Abstract
Objectives
The aim of this study was to optimise dexmedetomidine and alfaxalone dosing, for intramuscular administration with butorphanol, to perform minor surgeries in cats.
Methods
Initially, cats were assigned to one of five groups, each composed of six animals and receiving, in addition to 0.3 mg/kg butorphanol intramuscularly, one of the following: (A) 0.005 mg/kg dexmedetomidine, 2 mg/kg alfaxalone; (B) 0.008 mg/kg dexmedetomidine, 1.5 mg/kg alfaxalone; (C) 0.012 mg/kg dexmedetomidine, 1 mg/kg alfaxalone; (D) 0.005 mg/kg dexmedetomidine, 1 mg/kg alfaxalone; and (E) 0.012 mg/kg dexmedetomidine, 2 mg/kg alfaxalone. Thereafter, a modified ‘direct search’ method, conducted in a stepwise manner, was used to optimise drug dosing. The quality of anaesthesia was evaluated on the basis of composite scores (one for anaesthesia and one for recovery), visual analogue scales and the propofol requirement to suppress spontaneous movements. The medians or means of these variables were used to rank the treatments; ‘unsatisfactory’ and ‘promising’ combinations were identified to calculate, through the equation first described by Berenbaum in 1990, new dexmedetomidine and alfaxalone doses to be tested in the next step. At each step, five combinations (one new plus the best previous four) were tested.
Results
None of the tested combinations resulted in adverse effects. Four steps and 120 animals were necessary to identify the optimal drug combination (0.014 mg/kg dexmedetomidine, 2.5 mg/kg alfaxalone and 0.3 mg/kg butorphanol).
Conclusions and relevance
The investigated drug mixture, at the doses found with the optimisation method, is suitable for cats undergoing minor clinical procedures.
Introduction
In several Swiss veterinary practices, intramuscular (IM) injectable anaesthesia is preferred over intravenous (IV) and inhalational techniques for cats undergoing minor clinical procedures. The rationale for this tendency may be the common belief, among general practitioners, that deep sedation is safer than inhalational anaesthesia. As an alternative explanation, owing to the uncooperative nature of cats, IV catheterisation can be challenging in non-sedated cats and this may prevent veterinarians from attempting physical restraint.
To date, several drug combinations have been investigated for feline IM anaesthesia. Commonly used protocols often include an a2-adrenoreceptor agonist to produce reliable sedation, an opioid derivative to provide some analgesia, and ketamine owing to its anaesthetic and analgesic effects.
Among a2-adrenoreceptor agonists, dexmedetomidine – the active enantiomer of racemic medetomidine – has recently been licensed for administration to cats in some countries and has gained wide popularity in general practice owing to its convenience, ease of administration and possibility to antagonise its effects. Several investigators found that dexmedetomidine is a reliable sedative in cats and produces dose-dependent analgesia and muscle relaxation, with acceptable side effects.1,2 These features make this compound suitable as sole agent for minor clinical procedures associated with mild nociceptive stimulation.
Butorphanol is a synthetic opioid with agonistic activity at κ-opioid receptors and antagonistic effects at µ receptors. 3 Because in most countries it is not listed among controlled substances, it is very often preferred over more potent opioid derivatives by many practitioners in Europe. After IM administration, butorphanol has been found to decrease the thermal nociceptive threshold and produce short lasting and variable analgesia in cats.4,5 Its combination with dexmedetomidine has been shown to result in greater sedation and more profound muscle relaxation, but not faster onset of recumbency, than dexmedetomidine alone in experimental cats. 6
Ketamine is a dissociative anaesthetic with high bioavailability and short onset of action after IM administration. 7 Its use has been widely investigated in cats.8–12 However, its addition to dexmedetomidine and butorphanol was found to prolong the time to recovery. 6
As an alternative to ketamine, the inclusion of alfaxalone in α2-adrenoreceptor agonist–opioid combinations may offer some advantages for cats undergoing minor procedures, namely short duration of the anaesthetic effect, high therapeutic index and excellent muscle relaxation. 13 Additionally, clinical studies investigating the effects of subcutaneous alfaxalone prepared in 2-hydroxypropyl-β cyclodextrin solution seem to indicate a rapid systemic absorption.14,15
On the basis of the current literature, the dexmedetomidine–butorphanol–alfaxalone IM combination can be considered a promising anaesthetic technique for cats undergoing minor clinical procedures. The optimal dose combination of these three drugs is unknown and unlikely to be identified by randomised controlled study, owing to the numbers of animals required for all possible combinations to be investigated. For this reason, we applied a modified ‘direct search’ model, 16 an innovative optimisation method based on a mathematical algorithm, the main advantage of which is that a limited number of dose combinations – and therefore a limited number of patients – require investigation.
The aim of the current study was to optimise, via an innovative experimental design, the doses of IM dexmedetomidine, butorphanol and alfaxalone, in terms of reliability and rapid onset of anaesthesia and absence of perianaesthetic complications, in domestic cats undergoing minor surgical procedures.
Materials and methods
Study design and ethical approval
This study was designed as an investigator-blind, randomised, prospective clinical trial, and performed with the permission of the local ethics committee for animal experimentation (Canton of Berne, Switzerland, license number 22197) and written informed owner consent.
