This study was conducted at Queen Mary Hospital in Hong Kong, China. The study was approved by the Institutional Review Board of The University of Hong Kong/Hospital Authority Hong Kong West Cluster (UW 18–176), and registered at clinicaltrial.gov prior to patient recruitment on 24/07/2018 (NCT03597997). Written consent was obtained from all patients participating in the trial.
This study was prospective, double-blinded randomized controlled, conducted in accordance to the CONSORT guideline. All methods were carried out in accordance with Declaration of Helsinki. Patients aged between 18 and 80 years old with an American Society of Anesthesiologist (ASA) physical status of I-III scheduled for elective hepatectomy (left or right hepatectomy, segmentectomy, or wedge resection) were eligible. Exclusion criteria included the following: 1) known drug allergy to propofol, opioids, and non-steroidal anti-inflammatory drugs (NSAIDs) or COX-2 inhibitors, paracetamol, ketamine; 2) alcohol or drug abuse; 3) impaired renal function, defined as preoperative serum creatinine level over 120µmol/L; 4) impaired or retarded mental state; 5) body mass index (BMI) > 35kg/m2; 6) history of chronic pain; 7) pregnancy; 8) local infection; 9) history of psychosis; 10) delirium; 11) chronic opioid user; or 12) patient refusal (15).
Eligible patients were approached in the general ward before the operation for recruitment. The anaesthetic options were explained, and the patient was recruited into the study if s/he agreed. The patients were then randomized into one of two groups: i) SEVO group, anaesthetized by inhalational anaesthesia using sevoflurane; or ii) TIVA group, anaesthetized using propofol total intravenous anaesthesia. Patients recruited for hepatectomy were stratified in randomization. A computer-generated random sequence was used to select the allocation order, which was concealed in opaque envelopes, and opened at the time of intervention administration. Patients were not aware of the types of anaesthesia they received. A separate blinded investigator assessed the patients after the operation. The anaesthetist providing general anaesthesia was aware of the allocation, but s/he was not involved in data collection. Assessment was done at the general ward. Fasting for patients started at midnight before operation (15). Premedication was not prescribed.
On arrival to the operation theatre, a 20- or 22-gauge intravenous cannula was inserted. Standard monitoring with pulse oximeter, non-invasive blood pressure, and three lead electrocardiograms was applied prior to induction. For patients in the SEVO group, general anaesthesia was induced with propofol 1.5-3mg/kg, remifentanil 0.5-1mcg/kg, and rocuronium 0.6-1mg/kg or atracurium 0.5mg/kg given intravenously (IV). Sevoflurane, air and oxygen was used for maintenance of general anaesthesia. FiO2 was be kept between 35–50%. Bispectral Index (BIS) monitoring was applied and the level of anaesthesia was titrated to maintain a BIS value of between 40–60. Intravenous remifentanil infusion between 0.1–0.2 mcg/kg/min was given during surgery and titrated to provide optimal haemodynamic parameters. Ondansetron 4mg IV was given 30 minutes before the end of surgery. Sevoflurane and remifentanil infusion were switched off at the end of the procedure. Reversal of muscle relaxation was obtained with neostigmine 50mcg/kg IV and atropine 20mcg/kg IV after the operation. Patients were transferred to the post-anaesthetic care unit (PACU) for monitoring for at least 30 minutes. The anaesthetic and analgesic protocol for patients in the TIVA group was the same as the SEVO group. The only difference was that the induction and maintenance of general anaesthesia was performed using propofol total intravenous anaesthesia. Sevoflurane was not used, and oxygen and air were given. Target controlled infusion (TCI) with modified Marsh effect site model (Fresenius Kabi) was used for induction and maintenance of general anaesthesia. Level of anaesthesia was titrated to produce a BIS value of between 40–60. During induction of general anaesthesia, remifentanil 0.5-1mcg/kg, and rocuronium 0.6-1mg/kg or atracurium 0.5mg/kg were given IV. Remifentanil infusion was given as per SEVO group (15).
For both groups of patients, morphine sulphate at a bolus dose of 0.1mg/kg IV was given before skin incision. Additional 0.1mg/kg of morphine sulphate IV could be given in divided doses when the surgery continued for more than 2 hours at the discretion of the attending anaesthetist. Ketamine 0.5-1mg/kg IV was given before skin incision. Patients received local wound infiltration with up to 2mg/kg of levobupivacaine during wound closure (15).
