This was a multicenter, randomized, double-blinded, parallel-group, fentanyl-controlled non-inferiority phase 3 study in patients requiring respiratory management in intensive care to evaluate the efficacy, safety, and pharmacokinetics of continuous intravenous administration of remifentanil (using MR13A11A as the investigational drug). The study was conducted at 29 sites in Japan from December 2019 to December 2020 (Additional file 1: Table S1).

The study protocol and the informed consent form were approved by the Ethics Review Board of Hamamatsu University School of Medicine, Hamamatsu, Japan, (Approval No: 709) and at each hospital listed in Additional file 1: Table S1. All patients gave written informed consent before initiation of any study-specific procedures. The study was conducted in accordance with the ethical principles originating in or derived from the Declaration of Helsinki and Good Clinical Practice guidelines. The study was designed and conducted by the sponsor in collaboration with the principal investigators. Maruishi Pharmaceutical Co., Ltd. monitored study conduct, collected the data, and performed the statistical analyses.

Inclusion criteria were male or female patients aged ≥ 20 years who required respiratory management due to intubation or tracheostomy for 6 h to 10 days in intensive care and who were anticipated to require pain relief. The patients entering the ICU after surgery were required to have an American Society of Anesthesiologist Physical Status (ASA) status of I–III at the pre-operative assessment. If a patient was a female of childbearing potential, she was excluded if pregnant, possibly pregnant, or lactating. Approximately 15% of the patients who were recruited from ICU had internal medicine as much as possible. Key exclusion criteria were patients with severe damage to the central nervous system, with a neurological disease that made pain/sedation assessment difficult as adjusted by the investigator, who had received nalmefene within one week prior to the study drug administration, who required local anesthetic, epidural or intrathecal administration of analgesics, or nerve block, those with contraindications to muscle relaxants, who were likely to suffer from respiratory depression such as coma due to head injury or brain tumor, those with a history of seizure or asthma, and patients who were assessed as suffering from severe illness, that might cause death within 24 h, and those with massive bleeding who were being considered for reoperation.

Patients who met the eligibility criteria were randomly assigned in a 1:1 ratio to receive either remifentanil or fentanyl. Treatment allocation of patients was initiated via an envelope method. Remifentanil and fentanyl can be distinguished by the appearance of their vial; therefore, to ensure masking was maintained for the patients, investigators, site staff, assessors, and sponsor, only unblinded persons appointed by the investigator prepared the blinded administration solution. The results of PK measurement were not reported to the investigators and sponsor before the primary analysis.

