Version 16

Statistical analyses were conducted using SPSS version 16.0 for Windows (SPSS, Chicago, IL) and MedCalc version 16.8 (http://www.medcalc.be). A P value < .05 was considered statistically significant. Continuous variables are shown as the mean ± standard deviation (SD), and categorical variables as case number and percentage. Comparisons between groups were performed using Student t test for continuous variables, and the Chi-squared test for categorical data. Partial correlation analysis was used to assess associations between the obesity indices and metabolic risk factors.
Adjusted odds ratios (ORs) and 95% confidence intervals (95% CIs) of 1-SD incremental increase in the obesity indices in association with MetS and its other components were calculated by multiple logistic regression analyses. Age, education level, smoking status, alcohol consumption, physical exercise, menopausal status (only women), and family history of the corresponding condition were used as confounders for multivariate analysis. ROC curves were plotted and analyzed using MedCalc version 16.8 (http://www.medcalc.be). Sensitivities and specificities were calculated and AUCs were compared to show the efficacy of the various obesity indices to discriminate subjects with and without MetS.

Based to the skewed distribution of the parameters most data are demonstrated as median and range. Mann–Whitney U or chi-squared test were conducted for two-group analysis while Kruskal–Wallis test was used for analysis between multiple groups. Subgroup analysis was illustrated as box plots including median, quartiles, range, and extreme values of the given data with whiskers ranging from the minimum to the maximum value. Statistic outliers defined as a value that is smaller than the lower quartile minus 1.5 times the interquartile range or larger than the upper quartile plus 1.5 times the interquartile range were displayed as separate points [17 (link)]. Spearman linear correlation analysis was performed to assess differences between various variables. Kaplan–Meier curves were generated to estimate the survival function and a log-rank test was performed to assess differences between groups [18 (link)]. Receiver operating characteristic (ROC) curves were generated by plotting sensitivity against 1-specificity evaluating the value of a predictive marker or a composite score [19 (link)]. Differences between ROC curves were analyzed as previously described [20 (link)]. SPSS Version 23 (SPSS, Chicago, IL, USA) and MedCalc Version 16 (MedCalc Software, Ostend, Belgium) were used for all statistical analyses.

The sample size was estimated based on the principle of detecting a 5% difference in the prevalence of recurrent goiter with a 90% probability at p < 0.05. The univariate relation between patient characteristics and the development of goiter recurrence and the need for reoperation were examined. The statistical significance of categorical variables was evaluated by the χ² test and F test, whereas the Student’s t-test was used to evaluate continuous variables. Ten-year recurrence-free survival was calculated using the Kaplan–Meier method, with the log rank test for comparison between study groups. All the data were entered onto a dedicated spreadsheet (Microsoft Excel 2010; Microsoft Corporation, San Jose, CA, USA) by a medical assistant and then analyzed by a statistician (MedCalc, version 16, Belgium). p < 0.05 was considered to indicate significance.