Propranolol as an Anxiolytic to Reduce the Use of Sedatives From Critically-ill Adults Receiving Mechanical Ventilation

Background and Rationale:

Critically ill patients often require sedation for comfort and to tolerate mechanical ventilation. There are internationally accepted guidelines for sedation of critically ill patients, and dexmedetomidine and propofol are recommended as sedative agents over alternatives, such as benzodiazepines. There is now a worrying shortage of propofol in Canada, the European Union, and parts of the US. The supply of dexmedetomidine, an expensive alternative to propofol, is now considered to be at risk. It is unsuitable as monotherapy for deep sedation, and most hospitals in Canada limit use to 48h, due to high costs. This therefore makes it inadequate for COVID-19 treatment. Other sedative agents for mechanical ventilation include benzodiazepines (e.g. midazolam), but these are associated with higher mortality from sepsis, higher incidence of delirium, and longer length of stay in the ICU. That notwithstanding, midazolam is also currently in short supply. Shortages of sedative medications will be as impactful on critical care practice as shortages of mechanical ventilators; propofol, midazolam, and dexmedetomidine were all listed on March 31, 2020 as "Tier 3 Shortages" by Health Canada - having "the greatest potential impact on Canada's drug supply and health care system….based on low availability of alternative supplies, ingredients or therapies."

In critically-ill patients, the sympathetic nervous system can become hyperactive, producing neurotransmitters such as norepinephrine (NE) to increase blood pressure and heart rate. In the brain, a pontine nucleus called the locus ceruleus (LC) provides the majority of brain NE. LC adrenergic input to the medial septal area and medial preoptic area of the forebrain mediates arousal.

Sympatholytics have previously been prescribed to manage agitation in critical illness. Dexmedetomidine is one such example, an intravenous alpha-2 agonist, which readily penetrates the central nervous system. It has an anti-noradrenergic effect in the locus ceruleus, consequently increasing inhibitory GABA neuron activity in the forebrain. Clonidine, is an oral and intravenous agent that is highly lipid soluble, with good penetration of the Central Nervous System, has been well studied in the ICU environment. However a recent meta-analysis showed clonidine use did not reduce length of stay or length of ventilation for critically ill patients, and only led to a small reduction in the use of opioids. Propranolol is a non-selective beta-adrenergic antagonist, approved for treating hypertension, angina, arrhythmias, migraines and pheochromocytoma in Canada. It has also been used off label to treat anxiety disorders such as Post-Traumatic Stress Disorder. It is a lipophilic molecule, crosses the blood brain barrier, and can block the locus ceruleus' ability to activate the forebrain, similar to alpha 2 agonists. In rat models, propranolol delays arousal from a state of anaesthesia.

An extensive search of PubMed and Web of Science did not reveal published trials of propranolol as a sedative agent in critical illness or mechanical ventilation. However, propranolol was shown to significantly reduce agitation in a randomized controlled trial in patients with traumatic brain injury. Observational studies in traumatic brain injury have shown that propranolol use is associated with a shorter length of hospital and ICU stay, and possibly a lower mortality risk, without reports of significant side effects. A meta-analysis of 10 randomized controlled trials in severely burned patients found propranolol reduced hospital length of stay. Other studies have shown that propranolol may have beneficial effects on catabolism in critical illness, and improve cardiac function and survival in animal models of cardiac resuscitation. This study team published a single centre retrospective study of 64 mechanically-ventilated patients which found that the initiation of propranolol was associated with an 86% reduction in propofol dose, and an approximately 50% reduction in midazolam dose, while maintaining the desired sedation target.

If propranolol reduces sedative doses required by mechanically ventilated patients (either with or without COVID illness) to the degree seen in this team's retrospective study, this would effectively reduce usage of sedatives two- to seven-fold. The cost of the average daily dose of propranolol used in the retrospective study (~120mg) would be $0.27 (pharmacy cost at The Ottawa Hospital), and propranolol is in abundant supply in Canada and around the world.

Based on this data, propranolol may enable critical care providers to sedate mechanically ventilated patients using their usual approaches except with substantially lower doses of sedatives. This would not only extend the use of currently limited supply of sedatives, it would significantly reduce medication costs in critical care units in Canada and worldwide.

Trial Objective:

Does the addition of propranolol to a standard sedation regimen reduce the dose of sedative needed in critically ill patients requiring mechanical ventilation?

