Study selection and characteristics
As shown in Figure 1, of the 1788 records identified in the electronic database, 1435 records were screened based on title and abstract after removing all duplicate records; 48 full-text articles were assessed for eligibility, among which 39 full-text articles were excluded because of overlapping participants,28, 29 including participants with history of dementia or moderate to severe cognitive dysfunction,12, 13 and other reasons . Finally, 9 studies were included in the quantitative synthesis.10, 11, 15-19, 30, 31
As presented in Supplementary Table S3, of the 9 included studies, 6 were conducted in Europe, 2 in North America, 1 in Asia; all studies were observational; 2 were prospective design, and 7 were retrospective design; a total of 613,920 patients were included; the follow-up time were varied from about 2 to 10 years; 1 study included patients with the history of first-ever ischaemic stroke only, and the others included mixed patients; 4 studies set the observational window varied from 6 months to 4 years. The quality of included studies was high, reflected by a NOS score of 9 (Supplementary Table S4).