Anthropometric and Laboratory Data
The data for each registered patient was weight loss (kg), measured using a calibrated digital scale, body mass index (BMI), and height and weight measurements were recorded twice at each visit to increase accuracy and reliability); bag size, measured using the graduation endoscope in relation to length (all bags that allowed the device to be retrovisioned were considered wide).
All outcome variables, except blood tests (total plasma cholesterol, high density lipoprotein cholesterol (HDL) and HbA1c) were determined at randomization and in the second, fourth and sixth months after randomization.
Statistical analysis
The statistical analysis of the data has performed by an external collaborator (Dr. Idiberto José Zotarelli Filho) and interpreted by the principal investigator (Dr. Ricardo Fittipaldi-Fernandez). For data analysis, a database has built on the Microsoft Excel spreadsheet which has exported to the Minitab 18® statistical program (version 18, Minitab, LLC, State College, Pennsylvania, USA) (Minitab®)[22]and OriginPro ® 9 (DPR Group, Inc., Northampton, Massachusetts, USA).[23] A common descriptive statistical analysis has performed, obtaining the values of total “n”, mean and standard deviation, confidence interval (CI), and percentage for all the variables. Normality test has performed for all variables. One-Way test (ANOVA) has applied, adopting the α-level less than 0.05 with a statistical difference for 95% CI. The R-sq (R2) value has also analyzed.