhagic model. In VCI studies, 11 different models were found, with global hypoperfusion being the most common, applied in 11 (30%) articles. Of studies analyzed, the median number of tests applied per trial was 1 in both cases with only 9% of stroke trials and 16% of VCI trials reporting functional measures as primary end points. There were 74 different assessments described in stroke trials, consisting of functional tasks and neurological deficit scores (NDS). Rotarod was the most frequent assessment (15% of trials), and Bederson’s the most frequently used NDS (39% of trials). In VCI studies, there were 19 different assessments employed, with Morris Water Maze (MWM) being most popular (56% of studies). Marked inconsistencies were found in physical and functional characteristics of all assessments reported.
There is heterogeneity amongst and lack of prioritizing the use of outcome measures in pre-clinical trials. These inconsistencies compromise comparison and meta-analysis. Thus, a consensus of core-set functional outcome assessments could improve consistency in animal stroke and VCI models.