All analyzes were performed with SAS V.9.10 software (SAS, Cary, NC, USA). Regarding the sample size estimation, we assumed a 0.15 mm difference in the CIMT between the exenatide and insulin group based on previous liraglutide and DPP-IV studies [19 (link), 20 (link)]. Considering a standard deviation of 0.2 mm, 28 patients per treatment group were needed for the primary endpoint analysis to warrant a power of 80% with a two-sided significance level of 0.05. Assuming a drop-out rate of 15%, the sample size needed for each group would be 33, for a total sample size of 66 patients.
The full analysis set was used for statistical analysis. For the primary endpoints, the least-squares mean change from baseline to 52 weeks and associated 95% confidence intervals and P values for exenatide versus insulin were derived from a mixed model for repeated measures with age, sex, duration of T2DM and CIMT at baseline as fixed covariates. Normally distributed data are expressed as means and standard deviations, and the t-test was used for comparison. Skewed data were compared with a nonparametric test. The count data are expressed as proportions, and the frequencies were compared with a Chi-square test. A P < 0.05 indicated a statistically significant difference. All statistical analyses were carried out with IBM SPSS statistical software V.22 for Windows (IBM Corp., Armonk, NY, USA).
The full analysis set was used for statistical analysis. For the primary endpoints, the least-squares mean change from baseline to 52 weeks and associated 95% confidence intervals and P values for exenatide versus insulin were derived from a mixed model for repeated measures with age, sex, duration of T2DM and CIMT at baseline as fixed covariates. Normally distributed data are expressed as means and standard deviations, and the t-test was used for comparison. Skewed data were compared with a nonparametric test. The count data are expressed as proportions, and the frequencies were compared with a Chi-square test. A P < 0.05 indicated a statistically significant difference. All statistical analyses were carried out with IBM SPSS statistical software V.22 for Windows (IBM Corp., Armonk, NY, USA).