Statistical analysis

We will use intention-to-treat (ITT) and per-protocol (PP) populations in the analysis. The ITT population consists of all individuals who will be randomized, whereas the PP population consists all participants who complete the 8-week intervention. The data will be revised randomly to check accuracy and completeness. All data will be reported as mean ± SD. The percentage changes for each variable will be computed by the following formula:

where E and B are the end value and the baseline value of variable, respectively. The data normality will be analyzed using the Kolmogorov-Smirnov test. To compare parametric continuous data between and within the groups, the independent sample t test and the paired sample t test will be used, respectively. In addition, to compare the differences in asymmetric variables between and within the groups, the Mann-Whitney test and Wilcoxon test will be applied, respectively. The analysis of covariance (ANCOVA) test will be used to control confounding variables. To evaluate the association between changes in fibroblast growth factors (FGF19, FGF21), β-Klotho co-receptor concentrations and other variables, linear regression models will be used. SPSS version 21 (IBM, Armonk, NY, USA) will be used for data analysis. A p value < 0.05 will be considered statistically significant.