Patient selection
The Institutional Review Board (IRB) at The University of Texas MD Anderson Cancer Center approved this study. The informed consent was waived by IRB because this is a retrospective study. A cancer registry including records of 4011 adult (age ≥ 18 y at the time of surgery) patients with primary diagnoses of OCC or OPC who subsequently underwent resection of the primary OCC or OPC tumor at MD Anderson from 2007 to 2012 was used as the primary study database. Patients who underwent tonsillectomy for positive cervical lymphadenopathy with unclear primary disease, whose cases had been used for teaching of flexible endoscopy for tracheal intubation and patients who were <18 years of age at the time of HNRT were excluded from this study. In addition, because the majority of the patients in the registry received intensity modulated radiation therapy, those who underwent brachytherapy or proton therapy were excluded in order to ensure the uniformity of the sample.
All the patients in the registry who qualified for the study were assigned into 1 of 2 groups according to whether they had received HNRT: the HNRT group, which consisted of patients who received HNRT before tumor resection, and the non-HNRT group, who underwent upfront resection without having received HNRT. Because some of the data for this study were embedded in the notes describing tracheal intubation in patients’ medical records, requiring a manual search, we used a matching strategy to avoid having to hand-search the entire registry. We first identified the HNRT group as described above; then, we used the exact matching method to select matched controls from the non-HNRT group. The control patients were selected according to age, sex, and body mass index (BMI) to match the patients in the HNRT group at a ratio of 1:1. The matching range for age was ± 5 y. For BMI, the 2 groups were matched at 6 levels: ≤ 18.5 kg/m2, 18.6–25.0 kg/m2, 25.1–30.0 kg/m2, 30.1–35.0 kg/m2, 35.1–40.0 kg/m2 and ≥ 40.1 kg/m2. Each patient in the HNRT group was successfully matched with a non-HNRT control patient.
Data collection
The following data were electronically retrieved from the patients’ medical records: age, sex, BMI, American Society of Anesthesiologists physical status score (ASA score), airway assessment (mouth opening, neck range of motion, edentulous and MP scores), cancer diagnosis, type of surgery, and whether the patient had a history of HNRT. Data on radiotherapy and on the method of tracheal intubation were manually retrieved from the records after matching. The missing data in the primary data sets were manually searched for and placed in the corresponding data set. When a patient had multiple surgeries after HNRT, the data for the first surgery with tracheal intubation after HNRT were used. For airway assessment, most anesthesia providers in our practice considered patients to have trismus if they could open their mouths < 2 finger breaths (typically < 3.5 cm). However, no standardized criteria were used to measure neck extension, so the neck range of motion was based on providers’ subjective judgment. In this study, edentulous referred to a patient with complete upper, lower, or whole-mouth removable dentures. We included the grade of laryngeal view during tracheal intubation in the analysis to reflect the intubation effort according to the Cormack-Lehane system. 5 The grade of laryngeal view mainly applied to the patients who had been intubated via either direct laryngoscopy or video laryngoscopy of any type. For flexible endoscopy trachea intubation, grade I was used for data analysis.
Statistical analysis
Summary statistics, including mean, standard deviation, median, and range, were calculated for continuous variables, such as age, BMI, and the interval between radiotherapy and surgery. Frequencies and percentages were used to summarize data for categorical variables, such as sex, BMI, Mallampati (MP) score, mouth opening, neck range of motion, cancer stage, radiation status, and airway intubation status. The t test was used to evaluate differences in continuous variables between the 2 patient groups. Fisher’s exact test or a chi-square test was used to test for associations between radiation status (HNRT or control) and patient characteristics that may be associated with difficulty of tracheal intubation, including sex, BMI, MP score, mouth opening, neck range of motion, cancer stage, intubation difficulty status (difficult or easy) and patient characteristics. Odds ratio and its confidence interval were used to assess the effect of radiation status on intubation difficulty status. A P value of less than 0.05 was considered statistically significant. The statistical software SAS 9.3 (SAS, Cary, NC) was used for all the analyses.