This research was approved by the Medical Ethics Committee of the First Affiliated Hospital of Wenzhou Medical University and registered with the Clinical Trial Registration Center of China (ChiCTR2200061468, 25/06/2022). Patients who were 18 to 80 years old, had American Society of Anesthesiologists (ASA) physical status I–III, and were scheduled for an ERCP procedure between June 2022 and August 2022 were eligible to be enrolled in this prospective, randomized, single-blind study. Exclusion criteria included ASA physical status IV–V, refusal to participate, pregnant or breast-feeding patients, a history of allergy to study medication, and long-term sedative or narcotic analgesic drug abuse. All patients provided written informed consent.

Patients fasted for a minimum of 6 h prior to ERCP. After intravenous access was obtained, an infusion of 500 mL of Ringer’s solution was initiated at a rate of 250 mL h−1. Patients were positioned in a semi-prone position, and 2 L min−1 of oxygen was administered through a nasal cannula. As premedication, patients were given laryngopharynx topical anesthesia with 2% lidocaine hydrochloride mucilage and 1% dyclonine hydrochloride mucilage, an intravenous dose of 0.2 μg kg−1 sufentanil, and 0.02 mg kg−1 midazolam. Additionally, the DEX group also received an initial bolus of 0.6 μg kg−1 dexmedetomidine over 2 min, followed by a dexmedetomidine infusion at 1.2 μg kg−1 h−1. The PRO group received an initial bolus of 1–2 mg kg−1 propofol over 30 s followed by a propofol infusion at 2–3 mg kg−1 h−1. We targeted a sedation level on the Ramsay Sedation Scale (RSS, Table 1) of ≥ 4. In case of RSS < 4 or intolerance to the procedure, the patients were administered 0.1 μg kg−1 sufentanil and 0.01 mg kg−1 midazolam as rescue drugs. Continuous infusion of propofol or dexmedetomidine were withheld if BIS values < 45.

Age, gender, and body mass index were measured prior to surgery as baseline demographic data. The sedation level was evaluated using the RSS for clinical scoring and the bispectral index (BIS) as an objective tool. During the procedure, heart rate (HR), oxygen saturation (SpO2), respiratory rate (RR), mean blood pressure (MAP), RSS, and BIS levels were monitored and recorded at the following time points: 5 min before sedation (baseline, T0), 5 min after sedation (T1), 0, 5, 10, 15, and 20 min after starting ERCP (T2–T6), and post-procedure 0, 5, and 10 min (T7–T9). Patients were observed in the recovery unit for at least 30 min after the procedure. The recovery status was evaluated using the modified Aldrete Score [13]. Patients were ready for discharge when this score reached at least 9 without significant adverse effects such as nausea and dizziness.

Time to achieve RSS ≥ 4 was recorded as the onset time of targeted sedation. Recovery time was measured from the conclusion of ERCP until a modified Aldrete score of 9 was attained. Rescue drug injections were recorded. Immediately after the procedure, endoscopists were requested to assess their level of satisfaction: 1) satisfied, 2) moderately dissatisfied, 3) severely dissatisfied, and 4) unbearable. On the day following the procedure, the satisfaction score, pain, and procedure memory of the patients were evaluated by an anesthesia resident as follows: Satisfaction score: 1) comfortable, 2) mild discomfort, 3) severe discomfort, and 4) unbearable; there were four levels of pain: 1) no pain, 2) mild pain, 3) moderate pain, and 4) severe pain; Procedure memory: 1) I do not remember any part of the procedure; 2) I remember some parts of the procedure; and 3) I remember the entire procedure.

In terms of sedation regimen safety, hemodynamic and respiratory variables such as MAP, HR, SpO2, and RR were compared between groups. Sedation-related adverse events (SRAEs) and ERCP-related adverse events were among the perioperative adverse events. SRAEs were defined as hypoxemia (SpO2 < 90% for > 10 s), respiratory depression (RR < 10 bpm), hypotension (MAP < 65 mmHg), and bradycardia (HR < 45 beats min−1). Any sign of respiratory depression prompted an intervention which consisted of: 1) patient stimulation; 2) withholding medication; 3) a chin lift or jaw thrust maneuver; 4) increasing oxygen supplementation. If hypoxemia was observed, supplemental oxygen was administered through the nasopharyngeal airway. Intravenous saline or vasoactive agents (ephedrine or atropine) were used to treat hypotension or bradycardia. Post-ERCP pancreatitis (PEP), bleeding, perforation, and infection were categorized as ERCP-related adverse events. Intubation of the trachea and/or mechanical ventilation were considered severe adverse events.

The primary outcome was the incidence of hypoxemia. The secondary outcomes included: (1) onset time of targeted sedation, recovery time, and rescue drug injections; (2) the satisfaction scores of endoscopists and patients, as well as the pain and procedure memory scores; (3) RSS, BIS, MAP, HR, SpO2, and RR at T0–T9; (4) adverse events recorded, which included respiratory depression, hypotension, bradycardia, and ERCP-related adverse events.

We hypothesized that sedation with dexmedetomidine for ERCP would reduce the incidence of hypoxemia by at least 50% when compared to propofol. According to a previous study by Jokelainen et al. [14], 60.7% of patients who underwent ERCP under propofol-based sedation experienced hypoxemia. To achieve a power ≥ 0.80 and an α level = 0.05, a sample size of 17 patients per group was required for the chi-squared test with Fisher’s exact test. The final sample size was set at 19 patients per group, assuming a 10% dropout rate.

The SPSS software version 26.0 (IBM, Armonk, NY, USA) was used to conduct statistical analyses. An independent sample t-test and nonparametric tests were used for continuous variables. Pearson’s chi-squared and Fisher’s exact tests were applied to variables that were expressed as numbers and/or percentages of the total for categorical variables. Continuous variables with normal distributions are presented as the mean ± standard deviation. The Kolmogorov–Smirnov test was used to compare non-normally distributed data, which are presented as the median and interquartile range (IQR). A P value < 0.05 was considered statistically significant.