Objective To provide an overview of meta-analysis on the effect of exercise on the cognitive function of patients with Alzheimer's disease (AD), and to put forward improvement suggestions.

Methods PubMed, Scopus, CNKI, and other databases were electronically retrieved. The retrieval time was from the date of database construction to December 31, 2020. According to the inclusion criteria and exclusion criteria, final articles were selected and relevant literature information was extracted by screening the abstract and full text. The methodological quality of the included meta-analysis was evaluated using AMSTAR2 tool, and the quality of the outcome indicators of the included meta-analysis was further graded according to the GRADE system.

Results A total of 9 meta-analyses were included. The AMSTAR2 evaluation results showed that 4 studies were in low methodological quality, and 5 studies were in extremely low quality. The GRADE evaluation results showed that among the 17 outcome indicators included, 3 were intermediate quality indicators, 9 were low ones and 5 were extremely low level ones. The main deficiencies of methodology quality were that no protocol was provided or registered, no details of literature screening, data extraction, and reasons for exclusion, no source of funding and related conflicts of interest for each included literature. In addition, the implementation process of systematic reviews was not rigorous. The main flaws of evidence quality included designable and implementation limitations, inconsistent results, insufficient precision, and possible publication bias.

Conclusions The meta-analysis indicates that exercise can improve the cognitive function of AD patients to some extent, but the methodological quality of relevant meta-analysis is relatively low and the quality of the evidence is generally low. It is suggested that the related authors should follow rigorous meta-analysis report specifications. Journal editorial departments should strictly control meta-analysis report. Universities and evidence-based medicine centers build registration scheme platforms with policy support, and offer courses on meta-analysis theory and practice, which will provide a more scientific, accurate and high-quality clinical trials evidence for the field of exercise prevention and treatment of AD.