Effect of butorphanol on visceral pain in patients undergoing gastrointestinal endoscopy: a randomized controlled trial

This dual-center, randomized, placebo-controlled study was conducted at Beijing Tiantan Hospital, Capital Medical University, and Beijing Daxing People’s Hospital between August 14th, 2020, and September 30th, 2021. This study was approved by the China Ethics Committee of Registering Clinical Trials (Registration number: ChiECRCT20200200) and registered at Clinicaltrials.gov (NCT04477733, 20/07/2020)) in July 2020. All patients or their legal representatives provided written informed consent. The study followed the Consolidated Standards of Reporting Trials (CONSORT) guidelines.

We recruited adult patients aged 18 to 65 with an American Society of Anesthesiologists physical status from I to III who underwent colonoscopy or gastroscopy. The exclusion criteria included patients with a body mass index > 30 kg/m², a history of depression, opioid dependence, poorly controlled hypertension (systolic blood pressure > 180 mmHg), myocardial infarction, severe liver disease, and significant abdominal pain before surgery; patients with sensory system or language dysfunctions who could not cooperate to complete the scale; and pregnant women.

Patients were randomly assigned to receive butorphanol (Group I) or normal saline (Group II) in a 1:1 ratio based on computer-generated stratified randomization numbers. The random numbers were sealed in separate opaque envelopes until the analysis was complete. Patients in Group I received 10 μg/kg butorphanol intravenously 3 min before the intravenous injection of propofol. The patients assigned to Group II received an identical volume of normal saline at the same infusion rate.

The investigators did not participate in other processes while preparing the research solution. A designated staff prepared the solution, enclosed in a dark syringe (volume: 5 ml) labeled "study solution," The two solutions seemed identical. Randomization was blinded to the participants, chief anesthesiologists, and outcome assessors.

The regimens were standardized in both groups. Patients were deprived of water for 2 h and fasted for 8 h before surgery. Based on the requirements of anesthesia and surgery, venous access (central vein of the upper limb) was established. As perioperative fasting and bowel preparation are believed to cause intravascular hypovolemia, preemptive intravenous fluid was infused with 10–15 ml/kg normal saline. Standard intraoperative monitoring, which included electrocardiogram (ECG), heart rate (HR), blood pressure (BP), respiratory rate (RR), and pulse oxygen saturation (SpO₂), was applied. Preinduction medication was not administered. The patient received preoxygenation with a mask filled with 100% oxygen (4–6 L/min) for at least 3–5 min before induction. The induction of anesthesia was performed with 1.5–2 mg/kg propofol as a bolus infused slowly (60 to 120 s) until the eyelash reflex disappeared. When the vital signs were stable, an endoscopic examination was started. If there was a body movement reaction during the examination, a dose of 0.5–1 mg/kg propofol was added. Moreover, the mean arterial pressure (MAP) and HR were maintained within ± 20% of the values before anesthesia induction during the procedure with vasoactive drugs such as dopamine and atropine. After the procedure, the patients were transferred to the postanesthetic care unit (PACU) and followed up regularly for any adverse events for at least 30 min. If the pain persisted after 30 min, regardless of the patient’s assigned group (Group I or Group II), a single dose (0.01 μg/kg) of sufentanil was given, and the patients could leave the PACU after the pain was relieved.

Baseline data were recorded, such as age, sex, height, weight, body mass index, ASA physical status, and type of operation. The MAP, HR, RR, and SpO₂ of the patients were recorded at seven time points: T0, at admission; T1, before anesthesia; T2, 5 min after propofol administration; T3, at the end of the operation; T4, after 5 min in the PACU; T5, after 10 min in the PACU; and T6, after 30 min in the PACU.

The primary outcome was the incidence of visceral pain after the procedure. Because there is no universal scale for quantifying visceral pain, we defined a visual analog scale (VAS) score of ≥ 1 10 min after recovery as visceral pain. The secondary outcomes were visceral pain at 20 and 30 min after recovery, propofol consumption, the incidence of injection pain caused by propofol, episodes of hypotension (defined as an MAP less than 60 mmHg or 30% of the values before the induction of anesthesia) or bradycardia (defined as an HR less than 60 bpm), the operation time, the recovery time, and the adverse events at 24 h after recovery, such as fatigue, nausea or vomiting, abdominal bloating, dizziness or headache, hypoxemia (blood oxygen saturation below 90% for more than a minute or requiring any airway intervention), and involuntary body movement.

