Statistical analysisCategorical var

Statistical analysisCategorical variables were presented as numbers andpercentages; continuous variables were described asmean and SD or median and interquartile range(IQR) depending on the distribution of data.Normality of data distribution was analysed bymeans of the chi-square test. Fisher’s exact test orchi-square test was used to compare categorical variables, where appropriate. Paired t-test or pairedWilcoxon test and unpaired t-test or Mann–Whitneytest were used to compare means/medians within agroup or between groups, respectively. Pearson test(for normally distributed data) or Spearman test (fornon-normally distributed data) were applied to analysecorrelations between two continuous variables. Alltests were two-sided with the significance level set atp < 0.05. Statistical analyses were performed withMedCalc statistical software 10.0.2.0 (MedCalc,Mariakerke, Belgium).The sample size was set with the use of dedicatedsoftware to allow a 1 kg of body weight to be a significant difference between the studied groups as a resultof the intervention. Based on previous studies, we haveassumed a mean difference of 0.7 kg in the controlgroup,11 and a mean difference of 1.7 kg in the studygroup with a SD of 2 kg. A probability of type-I errorof 0.05 and a power of the analysis of 80% required aminimal sample size of 126 subjects