*2.5. Statistical Analysis*

Means and standard deviations (SD) from eligible studies reported high and low MD adherence for each parameter of MetS were used. Wherever it was necessary, and data were presented as median, minimum or maximum values or 95% confidence intervals (CI), conversion to mean and SD was performed [51–54]. When values of FBG, TG and HDL cholesterol were presented as mmol/L, conversion to mg/dL was employed using the Omni calculator [55]. The inverse variance method was used in order to estimate the weight of each study. The random effects model was used due to higher methodological heterogeneity among the included studies [56,57]. Moreover, Hedge's g was used as effect size and standardized mean difference (SMD) as a summary statistic model due to the heterogenous scores using in included studies for the definition of low and high adherence to MD [56]. Estimation of heterogeneity was performed with Cochrane Q test (*p* < 0.1: existence of heterogeneity) and I2 statistic [56,57]. I2 values >50% indicated substantial heterogeneity across studies. Publication bias was assessed with funnel plots and Egger's test [53]. All statistical analyses were performed using the R software developed at Bell Laboratories (formerly AT&T, now Lucent Technologies version 4.0.2).
