For calculation of the study sample size, we used a priori G*power analysis (Faul, Erdfelder, Lang, Buchner, 2007, version 3, University of Dusseldorf, Germany) based on α (type I error) = 0.05 and β (type II error) = 0.2 (power = 80%). Thirty-eight patients per group were sufficient to find a difference of 25% in the incidence of EA between them. There was a predicted 5% dropout of cases, so 40 subjects were required in each group to detect such difference.
Analysis of the collected data was performed using the SPSS program for Windows (version 26, IBM Corp, Armonk, New York, USA). At first, data normality was checked using the Kolmogorov–Smirnov test. Qualitative or categorical variables were presented as numbers and percentages to be compared using the Chi-square test or Fisher's exact test. Continuous variables were expressed as mean ± standard deviation if normally distributed, and median (range) for nonnormal data. Both the groups were compared parametrically using Student's t-test, and nonparametrically with MannWhitney U-test. All data were considered statistically significant if P ≤ 0.05, where more significant results were obtained with smaller P values.
Analysis of the collected data was performed using the SPSS program for Windows (version 26, IBM Corp, Armonk, New York, USA). At first, data normality was checked using the Kolmogorov–Smirnov test. Qualitative or categorical variables were presented as numbers and percentages to be compared using the Chi-square test or Fisher's exact test. Continuous variables were expressed as mean ± standard deviation if normally distributed, and median (range) for nonnormal data. Both the groups were compared parametrically using Student's t-test, and nonparametrically with MannWhitney U-test. All data were considered statistically significant if P ≤ 0.05, where more significant results were obtained with smaller P values.