This was a retrospective cohort study. After obtaining approval of the ethics committee of our centre and the patients’ written consent, the oncology and anesthesia databases were analyzed.
With retrospective screening of the laryngeal database, anesthesia database, electronic medical records, and follow-up status, we identified patients who had partial laryngectomy or total laryngectomy for LSCC between January 1, 2006, and December 31, 2016. Among them, 881 cases that underwent partial laryngectomy or total laryngectomy were enrolled into this study. The therapy for each patient was based on the guideline of National Comprehensive Cancer Network, NCCN), and the tumor location and the status of the tumor nodal and distant metastasis were defined on the basis of the 3rd to 6th editions of the Union for International Cancer Control (UICC) classification system and the 2nd to 7th editions of the American Association of Cancer (AJCC) staging system. Patients were excluded based on the following criteria: (1) pulmonary metastasis before surgery, (2) non-squamous cell carcinoma confirmed by postoperative pathology and (3) lost to follow-up.
During the study period, anesthesia induction was performed with propofol (1.5-2 mg/kg) or etomidate (0.3 mg/kg), fentanyl (0.003 mg/kg) or sufentanil (0.5 ug/kg), and cisatracurium (0.15-0.2 mg/kg). Anesthesia was maintained with sevoflurane (minimal alveolar concentration = 0.7-1.5), remifentanil (0.05-0.15 ug/kg/min) and cisatracurium (1-2 ug/kg/min). Opioids (fentanyl) and non-steroidal anti-inflammatory (flurbiprofen or parecoxib) were used for postoperative analgesia. The use of orotracheal intubation mainly depended on the condition of the tumor, especially the tumor size. In addition, the use of this procedure depended on the preference of the surgeons and anaesthesiologists for patients with advanced stage LSCC. In the tracheotomy group, tracheotomy was performed before induction of general anesthesia and required local infiltration anesthesia with 2% lidocaine. For patients in the intubation group, laryngoscopic orotracheal intubation was performed following the anesthesia induction. In both groups, male patients were intubated with a tracheal tube number 7.0, and female patients with a number 6.5. Opioid dosage was converted to equianalgesic (morphine) dosage to allow comparison of the two patient groups.
Patients at Sun Yat-Sen University Cancer Center had their first clinical and radiological evaluation 1 month after the final treatment, then had subsequent evaluations every 3 months during the first year, 2-3 times during the second and third years, and annually thereafter. Pulmonary metastasis was defined by histological verification or imaging (computerized tomography or enhancement computerized tomography), which manifested as single or multiple nodules of different sizes scattered in the lungs that gradually grew over time, some with pleural effusion. Clinical and histological confirmation of LSCC more three months after the initial treatment was defined as local recurrences, whereas LSCC diagnosed within three months of the primary therapy was defined as residual tumor . Meanwhile, lymph node metastasis detected by colour Doppler ultrasonography or computerized tomography was defined as growing lymph nodes , and some tissue samples were submitted to pathology for confirmation. Death was considered to be related to LSCC when patients died during treatment, or within the first 30 postoperative days, or if the medical records or death certificate documented laryngeal cancer as the underlying cause of death. Generally, the patients’ status and death causes were confirmed from death certificates or follow-up data. Postoperative complications included surgical incision bleeding, laryngeal fistula and subcutaneous emphysema.
Sex, age, height, weight, body mass index (BMI), smoking history, alcohol consumption and historical diseases were documented for each patient in the hospital information system. Historical diseases were defined as previously described health conditions , including cardiac diseases, chronic kidney diseases, hypertension, diabetes mellitus, and history of other cancers. In addition, patients in the intubation group were 1:1 matched with those in the tracheotomy group, based on operative time, pathological stage (T1/2, T3/4), clinical stage (1-4), type of surgery (partial laryngectomy, total laryngectomy), preoperative radiotherapy (yes, no) and postoperative chemotherapy (yes, no).
Descriptive statistics were presented as percentage, and continuous variables were presented as the mean ± standard. Data were compared by χ²-test and two-sample t test. All imbalanced variables with a significance level of P < 0.05 on χ²-test or t test were used to calculate the propensity score by logistic regression. According to Dr. Austin’s  recommendations, nearest-neighbour matching method with a calliper of 0.02 (0.2 x standard deviation) was used to perform a one-to-one match without replacement. Kaplan-Meier curves, log-rank test and hazards regression were used to analyse pulmonary metastasis, local recurrences, lymphatic metastasis and overall survival. In addition, considering the clinical significance but also the sample capacity, Cox regression was also used to analyse the risk factors with P-value < 0.1 in Kaplan-Meier curves of metastasis and recurrences. The results were expressed as 95% CI. The statistical significance was set at P < 0.05 and all tests were two-sided. All statistical analyses were performed using SPSS statistics 24.0 software for Windows (SPSS Inc, Chicago, IL, USA).