151751-Najiba-Chargi

162 CHAPTER 9 STATISTICAL ANALYSIS Categorical data are represented as a number and percentage of the total. A test for normality (Kolmogorov-Smirnoff test) and histograms were used to assess whether continuous variables were normally distributed. Continuous data are represented as mean±SD if normally distrib - uted, andmedian±interquartile range if skewed. Fisher’s exact tests, Pearson Chi square tests, independent sample t tests, and Mann-Whitney U tests were used to assess group differences where appropriate. Univariable andmultivariable backward stepwise logistic regression anal - ysis was used to examine the association between vascular calcification and PCF. Parameters entered as covariates in regression analysis were chosen based on known or expected associa- tion with a PCF. Not all anatomical locations could be assessed in all patients, most commonly when CT imaging of the thoracic area was not available or because of dental artefacts. The missing data were considered missing at random. Multiple imputation of these missing sites was applied to replace the missing values for logistic regression analysis, using the median of 20 imputated datasets.22,23 All statistical analyses were performed using the IBM SPSS Statistics version 25.0 software package (Chicago, Illinois, USA). All analyses were two-sided and p ≤ 0.05 was considered significant. RESULTS Between January 2008 and June 2017, 245 patients underwent TL at our institution. Of these 245, 17 patients were excluded because there was no CT imaging available, and 4 patients were excluded because of inadequate quality of imaging. Therefore, 224 patients were included in this analysis. Median interval between imaging and TL was 27 days. PATIENT AND TREATMENT CHARACTERISTICS The 224 patients that were included for analysis had a mean age of 64.8 years. Patients were predominantly male (82.1%). During the study period, 105 patients (46.9%) underwent pri - mary TL, 108 patients (48.2%) underwent salvage TL, and 11 patients (4.9%) underwent a functional TL. Prior to total laryngectomy, 99 patients (44.2%) had undergone radiotherapy and 21 patients (9.4%) had undergone chemoradiotherapy. A PCF occurred in 62 patients (27.7%), which required surgical closure in 40 patients (64.5% of all PCF). Patient, disease and treatment related characteristics, and their relationship with the occurrence of a PCF are presented in Table 2. To summarize, patients who had a PCF more often had hypopharyngeal cancer, a dysfunctional larynx after treatment, sarcopenia, laryngectomy with pharyngectomy and flap closure of the neopharynx. Of note, additional lymph node dissection and prior treatment for head and neck cancer did not appear to be more common in patients with a PCF. Patients with a PCF were not significantly older and did not have a significantly lower BMI.