Background: Opioid-induced respiratory depression driven by ligand binding to mu-opioid receptors is a leading cause of opioid-related fatalities. Buprenorphine, a partial agonist, binds with high affinity to mu-opioid receptors but displays partial respiratory depression effects. The authors examined whether sustained buprenorphine plasma concentrations similar to those achieved with some extended-release injections used to treat opioid use disorder could reduce the frequency and magnitude of fentanyl-induced respiratory depression.

Methods: In this two-period crossover, single-centre study, 14 healthy volunteers (single-blind, randomized) and eight opioid-tolerant (OT) patients taking daily opioid doses ≥90 mg oral morphine equivalents (open-label) received continuous intravenous buprenorphine or placebo for 360 minutes, targeting buprenorphine plasma concentrations of 0.2 or 0.5 ng/mL in healthy volunteers and 1.0, 2.0 or 5.0 ng/mL in OT patients. Upon reaching target concentrations, participants received up to four escalating intravenous doses of fentanyl. The primary endpoint was change in isohypercapnic minute ventilation (V_E). Additionally, occurrence of apnea was recorded.

Results: Fentanyl-induced changes in V_E were smaller at higher buprenorphine plasma concentrations. In healthy volunteers, at target buprenorphine concentration of 0.5 ng/mL, the first and second fentanyl boluses reduced V_E by [LSmean (95% CI)] 26% (13-40%) and 47% (37-59%) compared to 51% (38-64%) and 79% (69-89%) during placebo infusion (p=0.001 and <.001, respectively). Discontinuations for apnea limited treatment comparisons beyond the second fentanyl injection. In OT patients, fentanyl reduced V_E up to 49% (21-76%) during buprenorphine infusion (all concentration groups combined) versus up to 100% (68-132%) during placebo infusion (p=0.006). In OT patients, the risk of experiencing apnea requiring verbal stimulation following fentanyl boluses was lower with buprenorphine than with placebo (odds ratio: 0.07; 95% CI: 0.0 to 0.3; p=0.001).

Interpretation: Results from this proof-of-principle study provide the first clinical evidence that high sustained plasma concentrations of buprenorphine may protect against respiratory depression induced by potent opioids like fentanyl.

This two-part, placebo-controlled crossover study was conducted in the Leiden University Medical Centre, The Netherlands, from March 2018 until January 2019, in accordance with the principles of the Declaration of Helsinki, the International Conference on Harmonisation Good Clinical Practice (ICH GCP), and ethical principles as referenced in EU Directive 2001/20/EC. The protocol (EudraCT 2017‐004858‐42) was approved by the Medical Review and Ethics Committee of the BEBO foundation (Assen, The Netherlands). This study is registered with trialregister.nl, number NL7028 (https://www.trialregister.nl/trial/7028). Due to administrative reasons, trial registry was completed in May 2018, after participant recruitment began.

Study population

The study enrolled healthy volunteers (Part A) and opioid-tolerant (OT) patients (Part B). All participants provided written informed consent prior to any study-related procedure. In Part A, male and female healthy volunteers, aged 18 to 45 years with a body mass index of 18 to 30 kg/m² who had no history of substance use disorder, were eligible. Exclusion criteria included history of any clinically relevant medical, psychiatric, or neurologic condition; positive pregnancy test; current substance use disorder according to the criteria of the Diagnostic and Statistical Manual of Mental Disorders, 5th edition; [22] smoking or having smoked in the last 6 months; alcohol consumption >20 units/week (men) or >13 units/week (women); use of any medication within 14 days or 5 half-lives before dosing; opioid use (including opioid antagonists) within 30 days before dosing; use of medication that induces/inhibits relevant cytochrome P450 enzymes; history of suicidal ideation within 30 days or suicide attempt within 6 months prior to informed consent; or any other condition that, in the opinion of the investigators, could interfere with the ability to participate in the study.

For Part B, male and female OT patients, aged 18 to 55 years, with a body mass index of 18 to 32 kg/m² using daily doses of opioids ≥ 90 mg oral morphine equivalents [23], and who were in stable condition based on their medical evaluation were eligible. All exclusion criteria were similar to Part A, except for modified alcohol consumption limits to >27 units/week (men) or >20 units/week (women); broadened nicotine permissions to no smoking on dosing days; and, specifically no use of buprenorphine within 10 days of the first study drug administration.

All eligibility criteria are provided in the study protocol, which is available as supporting information (S1 File). Medical screening was completed within 30 days of first study drug administration. Participants were recruited through national advertisements, out-patient clinics with expertise in the treatment of pain, and in collaboration with specialized opioid-abuse treatment clinics.

Design, randomisation and masking

Both Parts A and B included two study periods, during which participants received continuous intravenous infusion of buprenorphine or placebo co-administered with up to four escalating fentanyl doses. Healthy volunteers in Part A were randomly assigned to one of two treatment sequences and were single blinded to the assigned treatments (buprenorphine or placebo) with masking of infusion syringes. The randomisation schedule was generated by an independent statistician using SAS version 9.4. Part B was an open-label, single-sequence crossover study where subjects received placebo treatment and then buprenorphine treatment. Dose group allocation was performed by the investigators within dose ranges specified per protocol and based on emerging data to explore exposure-response effects.

Drugs and study procedures

Healthy volunteers in Part A were admitted the day prior to the experiment for each study period, with a washout of two weeks between periods. OT patients in Part B were admitted to the clinic 2-5 days before the first study period. To ensure washout of each patient’s usual opioids, tailored substitution schedules with oxycodone began a minimum of 48 hours before the first experiment, and the last dose of oxycodone was administered at least 15 hours before study drug administration. Because tolerance to opioid effects is poorly characterized in patients receiving long-term opioids, OT patients received placebo plus fentanyl challenges in Period 1 to optimise the fentanyl dose escalation before buprenorphine and fentanyl were co-administered in Period 2. Due to the short half-life of fentanyl, study periods were separated by at least 40 hours. During this washout period, patients again received oxycodone for opioid substitution.

