This single-center study was conducted by the Department of Anesthesiology, Peking University Hospital of Stomatology in Beijing, China. Ethics approval was received from the Biomedical Ethics Committee of Peking University Hospital of Stomatology (Number: PKUSSIRB–201734029) in December 2017. The trial is registered with the Chinese Clinical Trial Registry (Number: ChiCTR1800015347). This was a prospective trial with two parallel arms to test the hypothesis of whether conditioning with sevoflurane leads to fewer PPCs compared to those associated with propofol within 7 days after surgery. Eligible patients were enrolled and randomly assigned to receive one of two interventions, sevoflurane or propofol.
The study population comprised 220 patients between March 26, 2018, and March 25, 2019. Written informed consent was obtained before randomization from each patient. Patients were eligible for participation in the study if they were older than 18 years and younger than 80 years, were scheduled to undergo free flap (fibula or forearm) surgery with an expected duration of at least 4 h, and had a preoperative pulmonary complication risk index (Canet score) exceeding 26 points. The Canet score is a risk score for pulmonary complications, with a score of 26–44 representing moderate risk and that of 45 or greater representing high risk. 12
Patients were ineligible if they refused to participate in the clinical trial or had a body mass index of 35 or higher, severe chest wall malformation, acute phase of chronic obstructive pulmonary disease (AECOPD), acute phase of chronic bronchial inflammation, uncontrolled asthma, pulmonary artery stenosis, pulmonary hypertension and congestive heart failure, complex heart deformities, severe liver or kidney dysfunction, or a history of mental illness.
All patients were managed according to the same anesthesia protocol. Routine hemodynamic monitoring (continuous 5-lead electrocardiogram, pulse oximetry, and noninvasive blood pressure), as well as the bispectral index (BIS) (Covidien, USA), was performed and cannulation of the dorsalis pedis was completed immediately after anesthesia induction.
Anesthesia induction was carried out in both groups with 0.1 mg/kg penehyclidine hydrochloride, 0.05 mg/kg midazolam, 0.3 μg/kg sufentanil, 2 mg/kg propofol, and 0.6 mg/kg rocuronium. The parameters were volume-controlled ventilation, tidal volume (Vt) of 8 ml/kg, and fraction of inspiration O2 (FiO2)of 0.4–0.5; the respiratory rate was adjusted to maintain an end-tidal carbon dioxide concentration (ETCO2) between 35 and 45 mmHg.
In the propofol group, anesthesia was maintained by propofol as a target-controlled infusion (2 to 6 μ g /ml), while in the sevoflurane group, sevoflurane was applied with end-tidal concentrations of 2 to 5%. Analgesia was administered by applying target-controlled infusion of remifentanil up to 6 ng/ml and or boluses of sufentanil 0.2 to 0.5 μg/kg in accordance with patient needs. Muscle relaxation was achieved by intermittent injection of rocuronium bromide. The depth of anesthesia was to maintain a BIS between 40 and 60. In patients who underwent surgery for more than 4 h, cefuroxime sodium was used 30 minutes before surgery and for the fourth hour during surgery.
Randomization and blinding
Randomization was performed by an independent statistician, and random numbers generated by SAS 8.0 software were used to assign participants randomly (1:1) to either intravenous propofol or inhalational sevoflurane for maintenance of anesthesia. The codes were kept in sealed envelopes. Before surgery, these envelopes were provided to the attending anesthesiologist by a researcher not involved in patient care. Patients and surgeons did not know about the grouping during surgery and follow-up. In addition, the physicians who conducted follow-up examinations after surgery were blinded to the group allocation.
Postoperative patients undergoing tracheotomy (extubation on the fifth day) or endotracheal intubation (extubation the next morning) according to intraoral conditions were observed for a night in the post-anesthesia care unit (PACU). All patients were treated with dexmedetomidine sedation (4 µg/ml) in the PACU and intravenous patient-controlled analgesia (PCA) (48 h). If there were no special circumstances, patients returned to the ward the next morning.
Patients in the ward inhaled hydrocortisone three times a day up to discharge, and the vibrating sputum clearance device was used until the sixth day after surgery. Patients with forearm flaps were treated with cefuroxime sodium until the fifth day after surgery. Cefuroxime sodium and ornidazole were used up to 6 days for patients with fibula flaps after surgery.
Furthermore, when any pulmonary complication was suspected, bedside chest radiography (CXR) was also performed. The follow-up period was 7 days. Pulmonary complications were classified according to diagnostic criteria and assessed using the Clavien-Dindo classification, whereby grade 0 indicated no complication and grade V indicated death. 13,14 PPCs were diagnosed based on objective data such as blood gas analysis data, complete blood count, X-ray findings, and medical records in real time. Surgical complications (vascular crisis or hematoma), extrapulmonary complications, hospital stay, and mortality at 30 days were also recorded.
The sample size was calculated based on previous studies. The incidence of PPCs after a free flap surgery in the propofol group was 26%, compared to 54% in the sevoflurane group. 2 Considering that clinical research mostly involves studies with a small sample, the power is increased as much as possible to increase the credibility of the results. Therefore, in this study, 95% power was considered. In order to detect differences, it was necessary to include 90 patients per group with an alpha risk of 2.5% and a beta risk of 5% in a two-tailed comparison. The ratio of the two groups was 1:1. Considering a 20% loss rate, 110 pairs were enrolled.
Categorical variables were analyzed using the chi-squared test, continuity correction chi-squared test, or Fisher exact test based on sample size or frequency. The independent t-test was used for normally distributed continuous variables, and Mann-Whitney U test was used for non-parametric continuous variables. The relative risk and the 95% confidence interval of the differences were calculated for the primary outcome. Univariate logistic regression analysis was used to determine relevant baseline covariates associated with the primary outcome. If P values were less than 0.10 and were clinically relevant, then adjusted analyses were performed using a multivariate logistic regression model. Furthermore, the time to occurrence of the PPCs was compared using the Kaplan-Meier estimator, and the differences between groups were tested by the log-rank test. All analyses were performed using SPSS version 21.0. A two-sided P value of less than 0.05 was considered to indicate statistical significance.