Study design
This prospective randomized clinical trial was conducted and reported in accordance with CONSORT statement. Patients receiving primary total hip or knee arthroplasty from November 2017 to December 2018 were included. Institutional review board approval (2012-268) was obtained before patient enrollment. Written informed consent and research authorization were obtained from all patients before surgery. The study was conducted in compliance with the tenets of the Declaration of Helsinki, and the original protocol was registered at the Chinese Clinical Trial Registry (ChiCTR1800015896, April 27, 2018). The study was a little different from the registered protocol, regarding the exclusion criteria, samples size and intervention measures in Group B. Firstly, a 50% reduction in PONV was expected, which was the calculation basis of sample size in registered protocol. However, during the pilot study, patients in group B and C also experienced a high incidence of PONV, then we adjusted the expected reduction to 45% and the interventions in group B (oral mosapride instead of single dose of ondansetron) with the approval of monitoring committee.
Inclusion and exclusion criteria
Eligible patients included those at least 18 years old who were at risk of PONV (at least 1 score of Apfel), and scheduled for primary total hip or knee arthroplasty for end-stage joint diseases such as osteoarthritis, development dysplasia of hip, and osteonecrosis of the femoral head. Exclusion criteria included a history of intolerance of any drugs used in the current study, administration of another anti-emetic drug or systemic steroid within 24 h before surgery, allergy to experimental drugs or history of adverse reactions, diabetes with poor blood glucose control, history of steroid or immunosuppressive drug use within the previous 6 months, history of cardiac disease such as heart failure, heart block, ventricular arrhythmia or severe impairment of bowel motility, renal function or hepatic function.
Randomization and treatment
According to the inclusion and exclusion criteria, the researchers would enroll appropriate patients before surgery and obtain written informed consents. All included patients in this triple-blinded study were randomly allocated to three groups of prophylactic treatment for PONV using a computer-generated randomization list in a 1:1:1 ratio. The group assignments were performed by a research assistant and kept in opaque sealed envelopes that were only opened immediately before surgery.
- Group A received 8 mg ondansetron (4 ml) at the end of surgery; three doses of 2 ml normal saline at anesthetic induction and 6 and 24 hours later; and oral placebo 3 hours before surgery and three times per day after the operation.
- Group B received one dose of 10 mg dexamethasone (2 ml) at anesthetic induction, two doses of normal saline 6 and 24 hours later, 5 mg oral mosapride 3 hours before surgery and three times per day after the operation, and a dose of 4 ml normal saline at the end of surgery.
- Group C received three doses of 10 mg dexamethasone (2 ml) at anesthetic induction and 6 and 24 hours later, 5 mg oral mosapride 3 hours before surgery and three times per day after the operation, and a dose of 4 ml normal saline at the end of surgery.
The intraoperative dexamethasone and ondansetron were administered by an anesthesiologist, and the postoperative oral drugs were administered by nurses who were not involved in the study. The patients, surgeons, data collectors, and analysts were all blinded to the group assignments.
Anesthesia and perioperative pain management
All surgeries were performed under general anesthesia by the same team (F.X.P. and Z.K.Z.) using the standard medial parapatellar arthrotomy for total knee arthroplasty and posterolateral approach for total hip arthroplasty. A cemented posteriorly stabilized prosthesis was implanted in all patients undergoing total knee arthroplasty, and cementless acetabular and femoral components were used in all patients undergoing total hip arthroplasty.
All patients received the same anesthetic regimen and multimodal pain management protocol. Cefuroxime (1.5 g) was given intravenously as a prophylactic antibiotic prior to the first incision. Sufentanil (0.2 µg/kg), propofol (2 mg/kg), atracurium (1 mg/kg), and midazolam (2 mg) were used for anesthetic induction. Sevoflurane (1%–3%) and single bolus of sufentanil (0.1 µg/kg), and atracurium (0.5 mg/kg) were then used to maintain anesthesia during surgery. After prosthesis insertion, propofol (4 mg/kg. h) and remifentanil (0.1 µg/kg. min) were used to maintain anesthesia. The anesthetic drugs were discontinued before wound closure. All patients received preoperative oral hydration for up to 2 hours prior to surgery and adequate intravenous fluids intraoperatively.
