Subjects will be randomized to a placebo or treatment group in a 1:1 ratio in blocks of 10 using an online randomization tool. The randomization table numbers for the subjects who are randomized, but subsequently withdrew consent after enrollment or had their case cancelled will be also included. Only the research pharmacists will have access to the securely-stored randomized group allocation list, which will remain blinded until statistical analysis is completed. A research pharmacist will prepare 100 mL normal saline infusion bags, which contain either 1,000 mg IV acetaminophen or placebo. Identical in appearance, the bags will be then delivered to either the operating room satellite pharmacy or ICU.
The subjects will receive a total of 6 doses of 1,000 mg IV acetaminophen or placebo at the following time points: (1) immediately after anesthesia induction, but prior to the incision, (2) at the end of surgery with (3) four additional doses administered postoperatively in the ICU every 6 hours for the first 24 hours. The first ICU dose will be administered 6 hours after the arrival in the ICU. The intraoperative study drug doses will be administered by one of the investigators (SJ or PR) and postoperative ICU study drug doses will be administered by nursing staff.
Anesthesia and Pain Management All of the study subjects will undergo cardiac surgery under general anesthesia. The cardiac surgery procedures will include standard midline sternotomy and in cases involving coronary artery bypass grafting, harvesting of saphenous vein and internal mammary artery grafts. Cardiopulmonary bypass will be conducted using membrane oxygenation. All subjects will undergo standardized induction per current practice at our institution with lidocaine (1-1.5 mg/kg), fentanyl (5-10 μg/kg), etomidate (0.2 mg/kg) and either rocuronium (1 mg/kg) or vecuronium (0.1 mg/kg). Succinylcholine (1-1.5 mg/kg) will be used when rapid sequence induction is appropriate. The anesthesia will be maintained with isoflurane. The intraoperative analgesia will consist of fentanyl or hydromorphone boluses, which will be administered at the discretion of the anesthesia provider. Postoperative pain management will consist of nursing fentanyl boluses while patients were sedated and mechanically ventilated. Following extubation, patients will receive oral oxycodone (5 to 10 mg orally every 3 hours as needed) with intravenous hydromorphone or morphine boluses for breakthrough pain.
Opioid Consumption At the end of the surgery, subjects will be transferred to the ICU while intubated and sedated on a propofol infusion. The time of arrival to the ICU will be considered time 0 for the purpose of postoperative study drug dosing and further evaluations. The 24 hour opioid consumption will be obtained from the electronic medication administration record and will be expressed in morphine equivalents per standard opioid conversions.
Secondary Outcomes The secondary outcomes will include 48 hour postoperative opioid consumption. The ICU nursing staff will be trained to use pain diaries containing an eleven point (0 to 10) Numeric Rating Scale (NRS). Pain assessments at rest and with movement will be recorded by the nursing staff at 6, 12, 18 and 24 hours postoperatively while one of the investigators will record pain assessments at 48 hours postoperatively. At the same times postoperatively, patients will be also assessed for opioid related adverse effects including nausea and vomiting, pruritus, respiratory depression, dizziness and sedation using a numeric scale (0 = no adverse effects, 1 = mild adverse effects not requiring treatment, 2 = moderate adverse effects requiring treatment and 3 = severe adverse effects refractory to treatment).
In addition, the ICU nursing staff will be instructed to complete the "Extubation Criteria" checklist for each 2-hour time period until extubation and the "ICU Discharge Criteria" checklist for each 4-hour time period until discharge from the ICU. A short "Patient Satisfaction Survey" will be also completed during the 24 and 48 hour follow up visits. The extent to which subjects overall pain experience met their expectations will be converted to the Likert scale with the following values: not at all (1), a little (2), a fair amount (3), very much (4) and extremely well (5).
Wound hyperalgesia at 24 and 48 hours postoperatively will be measured using a method previously described by Stubhaug et al. In brief, the lateral right chest area will be first stimulated where no pain sensation is experienced using a 180 gram von Frey filament (# 6.45). The filament will be then moved toward the incision in 1 cm steps until a distinct change in perception (painful, sore, or sharp feeling) is reported and the remaining distance to the incision will be noted. If no change in sensation appears, stimulation will be stopped at 4 cm from the incision. Wound hyperalgesia will be determined by testing the right side of the chest along five horizontal lines vertically separated by 2 cm at right angles to the incision. Only the right side of the chest will be chosen to avoid confounding skin sensation changes from the left internal mammary artery harvesting. The distance (in cm) from the incision to the point where sensations changed will be measured 5 times and the average distance will be used as wound hyperalgesia.
Data and Statistical Analysis A power analysis determined that 35 patients in each group will be needed to have an 80% chance to detect a 30% difference in mean 24 hour cumulative opioid consumption for the first 24 hour period after surgery using 2-sided t-tests with a set α level of 0.05, and an anticipated dropout rate of less than 15%.
Fisher's exact or Chi-squared and t-tests will be used to compare patient characteristics and intraoperative variables between the placebo and treatment group. Outcome variables will be compared between groups using t-tests and analysis of covariance controlling for age, sex, body mass index (BMI) and intraoperative hydromorphone. Scatter plots will be used to display the data to look for interactions, outliers and skewed distributions. The best fit line will be assessed by the F-test on the interaction between age and treatment group in an analysis of covariance. The analyses will be performed using SPSS software, version 15 (SPSS Inc., Chicago, IL). All reported P values will be 2-sided, and a significance level of P \< 0.05 will be used.