Study Design:

Open-label, Randomized Controlled Trial (RCT), 1:1 allocation

Patient Population:

Participants are adult patients admitted to an intensive care unit who are anticipated to require mechanical ventilation >48h, who are requiring intravenous sedatives to achieve a sedation goal that is anticipated to be stable for >48h.

Intervention:

Patients randomized to the intervention arm will receive propranolol enterally at a starting dose of 20mg every 6 hours for two to four doses, and then re-assessed for upwards titration every 24 hours (+/- 6 hours) at 10mg dose increases depending on clinical response. The maximum dose to be used in the intervention group is 60mg every 6 hours. The decision to titrate propanolol will be made daily at rounds; this is why participants will receive 2-4 doses at 20mg (to accommodate for time from enrollment to rounds the next day) and titration will occur every 24 hours +/- 6 hours (to accommodate rounds timing). Titration timing is based on clinical rounds so that the decision to titrate can be made by the clinical team in a way that aligns with their clinical workflows and reflects actual clinical practice in the ICU.

Daily dose titration will be guided by hemodynamic markers indicative of the expected sympatholysis from propranolol. Upward titration of propranolol should coincide with a downward titration in sedatives until a minimum level of sedative infusion is reached (propofol <0.5mg/kg/h or midazolam <0.5mg/h). For each participant, the intervention will be administered from enrollment until approximately 48h after the patient is liberated from the mechanical ventilator, and the study will continue until death, ICU discharge, or 30 days after enrollment, whichever is first. Enrollment can begin as early as 24h after the start of a sedative infusion in the ICU.

Primary Outcome and Sample Size:

The primary outcome will be a comparison of the change in primary sedative dose from baseline to Day #3 of the study in the intervention group compared with the same change in the control group- the difference in differences. Given the global need for a substantial reduction in sedative consumption for mechanically ventilated patients, we will assume that a 70% reduction in sedative dose would be clinically meaningful, which corresponds to the difference between the sedative infusion doses required for study eligibility (ie. propofol >1.5 mg/kg/h) and the minimum doses below which propranolol would no longer be increased (ie. propofol <0.5 mg/kg/h). Since we do not have (or know of) any data with which to estimate the standard deviation of the change in sedative use, we base our sample size calculation on the Cohen's d statistic (also known as the standardized effect size). The standardized effect size is computed as the magnitude of the difference between groups divided by the pooled standard deviation of the response measure (in our case, the change in sedative dose from baseline to day 3). By widespread consensus, standardized effect sizes of 0.2, 0.5, and 0.8 are respectively considered to be small, moderate, and large effect sizes. Given we would need to see at least a moderately large change in sedative use (decrease of 70%) to effect a substantial improvement in the overall sedative supply, we choose to power our trial to detect a standardized effect of 0.5. In order to have 80% power to detect such a minimal clinically important difference, we require a total of 32 participants in each study arm. in a 1:1 randomized controlled trial. Allowing a 10% dropout rate (a previous study of sedation strategies at these sites reported a dropout rate of 3%30), we will aim to enroll 72 patients (36 control; 36 intervention).

Statistical Analysis:

The primary outcome will be analyzed as a difference in differences- the change in sedative dose from baseline to Day #3 (defined as the 24h period starting 60h after enrollment) in the intervention group vs the same change in the control group. The Mann-Whitney U test will be used as a nonparametric test of independent samples for this outcome, as well as the secondary outcomes that relate to sedative dosing on Day #3 compared to baseline. For secondary outcomes, groups will be compared using unpaired statistical tests- Chi-square for proportions, and unpaired t-tests or Mann-Whitney tests as appropriate for parametric and non-parametric data. The effect of age, sex and gender, and pre-ICU beta-blocker prescription will be analyzed as planned subgroup analyses.

Ethical Considerations:

A No Objection Letter from Health Canada and Research Ethics Board approval have been obtained to conduct this clinical trial. An established data safety monitoring committee (DSMC) will evaluate adverse events. Two interim safety analyses will be conducted by the DSMC and the study will be terminated if the DSMC determines the risk of adverse events, based on the rate of adverse events reported in this clinical trial to the committee, outweighs the potential benefits. Informed consent will be obtained from each participant or their legal substitute decision maker.