Comparison of the onset and offset times of two available formulations of rocuronium bromide in an animal model

Abstract

Rocuronium bromide is a neuromuscular blocker in widespread use in anaesthesia, emergency and intensive care. Reports of reduced efficacy of a new different formulation of rocuronium bromide were submitted to Medsafe, the New Zealand Medicines and Medical Devices Safety Authority, in 2020. Given the requirement for rapid and predictable paralysis for patient safety the efficacy of the two available formulations of rocuronium bromide was investigated in an animal model. After ethics committee approval, 19 rats were anaesthetised and paralysis, defined as loss of tibialis anterior flexion on direct electrical stimulation of the sciatic nerve, was assessed by mechanomyography in response to ED90 doses of rocuronium.

Paralysis was observed at a median of 12 seconds for the new different formulation: A, Hameln Pharma (interquartile range (IQR) 6–106 seconds) and 28 seconds for formulation B: Pfizer (IQR 12–68 seconds) P = 0.48. Offset of paralysis was observed after 293 seconds for formulation A (IQR 250–372 seconds) and 241 seconds for formulation B (IQR 220–263 seconds). While the differences observed were substantial, they were not statistically significant. Moreover, the direction of observed difference was towards a shorter median onset and longer offset for the newer formulation, a finding in the opposite direction to the initial clinical concern.

Relevance to the clinical situation is indeterminate given the study was stopped at low numbers for futility and limitations around the clinical applicability of animal pharmacokinetics and dynamics. Nevertheless our findings provide some reassurance that the newly available different formulation of this critical use medication does not exhibit a substantial increase in time to onset.

Introduction

Rocuronium bromide, an aminosteroid compound that causes competitive neuromuscular blockade, is widely used in anaesthesia, emergency and intensive care. It is used to produce skeletal muscle relaxation to facilitate access to the infraglottic airway, intubation, ventilation and immobility for surgery. It is important for patient safety in emergent and non-emergent settings that the interval between dosing and the onset of apnoea associated with muscle relaxants occurs in a dose-dependent fashion, is timely and predictable so that there is reliable paralysis sufficient to allow both the safe placement of airway devices and establishment of ventilation.

Recently a new and different formulation of rocuronium bromide (Hameln Pharma 10 mg/ml) became the predominant product available in the New Zealand market. Previously supply was predominantly a Pfizer preparation (DBL rocuronium bromide 10 mg/ml). Both preparations contain the bromide salt form of rocuronium, sodium chloride, sodium acetate and water. The Hameln Pharma formulation also contains acetic acid and/or sodium hydroxide and the Pfizer formulation contains acetic acid. There have been reports in different hospitals where the time of onset of neuromuscular blockade with the new formulation has been greater than predicted by known pharmacokinetics and experience with other preparations. This has raised questions over the use of this formulation among anaesthetists and intensive care specialists. These concerns have been raised on NZ Health Forum, formal reports submitted to Medsafe ¹ (NZ Medicines and Medical Devices Safety Authority) and reported in the general media. ²

On 16 March 2021 Medsafe reported no adverse effects reports since October 2020, and the monitoring period was stopped on 31 May 2021. ¹ Given the nature of the concerns raised, we aimed to investigate in an animal model whether the time to onset of neuromuscular blockade is different between the two available formulations of rocuronium bromide.

Methods

Location

The animal study was conducted at the Ruakura Animal Research Facility, Hamilton, New Zealand on 23–25 June 2021. All study protocols were reviewed and approved by the Ruakura animal ethics committee.

Animals

Nineteen female Sprague–Dawley rats were studied with an age range of 120–125 days and a weight range of 335–430 g. Animals were kept in single gender enclosures with no chance of pregnancy. Twelve-hour light–dark cycles (lights on/off at 07:00/19:00 hours) and climate control were maintained. Access to feed and water was freely allowed until the day of animal utilisation.

Animal manipulations

On the day of study animals were sedated with ketamine at 50 mg/kg (Mayne Pharma Ltd., Auckland, New Zealand), and xylazine at 4 mg/kg (Bayer Health Care, Leverkusen, Germany) via intraperitoneal injection. Animals were then placed on a warming board at 38°C. Following dissection of the anterior neck, a tracheostomy tube (14 G × 1.77 in; Insyte, BD, Switzerland) was placed under direct vision. An intravenous (IV) cannula was placed in the internal jugular vein using a catheter over needle technique (24 G × 0.75 in; Insyte, BD, Switzerland). Mechanical ventilation using a small animal ventilator (Inspira ASV, Harvard Apparatus, MA, USA) was then undertaken with oxygen as the inspired gas admixed with 2% isoflurane (Merial, Auckland, New Zealand) via a vaporiser (Somnosuite Small Animal Anesthesia System, Kent Scientific Corp., Torrington, CT, USA) to maintain anaesthesia. Anaesthesia and mechanical ventilation were continued for the duration of the experiment.

