*2.5. Statistical Analysis*

Statistical analysis was performed with Statistica TIBCO 13.3 software (TIBCO Software Inc., Palo Alto, CA, USA) and R (version 3.6.1). We used the Shapiro–Wilk test to verify the distribution of the data. As appropriate, continuous variables were reported as a median with 0.25–0.75 quartiles or as a mean with standard deviation. They were compared by using the Mann–Whitney U test or unpaired t-test, respectively. Categorical variables were given as percentages and compared by χ<sup>2</sup> test. The variables of the CBA assay were Box-Cox transformed, and one-way covariance analysis (ANCOVA) was performed to adjust for potential confounders, including age, sex, and BMI. To evaluate the relationship between continuous variables, a Spearman rank correlation test or Pearson correlation tests were performed as applied. The cutoff points for the oxidative stress parameters

were calculated based on the receiver operating characteristic (ROC) curve to estimate the odds ratio (OR) with a 95% confidence interval (CI). Independent determinants of cumulative in time concentration, K concentration, and rate factor R were established in multivariate linear regression models, built using a stepwise forward selection procedure, verified by Snedecor's F-distribution; R2 was evaluated as a measure of variance. Unconditional multivariate logistic regression model and one-way variance analysis (ANOVA) were used to analyze the independent impact of comorbidities, including hypertension, diabetes mellitus, hypercholesterolemia, and oral steroid and statin therapy on evaluating oxidative stress, respectively. Results that had a *p*-value less than 0.05 were considered statistically significant.