After entering ICU, pre-dose screening was performed under light sedation using sedatives as needed to maintain a Richmond Agitation-Sedation Scale (RASS) [16, 17] of -2 to 0. As a run-in treatment, the remifentanil group received placebo (physiological saline), and the fentanyl group received 1 to 2 μg/kg of fentanyl by slow intravenous bolus injection. The study drug was administered by continuous intravenous infusion during respiratory management under intubation or tracheostomy in the ICU. The duration of administration of the study drug was ≥ 6 h to ≤ 10 days. Dose titration was started from an infusion rate of 1 mL/h (remifentanil: 0.025 µg/kg /min in the remifentanil group, fentanyl: 0.1 µg/kg/h in the fentanyl group), and then escalated by 1 mL/h (remifentanil: 0.025 µg /kg/min, fentanyl: 0.1 µg/kg/h) until the target level [behavioral pain scale (BPS) [16, 18] ≤ 5 or numerical rating scale (NRS) ≤ 3] was achieved. After an effective infusion rate was obtained, the rate was adjusted to maintain the target analgesia level. If the target analgesic level was expected to be achieved without hindrance, further increases were continued. If the appropriate analgesic level was not obtained even after reaching the maximum infusion rate, fentanyl was administered as a rescue analgesic (open-label setting). If the target analgesic level was not achieved even after increasing the infusion rate four times in a row during 20 to 30 min in the maintenance phase, rescue use of fentanyl was to be considered. The maximum infusion rate was 20 mL/h (remifentanil: 0.5 µg/kg/min in the remifentanil group, fentanyl: 2 µg/kg/h in the fentanyl group) in the titration and maintenance phases. The infusion rate of the study drug was gradually reduced toward weaning from the ventilator while observing the general condition of the patient. The infusion rate was reduced by up to 25% until the end of treatment, at intervals of at least 10 min, and weaning from the ventilator was carried out after the end of administration. If the infusion rate was ≤ 1 mL/h (remifentanil: ≤ 0.025 µg/kg/min in the remifentanil group, fentanyl: ≤ 0.1 µg/kg/h in the fentanyl group) at the start of infusion reduction, or if the infusion rate of ≤ 1 mL/h (remifentanil: ≤ 0.025 µg/kg/min in the remifentanil group, fentanyl: ≤ 0.1 µg/kg/h in the fentanyl group) was reached during the infusion reduction, the administration was terminated without further reduction. However, further infusion reduction was allowed depending on the condition of the patients. The study drug was gradually switched to other analgesics for pain management as the infusion rate decreased. A 24-h follow-up period was provided after the end of administration. A sedative was allowed to be used as needed. The target sedation level was RASS ≤ 0, with an intended level of -2 to 0, since analgesia could not be assessed at a RASS of ≤ -3. If both the analgesia and sedation levels did not meet each target level, the analgesia level was preferred. Temporary use of muscle relaxants was permitted as needed, but continuous use was prohibited. Systemic analgesics were prohibited during the treatment and observation periods. Use of nalmefene was prohibited for one week before titration, and during treatment. Epidural or intrathecal administration of local anesthetics or analgesics, and nerve block were prohibited from the start of dosing to the start of dose reduction with the intention of weaning from the ventilator.

Analgesia was assessed using BPS. However, if the BPS-based assessment was inappropriate because the patients was clearly conscious, the NRS was used. Analgesia was assessed at pre-dosing and at 5, 10, 15, 20, and 30 min, 1, 2, and every 2 h after the start of dosing, 5-min prior to termination of dosing, and 10, 20, 30 min and 1 h after the termination of dosing. When the infusion rate was changed, rescue fentanyl was administered or the sedative infusion rate was changed, analgesia was assessed within 5 min before the change or the start of dosing, and 5, and 15 min after the change or the start of dosing. Sedation was assessed with the RASS using the same schedule of analgesia assessment. The validity/accuracy of the RASS/BPS/NRS assessment was ensured by requiring all investigators, subinvestigators and nurses to take training prior to conducting any assessment related to this trial.

The primary endpoint was the proportion of the patients who did not require rescue administration of analgesia between the start of dosing and the start of dose-reduction toward weaning from the ventilator. The secondary endpoints were the number of rescue dose of fentanyl and the total dose of such doses, doses of sedatives used during treatment, proportion of time where BPS was maintained ≤ 5 or NRS was ≤ 3, proportion of time where RASS was maintained at ≤ 0, the proportion of time where RASS was between -2 and 0, duration from the end of dosing to weaning from the ventilator, duration from the start of dose-reduction to weaning from the ventilator, the infusion rate of the study drug (mean, minimal, and maximum) and duration of treatment.

Safety was assessed according to adverse events (AEs), laboratory tests including hematology, blood biochemistry, and urinalysis, vital signs, percutaneous oxygen saturation (SpO2), end-tidal carbon dioxide (ETCO2), blood gas analyses including partial pressure of arterial oxygen, partial pressure of arterial carbone dioxide, arterial oxygen saturation, HCO3−, and pH, body weight, and 12-lead electrocardiogram (ECG). Vital signs (blood pressure, heart rate and respiratory rate), SpO2 and ETCO2 were assessed at the same time points as the BPS/NRS up to 1 h after the termination of study drug dosing, and at 6 and 24 h after the end of dosing.