Experimental protocol

Anaesthetised and instrumented rats were placed in the lateral position on the warmer board. A skin incision was made from the knee towards the ischial tuberosity; immediately below this incision was the gluteal muscle. This was elevated and a cut made. The gluteus muscle was split via blunt dissection in a plane parallel to the skin incision. Immediately deep to the muscle incision the sciatic nerve was blunt dissected and insulated from the surrounding muscle by a plastic cover placed deep to the nerve. The leg was held stable by an 18G needle affixed through the knee joint into a corkboard attached to the warming plate. A tie was placed around the foot of the rat at the level of the metatarsal pads and pulled until just taut while attached to a force transducer (MLT 0420 Force Transducer; AD Instruments, Dunedin, New Zealand), as per the method of Boucheix et al. ³ The force generated by foot movement was recorded as an electrical signal recorded in millivolts, recorded in a Powerlab 8/35 (AD Instruments, Dunedin, New Zealand). At the completion of the experiment animals were euthanised with intravenous pentobarbitone.

Measurement of time to neuromuscular blockade

A binary random number generator was used for group allocation. One of two formulations of rocuronium bromide (Hameln Pharma (A) or Pfizer (B)) was given diluted in saline to 0.5 mg/ml at 1.34 mg/kg (the previously determined ED90) ³ injected as a fast IV push, followed by a 0.5 ml saline flush to clear the dead space of the intravenous cannula. A stopwatch timer was started at the end of the injection. An MLA0320 animal nerve stimulating electrode (AD Instruments, Dunedin, New Zealand) was applied to the sciatic nerve manually and held in contact with the nerve for the duration of data acquisition. At the same time as the stopwatch was started, an electrical stimulus as a 5 V pulse at 0.16 Hz was commenced. The primary variable of interest was paralysis of the tibialis anterior as demonstrated by an absent positive deflection on the force transducer. The primary endpoint was measured using the electronic force transducer strips and was evaluated after the experiment by a study member (SO) blinded to study group allocation. Stopwatch recording of recovery times was also by a study member (LMS) blinded to study group allocation.

The initial planned time of observation was two minutes. The primary outcome was seen in the first three subjects within this timeframe, and in the fourth subject (the second subject from group B) the primary outcome was not seen. We added a further subject to group B prior to the first interim analysis so that the effect of excluding this subject could be estimated, and the time of observation was increased to ten minutes for subsequent subjects. With prolongation of the observation time, we included the time to recovery of positive deflection on the force transducer as a secondary outcome variable.

Statistical analysis

Baseline data are described as means with 95% confidence intervals and range as appropriate. Time to absent positive deflection on the force transducer and time to recovery of positive deflection are reported as medians and interquartile ranges (IQRs). We are not aware of any previous description of the standard deviation (SD) of the primary outcome variable in rats. Previous work has indicated that a clinically relevant difference in onset time for rocuronium bromide in humans is about 30 seconds, equivalent to one SD in young human subjects. ⁴ Our study was powered for a situation in which the difference of interest was equal to the SD of the variable measured using the human value of 30 seconds for each. On this basis we planned for 15 subjects per group to provide 80% power at an alpha of 0.05. The Ruakura animal ethics committee requested that we undertake and report on interim analyses to minimise animal use. We used machine simulation generated sequential stopping rules for small trials from Fitts, ⁵ to plan interim analyses after 9 and 12 subjects in each group. These rules determine P values which permit stopping at interim analysis for a positive result at a threshold low P value and an upper bound P value where the experiment can be halted as the real effect size would be expected to be very small or absent. ⁵ Applying these rules a P value of 0.02 at interim analysis saw the study stop for a positive result with a preserved overall alpha of 0.05, and a P value of 0.25 led to cessation for futility with preserved overall statistical power. Time to recovery in those subjects where this was measured is reported as a secondary outcome variable.

Results

Baseline characteristics

Rats were 149–158 days old at the time of the experiment. Individual birthdates were not recorded for each rat so comparison of distribution between groups was not made. Rats administered formulation B were on average 28 g heavier than those administered the formulation A (370 versus 398 g, mean difference 28 (0–56) g, P = 0.05). While this was statistically significant in the absence of any difference in experimental conditions we attributed this to type I error. The experiment was stopped at the first interim analysis after nine subjects in group A and ten subjects in group B as the P value was above the predetermined threshold for futility.

Primary endpoint: time to absent positive deflection on force transducer

The primary outcome measure was not significantly different between groups. The distribution of the primary endpoint in both groups was not consistent with a normal distribution, as such a non-parametric analysis was used.

The median time to absent positive deflection was 12 seconds (IQR 6–106 seconds) for formulation A and 28 seconds for formulation B (IQR 12–68 seconds), P = 0.48, Mann–Whitney.