We used PASS software to calculate the necessary sample size for this study. A cohort study reported that 45% of participants (124 of 277) undergoing colonoscopy and gastroscopy complained of pain during follow-up [16]. The incidence of pain after colonoscopy was higher in two other randomized controlled studies, at 47.8% and 51%, respectively [17, 18]. In our pretrial test, the incidence of visceral pain after colonoscopy or gastroscopy was 54%. Combining all the above data, we estimated that the incidence of pain in Group II was 50%, with a 20% reduction in Group I. Thus, 186 people were included in this trial (power = 80%, and α = 0.05). The ratio of the two sets of samples was 1:1. Considering an overall withdrawal rate of 10%, the sample size was estimated to be 206 patients (103 patients per group).

The data were analyzed using SPSS 26.0, and the figures were created by GraphPad Prism 9.0. All analyses were based on the intention-to-treat (ITT) principle. The Kolmogorov–Smirnov test was used to analyze continuous outcomes to judge the normality of their distributions. Normally distributed continuous variables were summarized as the mean value ± standard deviation and were compared using independent t tests. Skewed continuous variables were summarized as the median value and interquartile range and were compared using the Mann–Whitney U test. As appropriate, categorical variables were summarized as the number and percentage and compared using the chi-square or Fisher’s exact test. The primary endpoint was the incidence of visceral pain in the recovery room, and the chi-square test was used to compare the differences between the two groups. The risk ratios and 95% confidence intervals were reported for the primary and secondary outcomes. MAP, HR, RR, and SpO₂ were compared four times between the groups at T0, T1, T2, and T3 using a two-sample Student’s t test. Bonferroni correction was used to justify the P values for these three variables, and an α level of 0.0125 was considered statistically significant. Moreover, an α level of 0.05 was considered statistically significant for the remaining variables.

In the ITT population, 206 patients were consecutively enrolled at two hospitals between August 14th, 2020, and September 30th, 2021. Eventually, 203 patients were randomly assigned to Group I (n = 102) or Group II (n = 101). At the primary time point, nine patients were lost to follow-up. A total of 194 patients in the two groups completed the study (Fig. 1). The data from all 194 patients were included in the analysis. The demographics, baseline assessment, and intraoperative details were similar between the two groups and are presented in Table 1.

The incidence of visceral pain at 10 min after recovery was significantly lower with butorphanol (31.5% vs. 68.5%, respectively; RR: 2.738, 95% CI [1.409–5.319], χ2 = 9.157, P = 0.002; see Table 2). The significant difference in Fig. 2 and Table 2 shows the pain level and distribution of visceral pain (P = 0.006).

The incidence of visceral pain at 20 and 30 min after recovery showed a similar change (Fig. 2, Table 2). Butorphanol infusion allowed for an apparent reduction in propofol consumption (200 vs. 250, respectively; Z = -2.720, P = 0.007), which was more pronounced in patients undergoing colonoscopy than in those undergoing gastro-colonoscopy (Z = -6.999, P < 0.001; Online Resource 1). This may be because of the prolonged operation time in gastro-colonoscopy (Z = -3.095, P = 0.002; Online Resource 2). However, there were no significant differences in episodes of bradycardia or hypotension, operation time, recovery time, or incidence of injection pain between the two groups (P > 0.05, Table 2).

Perioperative monitoring parameters (MAP, SpO2 and RR).

The MAP, SpO_2, and RR were not significantly different between the two groups of patients at the seven time points. Only after 5 min of propofol administration (T2, t = -2.716, P = 0.007) and at the end of the operation (T3, t = -2.261, P = 0.025) was the HR of Group I slightly lower than that of Group II, and the other time points were significantly different (Fig. 3).

Adverse events.

Only one patient in the butorphanol group developed fatigue during the operation and recovery room. After 24 h of recovery in the butorphanol group, the number of people with nausea or vomiting, abdominal bloating, dizziness, or headache was 2, 7, and 5, respectively. In Group II, only seven people had abdominal pain and bloating. No cases of hypoxemia or body movements were reported. There were no significant differences between the two groups in adverse events (χ² = 4.345, P = 0.182).