2.2. Data Extraction and Quality Assessment

For each included study, the following data were extracted: name of the first author, date of publication, study location, exposure type, time of follow-up, number of participants, events of type 2 diabetes, adjusted factors, and RRs with 95% Confidence Interval (CI). Both unadjusted and adjusted risk estimates were extracted, and we would perform meta-analyses on unadjusted and adjusted risk estimates separately. The quality assessment was performed by Newcastle Ottawa scale (NOS) [20]. Methodological quality was mainly assessed on the selection of participants, the comparability of exposure group and non-exposure group, and the ascertainment of outcomes. Studies with 7–9 stars were defined to have A-level quality, those with 4–6 stars were defined to have B-level quality, and those with 0–3 stars were defined to have C-level quality.