Interventions to reduce cognitive decline post-stroke: an umbrella review

INTRODUCTION: Post-stroke cognitive impairment (PSCI) remains a common problem that can lead to significant disability and dependency. Previous work suggested that methodologically robust trials were lacking and that there is continued uncertainty on the benefits and limitations of pharmacological and non-pharmacological interventions. This umbrella review aims to update and synthesise evidence on pharmacological and non-pharmacological interventions for PSCI. PATIENTS AND METHODS: This review was pre-registered on PROSPERO (CRD42024611909) with searches from 2019 to 2024. Eligible studies were systematic reviews (SRs) of randomised controlled trials (RCTs) evaluating pharmacological or non-pharmacological interventions in adult stroke survivors (≥18 years) of any severity and reporting change in global cognition (eg, Mini-Mental State Examination). AMSTAR2 was used for risk of bias. RESULTS: Of 4889 records identified, 69 SRs (global cognition data from 747 unique RCTs; 57,176 participants) met inclusion criteria. Interventions were grouped into 37 categories including both single (eg, acupuncture, exercise, cognitive rehabilitation) and combined modalities and assessed at categorical level. Most intervention types were associated with statistically significant improvements in global cognition. Substantial heterogeneity and wide prediction intervals indicate that effect sizes may vary considerably across populations and settings. DISCUSSION: This review identifies promising evidence that several interventions may produce clinically meaningful improvements in global cognition post-stroke. However, confidence in this evidence is low: despite a substantial increase in the number of RCTs most included reviews were rated as critically low in quality with methodological limitations. CONCLUSION: Current evidence is insufficient to support firm recommendations for the inclusion of specific interventions in guidance.