We conducted a single-center, prospective, randomized, double-blind study from August 2015 to December 2017 at Seoul National University Dental Hospital, Republic of Korea. Permission to conduct this study was granted by the institutional review board of Seoul National University Dental Hospital (approval number, CME15001). Written informed consent was obtained from all participants. All aspects of participant privacy and confidentiality were preserved. The investigation was registered with the Clinical Research Information Service (https://cris.nih.go.kr/KCT0001592) and performed according to the guidelines for the proper conduct of medical research on human participants.

Patients undergoing bimaxillary orthognathic surgery with intravenous PCA for postoperative pain control were enrolled. All patients had an American Society of Anesthesiologists (ASA) patient status of I or II and were between 20 and 40 years of age. Exclusion criteria were as follows: (1) urgent or emergent case, (2) re-do case, (3) ASA status of at least III, (4) allergic to fentanyl or nefopam, (5) history of drug abuse, (6) chronic pain (≥ 3 months), (7) use of analgesic or hypnotic medication within 2 weeks, (8) hepatic, renal, or cardiac insufficiency, (9) pulmonary diseases such as chronic obstructive pulmonary disease, asthma, and upper respiratory infection within 2 weeks, (10) smokers, (11) pregnant or breastfeeding, (12) refused to participate, (13) diabetes or neuropathic diseases, (14) unable to use the PCA device, (15) others who the investigator judged to be inappropriate candidates for participation in the clinical study. All patients received a general explanation of the study process, including instruction in the use of the visual analogue scale (VAS) ranging from 0 (no pain) to 10 (worst pain imaginable). The patients were also carefully instructed in use of the Accumate® 1100 PCA device (Wooyoung Medical Co. Ltd., Jincheon, Korea) that was used in the study.

The patients were randomly assigned to one of two groups using a computer-generated random number table (available at http://www.randomization.com/). The patients were allocated to receive either nefopam-based (nonopioid; N group) or fentanyl-based (F group) PCA for postoperative pain control using sealed envelopes containing the treatment options (N and F) before induction of anesthesia. All anesthesiologists, surgeons, nurses, and patients were blinded to the study. All data were collected by trained observers who also were blinded and did not participate in patient care.

All patients arrived at the operation room without premedication and were not given any preanesthetic medications. Anesthesia was induced after establishing routine patient monitoring (pulse oximetry, electrocardiography, and noninvasive blood pressure monitoring) and bispectral index monitoring. Following preoxygenation, anesthesia was induced with a 5 µg/mL effect site concentration of propofol and 5 ng/mL of remifentanil using an Orchestra Base Primea target-controlled infusion system (Fresinius Kabi, Bad Homburg, Germany). Anesthesia was maintained with propofol (2–4 µg/mL) and remifentanil (3–10 ng/mL) depending on bispectral index level (bispectral index was maintained between 40–60) and vital signs. Rocuronium (0.6 mg/kg) was administered for muscle relaxation following anesthesia induction and additional doses were administered as necessary (self-respiration is restored or the operator wished). After nasotracheal intubation, patients were ventilated with 50% oxygen in air. The tidal volume was 6–8 mL/kg (lean body mass), and positive end expiratory pressure was utilized as necessary. Respiratory rate was adjusted to maintain end-tidal carbon dioxide partial pressure from 30 to 35 mmHg. Invasive blood pressure monitoring and arterial blood gas analysis were performed by placing the catheter on the dorsalis pedis artery. Upon completion of the operation, residual neuromuscular paralysis was reversed using sugammadex (200 mg) and the PCA (Group N, 120 mg nefopam; Group F, 700 mcg fentanyl in total normal saline 120 mL) machine was connected. After recovery of self-respiration, patients were transferred to a postanesthetic care unit. After confirmation of complete recovery of consciousness, self-respiration (with sustained spontaneous respiration rate > 12/min), and an open airway, the tracheal tube was removed.