PK was assessed in 24 patients whose body mass index was < 25 in the remifentanil group. Arterial blood samples for pharmacokinetic analysis were collected 1 h after the start of dosing, immediately before the start of dose reduction leading into weaning from the ventilator, at the end of dosing, and 1, 2, 3, 5, 7, 10, 20, and 60 min after the end of dosing. Remifentanil concentrations in arterial blood were measured using a liquid chromatography–tandem mass spectrometry method [19]. The quantification limit was < 0.05 ng/mL.

The sample size was established based on the expected proportion of patients who would not require rescue fentanyl. We conservatively expected 85% of patients would not need a rescue analgesic in the remifentanil and fentanyl groups based on the reported efficacy proportion of ≥ 95% [8]. Placebo effect was assumed to be 50% based on the proportions of non-rescue analgesics in the clinical studies of sedatives [20,21,22,23]. The non-inferiority margin for this study was set at 15%. To show non-inferiority of remifentanil to fentanyl under these conditions, with a one-sided alpha level of 2.5% and 80% of power, each treatment group should consist of 90 participants. In addition, the number of participants needed to verify that the effective proportion of remifentanil exceeded a threshold of 70% with the expected efficacy of 85%, significance level 5% on both sides, and power of 80% was 65. Therefore, this study planned to enroll 90 patients/group (total of 180 patients). For PK assessment, 20 participants were selected in order to calculate the PK parameters.

Safety analysis was performed for subjects in the safety analysis set (SS), and efficacy was analyzed primarily for the full analysis set (FAS) and secondarily in the per protocol set (PPS). The SS consisted of subjects who received the study drug at least once excluding those who were good clinical practice (GCP) non-compliant. The FAS included the SS subjects excluding those who had no primary data-assessment endpoint. The PPS was defined as the subset of subjects in the FAS excluding those with deviations from the protocol determined to affect efficacy assessment.

For the primary endpoint, the proportion of patients who did not require rescue administration of analgesia and its 95% confidence interval (CI) were calculated in each group. In the remifentanil group, the lower limit of the 95% CI was compared with the threshold of 70%. The difference and its 95% CI (Wald method) in the effective proportions between the groups was calculated. Non-inferiority of remifentanil to fentanyl was verified by the Δ-addition method (non-inferiority margin Δ = 15%). For sensitivity analysis, the primary endpoint was evaluated in the PPS population. The primary endpoint was also evaluated in the subpopulation that included sex (male and female) and age (< 65 years and ≥ 65 years), type of ICU patients (internal medicine patients and post-operative ones), and concomitant sedation (dexmedetomidine and propofol). For secondary endpoints, descriptive statistics were shown for each group and the 95% CI of the mean was calculated. In addition, the difference in the mean and its 95% CI were calculated, and compared between the groups using t-test. The number of doses in the remifentanil group was compared with that in the fentanyl group by the Wilcoxon rank sum test.

For safety assessment, events that occurred after administration of the study drug were analyzed. The number of AEs or adverse drug reactions (ADRs), defined as AEs other than those for which a causal relationship was not ruled out, and number of subjects who experienced AEs were shown by group. The difference in AEs and ADRs between the groups were analyzed by Fisher’s exact test. AEs were coded using the Medical Dictionary for Regulatory Activities Japanese edition ver. 23.1.

PK assessment, descriptive statistics including number of subjects, mean, standard deviation (SD) for remifentanil concentration in arterial blood were shown for each measurement time. Descriptive statistics including number of subjects, mean, median, geometric mean, SD, coefficient of variation (CV), minimum, and maximum of the PK parameters [non-compartment model: elimination half-life (t1/2), area under the concentration–time curve from time zero to time t (AUC0-t), AUC from time zero to infinity (AUC0-inf), maximum concentration (Cmax), total body clearance (CL), distribution volume at steady state (Vss)] were calculated.

For missing data, no data complementation was performed related to efficacy, safety and pharmacokinetic analysis. All statistical tests were performed using a two-sided significance level of 0.05. All analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).