On the morning of each study period, an intravenous line was placed for administration of study medication and an arterial line was placed for blood sampling in the opposite arm. Isohypercapnic ventilation was measured for approximately 6 hours using the dynamic end-tidal forcing technique, as described elsewhere [20, 21], allowing the investigator to direct ventilation towards pre-defined end-tidal PCO₂ (7 kPa) and end-tidal PO₂ (14.5 kPa) values. A combination of oxygen, carbon dioxide, and nitrogen was delivered to the participants through a face mask and inspired minute ventilation was measured by pneumotachography. A finger probe with pulse oximeter was used for continuous surveillance of arterial oxygen saturation (SpO₂). Ventilation data were captured as one-minute breath-to-breath averages.  

Intravenous infusion with buprenorphine (Indivior UK Ltd., UK) or placebo started once baseline minute ventilation (V_E) had stabilized at 20 ± 2 L/min (about 4-fold above normal resting V_E). In healthy volunteers, infusion rates of 0.02 mg/70 kg/h and 0.05 mg/70 kg/h buprenorphine were selected to target plasma concentrations of 0.2 ng/mL and 0.5 ng/mL, respectively. In OT patients, 10-fold higher buprenorphine infusion rates were administered: 0.1, 0.2 or 0.5 mg/70 kg/h targeting plasma concentrations of 1.0, 2.0 or 5.0 ng/mL, respectively. In both healthy volunteers and OT patients, a 10-fold higher infusion rate was used over the first 15 minutes to speed attainment of steady-state buprenorphine concentrations at the site of action. In order to manage possible gastrointestinal side effects, all participants received 4 mg of ondansetron prior to infusion.

At 120, 180, 240, and 300 minutes after the start of the buprenorphine or placebo infusion, escalating intravenous fentanyl doses (Hameln Pharmaceuticals Ltd., UK) were administered over 90 seconds. The planned fentanyl doses in healthy volunteers were 0.075 mg/70 kg, 0.15 mg/70 kg, 0.25 mg/70 kg and 0.35 mg/70 kg. In OT patients, the planned fentanyl doses were 0.25 mg/70 kg, 0.35 mg/70 kg, 0.50 mg/70 kg and 0.70 mg/70 kg.

Pharmacokinetic sampling

Arterial blood samples for analysis of buprenorphine and fentanyl plasma concentrations were collected at multiple timepoints over 540 minutes after the start of buprenorphine or placebo infusion (Figs 3 and 4). Buprenorphine and fentanyl plasma concentrations were assessed using liquid chromatography with tandem mass spectrometry (LC-MS/MS) methods validated over a range of 0.02 to 10.0 ng/mL for buprenorphine and 0.1 to 50.0 ng/mL for fentanyl.

Pharmacodynamic and pharmacokinetic outcomes

The primary study endpoint was maximum decrease in minute ventilation, defined as the minimum value of isohypercapnic V_E observed during each fentanyl dosing period compared to pre-fentanyl baseline. The pre-fentanyl baseline value was defined as the average of the last 5 minutes prior to the first fentanyl dose. Secondary endpoints included the number and percentage of participants who experienced apnea (defined as ≥20 s loss of respiratory activity) and required verbal stimulation to breath after a fentanyl dose.

Buprenorphine average plasma concentration at steady-state (C_avg) was calculated as the area under the plasma concentration-time curve between 120 and 360 minutes after the start of buprenorphine infusion divided by the time interval. Treatment-emergent adverse events (TEAEs) were recorded from time of first screening visit through the end of the last visit. Fentanyl dose escalation was halted if a subject did not breathe for a prolonged period or SpO₂ dropped below 85% despite active verbal stimulation by the investigator, or if the investigators deemed necessary (i.e. other TEAEs). Drug plasma concentrations and safety measures (SpO2, TEAEs) were exploratory endpoints.

Statistical analysis

To reduce the impact of measurement noise on V_E measures, analysis of V_E endpoints was conducted after post-hoc adjustment of data sets. The adjusted data sets reflect imputations based on clinical notes to account for the impact of concurrent clinical events such as facemask removal, urinating with facemask on and severe itching. Stimulated, nonspontaneous breathing data were set at zero (apnea) for analyses on ventilation data.

Maximum percent decreases in V_E relative to baseline were compared between treatment groups using a mixed effects model with treatment as a fixed effect. For Part B, all buprenorphine concentration groups were combined to perform the treatment comparison. Maximum percent decreases in V_E were assessed within the first 10 minutes after each fentanyl bolus to minimise the impact of random variation evident over the full 60-minute intervals. Secondary endpoints were compared between treatment groups by exact conditional logistic regression and Fisher’s exact test. The exploratory safety endpoint (SpO₂) was analysed in a similar manner to changes in V_E.

The primary and secondary endpoint analyses were performed on participants who received at least 1 dose of fentanyl and had at least 1 post-dose assessment, excluding one subject who received the wrong buprenorphine infusion rate. TEAEs were summarized for participants who received at least one dose of study medication. The buprenorphine and fentanyl plasma concentrations were summarized for all subjects who received at least 1 dose of the medication and had an adequate number of pharmacokinetic samples collected.

Given the nature of this study, no sample size or power calculations were performed, and statistical testing was descriptive. Statistical analyses were performed using SAS version 9.4.