Twenty milligrams of nefopam provides an analgesic effect equivalent to 6–12 mg morphine, and 10 mg of morphine provides equivalent analgesia to 100 µg fentanyl [21]. Therefore, PCA was set as follows: 120 mg of nefopam for the N group and 700 µg of fentanyl for the F group in total volumes of 120 mL with normal saline. Continuous basal infusion of nefopam or fentanyl was provided (1 mL/hour). The PCA machine allowed for a 1 mL bolus, with a 15 minutes lockout time, and a 5 mL maximum per hour. Drugs were infused using an Accumate 1100®. The PCA device time was synchronized with Korea Standard Time at the national metrology institute. All labels on the PCA devices were hidden so that patients and medical staff who had direct contact with the patients did not know which drug was being administered. If postoperative pain was not controlled by continuous PCA infusion, ketorolac was administered intravenously as a rescue analgesic. If PONV occurred, antiemetics were administered.

The primary outcome was the incidence and number of PONV events after surgery. PONV was assessed using a 4-point ordinal scale (0 = none, 1 = nausea, 2 = retching, 3 = vomiting) upon arrival at the Post-anesthesia care unit 0.5, 1, 8, 24, 48, and 72 hours after surgery. Nausea was defined as a subjectively unpleasant sensation associated with awareness of the urge to vomit. Retching was defined as labored, spasmodic, rhythmic contraction of the respiratory muscles without expulsion of gastric contents. Vomiting was defined as the forceful expulsion of gastric contents from the mouth.

The secondary outcomes were pain score, sedation, shivering, difficulty sleeping, and other unwanted symptoms. Pain was assessed using the visual analogue scale (VAS) ranging from 0 (no pain) to 10 (worst pain imaginable) at 0.5, 1, 8, 24, 48, and 72 hours after surgery, both at rest (VASr) and during movement (VASm). Total PCA volume, need for rescue analgesics, and dose of rescue analgesics and antiemetic injections were recorded at 0.5, 1, 8, 24, 48, and 72 hours after surgery. The PCA discontinued if the respiratory rate was < 12/min, oxygen saturation was < 95%, sedation score was ≥ 4, or when other side effects caused by the PCA were suspected. If use of the PCA was discontinued, the reason for discontinuation, and the time at which discontinuation occurred was noted by medical personnel who did not have information about this study. The occurrence of side effects was assessed by checking for sedation, shivering, difficulty in sleeping, constipation, hypotension, and other unwanted symptoms for 72 hours postoperatively. Degree of sedation was scored as follows using the Ramsay sedation assessment scale [22]: 1 = anxious, agitated, or both; 2 = cooperative, oriented, and tranquil; 3 = responds to commands only; 4 = a brisk response to a light glabellar tap; 5 = a sluggish response to a light glabellar tap; 6 = no response. Patient satisfaction scores were also evaluated using a 5-point Likert scale. The Likert scale was scored as follows: 1 = not at all satisfied; 2 = slightly satisfied; 3 = moderately satisfied; 4 = very satisfied; 5 = completely satisfied. All outcomes were collected and analyzed by a biostatistician who was not a participant in the study and only knew details of study design.

After the IV PCA device was collected, information regarding patient self-administration (the time of starting the device, the number of bolus attempts, the amount of bolus drug delivery, and the end of device usage) was downloaded to personal computers loaded with an IV PCA device-specific download program, Eventlog Viewer version 1.1 for Accumate1100 (Wooyoung medical, Inc., Seoul, Korea).

To determine appropriate sample sizes, we assumed that incidences of PONV in the F group and the N group of 25% and 10%, respectively, would be significant on the basis of previous publication [16]. Seventy-nine patients per group were required for a power of 80% with a type I error rate of 5%. Considering a drop-out rate of 10%, 174 patients were planned to be included. However, interim analysis after including 89 patients showed a high incidence of PONV in the N group, and the study was stopped.

Measured values were presented as mean ± SD or as a number (%). Intra-group distributions of variables were evaluated for normality using the Kolmogorov-Smirnov and histogram tests. Fisher's exact test or χ² test was used to evaluate categorical variables and Student's t-test or Mann-Whitney U test were used to evaluate continuous variables. P < 0.05 was considered statistically significant. SPSS version 25.0 for Windows was used for statistical analyses.