In December 2013, the University of Saskatchewan Biomedical Research Ethics Board gave approval for this single-centre randomized double-blind placebo-controlled parallel-group trial. This project was completed at our tertiary care hospital where approximately 700 total knee replacements are performed annually. Inclusion criteria were adults aged 18-85 yr, American Society of Anesthesiologists (ASA) physical status class I-III, and scheduled for elective unilateral primary total knee arthroplasty under spinal anesthesia. Exclusion criteria included patients with relative contraindications to dexmedetomidine (e.g., allergy, heart block, and significant renal and hepatic impairment), patients with contraindications to morphine or spinal anesthesia, patients preferring general anesthesia, and patients taking narcotics.
The primary outcome was the consumption of morphine delivered via a patient-controlled analgesia (PCA) pump (LifeCare PCA™ Infusion System, Hospira, Lake Forest, IL, USA) in the first 24 hr following surgery. Secondary outcomes included morphine consumption at six and 12 hr postoperatively, time to first analgesic request via PCA, resting visual analogue scale (VAS) [(0-10) where 0 = no pain, 10 = worst imaginable pain] scores at six, 12, and 24 hr, discharge readiness time from the postanesthesia care unit (PACU), patient satisfaction with their analgesia in the first 24 hr, hemodynamic and respiratory changes intraoperatively and in the PACU, and opioid-related adverse effects (i.e., nausea, vomiting, pruritus) in the first 24 hr. All the outcomes were pre-specified prior to the start of the trial; however, the research team did not update the protocol at clinicaltrials.gov prior to the study in order to reflect all these non-primary outcome decisions.
Prior to entering the operating room, a member of the research team recruited the patients and obtained written informed consent. A computerized random number generator (www.random.org) was used to randomize the patients to either the dexmedetomidine or the placebo group in a 1:1 ratio. Access to the random number sequence and preparation of the series of sealed numbered envelopes was entrusted to a research coordinator not involved in patient recruitment, clinical care, or data collection. The patient, anesthesiologist, surgeon, and research team member involved with the patient were blinded to the group allocation. Once a patient was enrolled in the study, an anesthesiologist, who was not involved in the project or patient care, prepared a syringe (containing either normal saline or dexmedetomidine) according to the instructions provided in the sealed numbered envelope. The anesthesiologist then provided this numbered syringe to the attending anesthesiologist involved in the patient’s care.
Prior to entering the operating room, all patients were familiarized with the VAS and instructed on the use of the PCA pump. They were instructed to press the PCA demand button if their pain was ≥ 4 on the VAS.
All patients received oral premedication with acetaminophen 975 mg and naproxen 500 mg as well as an intravenous bolus of Ringer’s Lactate 500 mL 30 min prior to their scheduled operation time.
In the operating room, electrocardiography, noninvasive blood pressure, and pulse oximetry monitors were applied. In the dexmedetomidine group, patients received a loading dose of intravenous dexmedetomidine 0.5 µg·kg−1over ten minutes, followed by an infusion of 0.5 µg·kg·hr−1 delivered by a Medfusion® 3500 infusion pump (Smiths Medical, St. Paul, MN, USA). In the placebo group, patients received a loading dose and infusion of the equivalent volume of normal saline. Oxygen was delivered to all patients at 3 L·min−1 via nasal prongs. Once the loading dose of the study drug was started, spinal anesthesia was obtained in the sitting position using 0.75% hyperbaric bupivacaine 12.75 mg and fentanyl 10 µg. The anesthesiologist providing care for the patient continually assessed the intraoperative level of sedation and targeted a moderate level of sedation defined by the ASA15 (i.e., depression of consciousness with maintenance of purposeful responses to verbal commands/ light tactile stimulation, an unassisted patent airway, adequate spontaneous ventilation, and cardiovascular function). Intravenous midazolam 0-4 mg was available to both groups intraoperatively in order to achieve this goal, and no other intraoperative narcotics or sedatives were to be used. Phenylephrine and ephedrine were available at the discretion of the anesthesiologist in the event of hemodynamic changes. Although there were no specific hemodynamic targets with respect to the administration of phenylephrine and ephedrine, all anesthesiologists were informed of the potential for bradycardia and hypotension with dexmedetomidine and were instructed to provide hemodynamic support as per their clinical judgement. No peripheral nerve blocks were performed on the study subjects, and the surgeons did not administer local anesthetic infiltration during the procedures.
Postoperatively, the infusion was discontinued once the final dressing was applied. Patients were then transferred to the PACU and PCA was initiated. Delivery of morphine was standardized to 1.5 mg with a lockout of eight minutes. This PCA setting differs from our initial protocol registered with clinicaltrials.gov to account for convenience of using the institution’s standardized PCA order forms. This decision was made prior to patient enrolment, and all patients enrolled in the study received the same PCA orders. No supplemental analgesics were ordered for the first 24 hr following surgery. Dimenhydrinate 50 mg and ondansetron 4 mg were available if a patient experienced postoperative nausea or vomiting. Patients were discharged from the PACU once discharge criteria were met as per the modified Aldrete scoring system.16
All patient data and satisfaction scores were collected for 24 hr postoperatively by a research member who was blinded to the group assignment and not involved with the perioperative or postoperative care of the patient. Patient satisfaction was assessed by directly asking the patients to rate their satisfaction with the adequacy of analgesia in the last 24 hr as excellent, good, acceptable, or poor. Nursing staff blinded to group allocation assessed and recorded the VAS scores.
Statistical analysis was performed using SigmaPlot® version 13.0 (Systat Software Inc., San Jose, CA, USA).
Sample size was calculated using a previously published study in which mean [standard deviation (SD)] 24-hr morphine consumption for total knee arthroplasty was 58.6 (27.3) mg.17 A 50% reduction in morphine consumption was considered clinically significant as well as consistent with previously published studies showing decreased use of opioid analgesics following a dexmedetomidine infusion in patients undergoing general anesthesia.12 Using a two-sided test with an alpha of 0.05 and a power of 0.9, the required sample size for a two-sample comparison of means was estimated at 19 patients per group. The decision was made to enrol 20 patients per group to allow for potential patients who drop out.
Data were tested for normality using the Shapiro-Wilk test with a cut-off of 0.05. Student’s t tests were used to analyze the outcomes in data with normal distribution, including our primary outcome. The Mann-Whitney U test was used for data that were not normally distributed, and a Wilcoxon rank-sum test was used to analyze patient satisfaction.