2.4 Statistical analyses
Baseline characteristics of study participants were expressed as mean ± standard deviation for continuous variables with a normal distribution, or median and interquartile range (IQR) for continuous variables with a skewed distribution. Categorical variables were summarized as count and percentage. UACR, FPG, 2h OGTT, HbA1c, fasting insulin, HOMA-IR, HOMA-β, TG, ALT, AST, γ-GGT, and MET-h/week were logarithmically transformed prior to analysis due to skewed distributions. Characteristics between groups were compared using one-way ANOVA. Comparisons between categorical variables were performed with the χ2 test. Correlations between the indices of glucose metabolism (FPG, 2h OGTT, HbA1c, fasting insulin, HOMMA-IR, and HOMA-β) with UACR and eGFR were examined with Pearson’s correlation analysis and multivariate linear regression. Cox proportional hazards analyses were used to calculate incidence of CKD, and the results were expressed as hazard ratio (HR) and 95% confidence interval (CI). Model 1 was unadjusted; Model 2 was adjusted for age, sex and BMI; Model 3 was further adjusted for current smoking status, current drinking status, physical activity level, SBP, γ-GGT, and LDL-C. The relations of indices of glucose metabolism with CKD were also examined in subgroups stratified by age (≥ 58 or < 58 years old), sex (male or female), degree of obesity categorized by BMI (normal, overweight, or obese), central obesity (yes or no), diabetes (yes or no), and hypertension (yes or no). Interactions were tested by including strata factors, the quartile of glucose metabolism index and the respective interaction terms (strata factors multiplied by quantiles of glucose metabolism index) simultaneously in the models.
All statistical analyses were performed using SAS version 9.3 software (SAS Institute Inc., Cary, NC, USA). All statistical tests were 2-sided, and values of P < 0.05 were considered statistically significant.