Our initial planned period of observation was 120 seconds. In the fourth subject our endpoint was not achieved, and we therefore decided to increase the observation period to 600 seconds or recovery of the positive deflection on force transduction. On unblinding, the fourth subject received the older formulation of rocuronium bromide. At interim analysis we recorded this subject’s time to loss of upstroke as the minimum time paralysis could have occurred (121 seconds, 1 second after the end of observation). The effect of using the lowest possible value for this subject was towards continuing the experiment for a delayed effect of formulation A, the newer formulation. The decision to stop for futility was unchanged when this subject was excluded from analysis, and when using an assumed value of 313 seconds (longer than any other subject).

Secondary endpoint: time to recovery of positive deflection

There was no significant difference in time to recovery of positive deflection in force transducer trace between groups. In subjects that received formulation A the median time was 293 seconds (IQR 250–372 seconds), in those that received formulation B the median time was 241 seconds (IQR 220–263 seconds).

Discussion

Our animal model did not demonstrate any statistically significant difference in any of the metrics of onset or offset for the two formulations of rocuronium bromide, having been stopped at a planned interim analysis for lack of observed difference. While a small animal study can clearly not directly extrapolate to the clinical situation, our data do not provide evidence for any increase in median time to effect for the new versus the older formulation.

While observed differences were not statistically significant, they were substantial in magnitude. The clinical relevance of these differences is unknown. Limitations such as small numbers, the use of human data to power an animal trial and outlier effects mean that this study may lack the power to exclude a real difference. Nevertheless, if there is such a difference our data suggest this would not be in the direction of a longer median time to onset of effect for formulation A versus formulation B, which was the clinical concern which led to this work. As such our findings may provide some reassurance.

Initially, there were multiple reports of lack of efficacy of a new formulation of rocuronium bromide submitted to Medsafe after a warning was published on the NZ Health Forum. It may be that confirmation bias in those who were aware of the initial report(s) of reduced efficacy contributed to their subsequent reporting when in fact there was no change in the efficacy of the neuromuscular relaxant with the new and different formulation.

Our study has other limitations. The dose of rocuronium bromide used clinically in adults is 1.9–3.2 times the ED95 (the dose causing 95% depression of twitch height). ⁶ This large dose is given to achieve high plasma concentrations to speed onset for a drug with relatively low potency at the neuromuscular junction. A lower dose was selected in this model to facilitate potential differentiation of effect between groups. Regardless, in eight of 19 subjects, the primary endpoint occurred within 12 seconds despite using the estimated ED90 dose of rocuronium bromide for rats. Our choice to use ED90 may have affected the spread of results as a proportion of rats would be predicted to have a delayed or absent primary outcome. There were two outliers with effect greater than 120 seconds in the group receiving formulation A and one outlier for formulation B. Differences between groups for times to onset and offset, while not statistically significant, were large relative to the measured values. The median time to effect for formulation A was, at 12 seconds, less than half that for formulation B, although this was not statistically significant. That the difference in medians favoured the newer formulation did, however, influence the decision to invoke the stopping rule at the interim analysis: the observed effect was in the opposite direction to that of the initial clinical concern. The 20% relative difference in time to offset between groups may be similarly influenced by small numbers and outlier effects. The time to offset at four to five minutes in our protocol is very different to the clinical scenario, which reinforces caveats around direct applicability of animal models to the clinical context. Should further animal work be considered, estimation of the ED50 or ED90 may be useful endpoints, although would require significantly larger subject numbers.

Temperature controlled storage of rocuronium bromide is critical to prevent storage lesion and ensure optimal drug effect. The manufacturers recommend rocuronium bromide is refrigerated between 2–8°C and once removed from the refrigerator stored at less than 30°C and used within 8 (Hameln Pharma) ⁷ or 12 weeks (Pfizer). ⁸ It is possible that during transit to New Zealand hospitals the batch of rocuronium bromide for which Medsafe received reports of concern may have had a disruption to cold chain storage. Global freight disruptions and the international shortage of refrigerated containers during the SARS-CoV-2 pandemic has caused logistic challenges in safe transit of foodstuffs and likely has created similar challenges for safe drug transit.^9,10 Exposure to significant heat or to temperatures less than 2°C could have affected drug efficacy and resulted in delayed or inadequate paralysis. A storage lesion due to incorrect storage in individual hospitals seems less likely given the reports of inadequate drug effect were from multiple different sites.

Conclusion

At an ED90 dose of rocuronium bromide in a rat model we did not find any statistically significant difference in the median time to onset of paralysis between the two formulations of the drug available in New Zealand. Relevance to the clinical situation is indeterminate given the study was stopped at low numbers for futility and limitations around the clinical applicability of animal pharmacokinetics and dynamics. Nevertheless, our findings provide some reassurance that the newly available different formulation of this critical use medication does not exhibit a substantial increase in time to onset.