Cognitive Assessments in Randomized Controlled Trials of Acute or Secondary Prevention Stroke Treatments 2011 to 2024: A Scoping Review.

Cognitive impairment after stroke is common and associated with poorer outcomes. However, cognition is rarely assessed in acute or secondary prevention randomized controlled trials (RCTs), and those that do have not been systematically synthesized. This review examines how often cognitive end points are used, specific assessments applied, domains tested, and rates of missing cognitive data. We performed a search on EMBASE, PsycINFO, MEDLINE and Cumulative Index to Nursing and Allied Health for RCTs involving adult (≥18 years) participants with stroke, published from 2011 to 2024. Abstracts and full-text publications were screened for stroke RCTs with cognitive end points. These were categorized into acute, secondary prevention, and rehabilitation trials. Descriptive statistics summarized the frequency of cognitive end points, domains, and missing data rates in acute and prevention trials. Of 12 822 screened studies, 980 met criteria for full-text screening and 406 were stroke RCTs with a cognitive end point. Among these, 43 were acute and secondary prevention RCTs eligible for data extraction. The Mini-Mental State Examination (22/43 studies, 52%) and Montreal Cognitive Assessment (18/43 studies, 42%) were used in the most RCTs. There were 66 distinct cognitive tasks used, with greatest diversity in memory tasks (19), executive tasks (11) and global screens (11). Mean missing data were 22.2% (SD: 10.4%, range 0%-62%). Cognitive tasks are infrequent outcomes in stroke RCTs. When used, the tasks and domains assessed vary widely and are heavily affected by missing data. More pragmatic approaches to measuring meaningful cognitive change in all clinical trial participants are needed.