Resting pain scores were checked every 5 minutes in the PACU. Morphine sulphate at a dose of 2mg IV was given every 5 minutes until the numerical rating scale (NRS) pain score was less than 4/10. Respiratory rate, oxygen saturation, Ramsay sedation scores, blood pressure and heart rate were monitored every 5 minutes in the PACU. Patient controlled analgesia (PCA) morphine was connected to the patient once the NRS pain score was less than 4/10. The following parameters were set for the PCA morphine machine: bolus 1mg of intravenous morphine sulphate with each patient demand, lockout duration of 5 minutes, and maximum dose limit of 0.1mg/kg of morphine sulphate per hour. When the patient resumed fluid diet in the ward on POD 0, regular oral dihydrocodeine was prescribed at a dose of 30mg three times a day for two days. Afterwards, dihydrocodeine was given as needed. Breakthrough pain was treated by intramuscular/subcutaneous morphine 0.1mg/kg every 4 hours as needed starting from POD 0 (15). Pain related parameters such as NRS pain scores at rest and during coughing, cumulative PCA morphine doses and side effects were recorded every 4 hours.
PCA morphine was given for at least 2 days. It was stopped on POD 2 if NRS pain scores during coughing were less than 4/10 and morphine consumption was low. PCA morphine was continued if NRS pain scores during coughing were equal or greater than 4. After discontinuation of PCA morphine, NRS pain scores (at rest and during coughing) and the dose and frequency of rescue analgesia used were recorded daily until patient discharge. Patient satisfaction with analgesia, where 0 corresponded to the least satisfaction, and 10 corresponded to the most satisfaction, was assessed on POD 1 (15). Quality of recovery 9 (QoR-9) was evaluated on POD 1 and POD 3, which contains 9 questions to assess postoperative changes in emotion, well-being, social function and physical disability (16).
Various outcomes were also assessed at 3 and 6 months after surgery. This included the presence or absence of pain, and NRS pain scores at rest and with coughing (15). Patient satisfaction with analgesia, health related quality of life (assessed using the 12-Item Short-Form (SF-12) Health Survey (17)) and psychological status (assessed using the Hospital Anxiety and Depression Scale (HADS) questionnaire) were also evaluated.
Statistical analysis
The primary outcome was the NRS pain score during coughing on POD 1. Our previous retrospective study on hepatobiliary and pancreatic surgery had shown that the mean (SD) of NRS pain score during coughing on POD 1 for TIVA group was 4.30 (1.99) (18). To detect a difference in NRS pain score of 1.3 out of 10 with a significance level of 0.05 at power of 80%, 37 patients per group were needed. An NRS pain score difference of 1.3 has been regarded as the value corresponding to minimal clinically significant change (19). The sample size was estimated using the methods described for clinical trials (20). To consider for possible dropouts, 45 patients were recruited into each group, giving a total of 90 patients.
For patient demographics and intraoperative analgesic consumption, continuous data were analysed using independent-samples t-test or Mann-Whitney U test, and categorical data were analysed using Pearson Chi-square test or Fisher’s Exact test. HRQOL domains for SF-12, satisfaction with analgesia and psychological condition were compared with Mann-Whitney U test. The standardized pain scores both at rest and during coughing were expressed as area under curve (AUC) weighted by the corresponding time interval. The weighted AUC was equivalent to a time weighted average of the pain scores for the specified time interval, and was of the same scale as the NRS (0–10). The difference in the standardized pain scores between the TIVA and SEVO groups were compared using the Mann-Whitney U test, and expressed as median (interquartile range). The cumulative PCA morphine consumption at 24, 48 and 72 hours after the surgery between the two groups was also compared using the Mann-Whitney U test. Incidence of adverse effects was compared using the Chi-square test or Fisher’s exact test as appropriate. IBM SPSS Statistics Version 27.0 (IBM Corp. USA) was used to analyse the above data. False discovery rate (FDR) is the expected proportion of the rejected null hypotheses that are actually true, and it is a recommended alternative in health-related studies (21, 22). Benjamini-Hochberg false discovery rates (FDRs) were applied to correct for multiple-hypothesis, and an FDR value of less than 0.1 was regarded as statistically significant (23).