After prosthesis insertion, a periarticular infiltration of 200 mg ropivacaine (100 mg/10 ml) in 60 ml normal saline was injected around the capsule before closure. One dose of 40 mg parecoxib was injected intravenously to manage pain. Postoperative pain control consisted of 50 mg of oral diclofenac sodium (Voltaren sustained-release tablets; Novartis Pharmaceuticals, Basel, Switzerland) every 12 hours. Breakthrough pain was recorded using a Numeric Rating Scale (NRS) ranging from 0 points (no pain) to 10 points (the most severe pain). Once the patients provided feedback that the pain was at a NRS score of >4 points, 10 mg oral oxycodone was administered. If the VAS score was >6 points, 50 mg of pethidine was given by intramuscular injection as required up to every 6 hours. No nerve block or intravenous patient-controlled analgesia was utilized perioperatively.
Outcome measurements
The primary outcome was the incidence of total PONV, and the secondary outcomes were complete response, times until first defecation and ambulation, postoperative appetite score, satisfaction score, length of hospital stay, blood glucose level, and complications. A blinded clinical investigator (Y.C.C.) recorded all episodes of nausea and vomiting, the severity of nausea, the antiemetic rescue requirement, and complete response to evaluate the antiemetic efficacy. Nausea was defined as a subjective unpleasant sensation associated with awareness of the urge to vomit, and vomiting was defined as the forceful expulsion of gastric contents from the mouth [12]. The incidence of total PONV was determined during the 48-hour study period by calculating the proportion of patients who experienced PONV at least once. Furthermore, newly developed PONV in each of the four periods (0–6, 6–12, 12–24, and 24–48 hours) was also calculated. Following institutional guidelines, 10 mg of intramuscular metoclopramide was used as a first-line antiemetic rescue treatment when patients experienced two or more episodes of PONV within 2 hours. This was followed by administration of 4 mg ondansetron intravenously when two consecutive boluses of metoclopramide alone administered at a 30-minute interval were ineffective
We used a standardized scoring algorithm to classify the severity of nausea during the 48-hour observation period into four degrees. Complete response was defined as no PONV or no requirement for a rescue antiemetic. Mild nausea was defined as the occurrence of nausea caused by an exogenous stimulus (drinking or movement). Moderate nausea was defined as the experience of nausea that required only one dose of a rescue antiemetic. Finally, severe nausea was defined as more than two emetic episodes or the need for more than one dose of a rescue antiemetic(15).
The blinded investigator (Y.C.C.) also recorded the time of first defecation, time of first ambulation, length of hospital stays, appetite score on postoperative days 0 to 2, and patients’ satisfaction score. The patients’ appetite was scored by comparison with the preoperative level (the morning on the day 1 before surgery); 1 point represented a worse appetite, 2 points represented no change in appetite from the preoperative state, and 3 points represented a better appetite. Patients’ satisfaction before discharge was estimated using a NRS that ranged from 0 points (extremely dissatisfied) to 10 points (very satisfied).
The fasting blood glucose level was measured in all patients on postoperative days 1 and 2, while the 2-hour postprandial blood glucose level was measured in patients with diabetes mellitus on postoperative day 1. All patients were followed up for 3 months postoperatively, and any complications such as prolonged QT syndrome, wound discharge, surgical site infection, or readmission were recorded.
Statistical analysis
We performed an a priori power analysis based on our preliminary results showing that the incidence of PONV was 49% in patients receiving ondansetron prophylaxis alone after total joint arthroplasty(12). We calculated that 303 patients (101 in each arm) were required to detect a 45% reduction in the incidence of PONV in Group C at an alpha level of 0.05 and power of 0.9 using a two-sided test. To allow for exclusions and dropouts, we enrolled 348 patients in the current trial.
The chi-square test or Fisher’s exact test was used to determine the statistical significance of differences in the categorical variables, such as the incidence of PONV and proportion of patients with complete response; if significant, multiple comparisons between groups were performed by Bonferroni’s corrected post hoc test (Z test). Continuous variables were analyzed with one-way analysis of variance if normality test was affirmative (body mass index, length of hospital stay, and times until first defecation or ambulation) or the Wilcoxon signed-rank test (patients’ appetite and satisfaction scores); if significant, multiple comparisons between groups were performed by the post hoc Tukey test. A p value of <0.05 was considered statistically significant. Statistical analysis was conducted using SPSS 21.0 (IBM Corp., Armonk, NY, USA).