Statistical analysis
The meta-analysis was performed using the RevMan software (Review Manager Version 5.3). For the outcomes of SBP, DBP, BMI, and lipids, Mean Difference (MD) and the 95% confidence interval (CI) were calculated using a fixed effects or random effects model. Where heterogeneity was substantially high (Chi2 P value of <0.1 and I2 value of >30%), we reported the outcomes using random-effect mode, otherwise the outcomes were reported using a fixed effect model. When standard error of the mean (SEM) or CI of means were reported instead of the SD, the SEM/CI were converted to SD. When median and intra-quartile range (IQR) were reported, the results were extracted as reported and are detailed in table 1. For the meta-analyses on CVD mortality, the number of deaths due to CVD and the total number of participants were used in the meta-analysis to analyse the risk difference. If the study reported the number as a percentage, then the number of participants/events was calculated based on the total sample size for each group. The methodological quality was assessed using the National Heart, Lung and Blood Institute (NHLBI) Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies.