Carbohydrate intake and the risk of stroke: dose-response meta-analysis of cohort studies
Description
The results of dose-response meta-analysis of cohort studies on carbohydrate intake and the risk of stroke
Files
Steps to reproduce
We conducted updated meta-analyses on stroke events and deaths incorporating the Guangzhou Biobank Cohort Study results following the protocol of Meta-analysis Of Observational Studies in Epidemiology. Literature searches were conducted using PubMed, MEDLINE and EMBASE databases including studies published on or before April 31, 2021 based on the following query syntax: ((carbohydrate) OR (low-carbohydrate-diet [Title/Abstract]) AND ((stroke [Title/Abstract]) OR (“ischaemic stroke” [Title/Abstract]) OR (“ischaemic stroke” [Title/Abstract]) OR (“haemorrhagic stroke” [Title/Abstract]) OR (“haemorrhagic stroke” [Title/Abstract])). We conducted updated meta-analyses on stroke events and deaths incorporating the Guangzhou Biobank Cohort Study results following the protocol of Meta-analysis Of Observational Studies in Epidemiology.1 Literature searches were conducted using PubMed, MEDLINE and EMBASE databases including studies published on or before April 31, 2021 based on the following query syntax: ((carbohydrate) OR (low-carbohydrate-diet [Title/Abstract]) AND ((stroke [Title/Abstract]) OR (“ischaemic stroke” [Title/Abstract]) OR (“ischaemic stroke” [Title/Abstract]) OR (“haemorrhagic stroke” [Title/Abstract]) OR (“haemorrhagic stroke” [Title/Abstract])). After literature searching and the removal of duplicates, titles and abstracts of the articles were screened for the following inclusion criteria: human, prospective cohort studies, published in English. For articles that met the inclusion criteria, the full text was retrieved. The Newcastle-Ottawa Scale was used for quality assessment of each prospective cohort study. As there were two meta-analyses including six studies totally published up to November 2017, we conducted an additional hand search for relevant articles based on references of these meta-analyses and its included articles. The dose-response association was examined using Greenland and Longnecker’s method. Non-linear dose-response relationship was assessed by restrict cubic splines with three knots combined with multivariate meta-analyses. For all meta-analyses, heterogeneity was assessed with the I² statistic and Cochran’s Q test.