Animals
One hundred and twenty client-owned adult cats admitted to the Veterinary Teaching Hospital of the University of Berne from January 2013 to January 2015 for minor surgical procedures (wound curettage followed by bandage/wound dressing change, or control radiographic examination followed by external pin removal) were enrolled in the study.
Food, but not water, was withheld for 12 h prior to anaesthesia. Cats underwent a routine preanaesthetic physical examination in order to assess their health status. In cooperative animals, in which venepuncture and blood sampling could be performed without sedation, basic blood parameters (haematocrit, total proteins, and electrolytes and serum creatinine levels) were also assessed. Exclusion criteria were pregnancy, systemic diseases, impaired cardiovascular function, elderly (>8 years) and American Society of Anesthesiologists’ risk classification grade >II.
The animals were randomly assigned to one of five treatment groups. A manual randomisation technique was used (drawing group assignment from an opaque envelope).
Treatment groups
Cats were injected intramuscularly with a mixture of dexmedetomidine (Dexdomitor; Pfizer), alfaxalone (Alfaxan; Vétoquinol) and butorphanol (Morphasol; Graeub), mixed in one syringe, at one of the following dose combinations:
group A: 0.005 mg/kg dexmedetomidine, 2 mg/kg alfaxalone and 0.3 mg/kg butorphanol;
group B: 0.008 mg/kg dexmedetomidine, 1.5 mg/kg alfaxalone and 0.3 mg/kg butorphanol;
group C: 0.012 mg/kg dexmedetomidine, 1 mg/kg alfaxalone and 0.3 mg/kg butorphanol;
group D: 0.005 mg/kg dexmedetomidine, 1 mg/kg alfaxalone and 0.3 mg/kg butorphanol;
group E: 0.012 mg/kg dexmedetomidine, 2 mg/kg alfaxalone and 0.3 mg/kg butorphanol.
Each treatment group was composed of six cats. The rationale for choosing this number is explained in the supplementary material. The drugs doses for groups A–E were selected on the basis of the following criteria: clinical experience of the study’s designer, existing literature and manufacturers’ recommendations for the species. In order to avoid iatrogenic muscular lesions, the volume of injectate exceeding 1 ml was equally divided into two injection sites.
Anaesthetic procedure
Prior to anaesthesia, baseline values for heart rate (HR), respiratory rate (RR) and rectal body temperature were recorded. To evaluate the baseline temperament and behaviour, each cat was assigned to one of three categories: tranquil and quiet, stressed or scared, and aggressive.
The time to anaesthesia, defined as the minutes elapsing from injection to lateral recumbency, was recorded.
The occurrence of undesired effects after IM injection, namely vomitus, hypersalivation, respiratory depression and/or increased muscular tone, was noted. A score ranging from 0–4 (undesired effects score) was assigned to each cat, with 0 being ‘none of the listed side effects was observed’ and 4 being ‘all the listed side effects were observed’.
Ten minutes after IM injection, a catheter was placed into the right or left cephalic vein, or in a peripheral vein of one hindlimb if judged more appropriate with respect to the clinical procedure.
The observer evaluated the degree of reaction to IV catheter placement by assigning a score, ranging from 0–4, with 0 being ‘no reaction’, 1 being ‘mild reaction’ (attempts to withdraw the limb), 2 being ‘moderate reaction’ (vocalisation and and/or hissing, movements, one person needed for physical restraint) and 3 being ‘aggressive reaction’ (vocalisation and/or hissing, attempts to bite, two people required for adequate physical restraint).
A score ranging from 0–5 (composite anaesthesia score), as described by Biermann et al, 11 was assigned every 5 mins until completion of the clinical procedures. The latter were started as soon as the cats became laterally recumbent.
During anaesthesia, the cats were instrumented with a pulse-oximeter (Microcap Plus; Oridion), an electrocardiograph (Schiller AT-4; Medical Device Depot) and a Doppler (Model 811-B; Parks Medical Electronics) for non-invasive arterial blood pressure measurement. The respiratory rate was determined by visual examination of the thorax. Arterial oxygen saturation, HR, RR and systolic arterial blood pressure (SAP) were manually recorded every 5 mins.
The depth of anaesthesia was assessed on the basis of commonly used clinical descriptors (presence or absence of corneal and palpebral reflexes and degree of muscular relaxation).
If major movements, defined as flexion/extension of the limbs and/or of the neck, and/or lifting of the head, were observed during the surgical procedure, propofol (Propofol 1%; Fresenius Kabi) was administered intravenously in steps of 1.5 mg/kg. The number of propofol boli given to each cat, as well as the time of administration, were recorded. All cats received oxygen supplementation by mask with the flow rate set to deliver 2 l/min, and a balanced crystalloids solution (Ringer-Lactat; Baxter) was administered intravenously at the rate of 5 ml/kg/h.
In the event of bradycardia (HR <100 beats per minute [bpm]) with normo- or hypotension (SAP <100 mmHg) and with or without ventricular escape rhythm, glycopyrrolate (Robinul; Sintetica SA), 0.01 mg/kg, was given intravenously. In the event of moderate sinus bradycardia (HR 99–70 bpm) accompanied by hypertension (SAP >150 mmHg), no anticholinergic was given and the rhythm was monitored closely to detect any change from the baseline. Finally, if severe bradycardia (HR <70 bpm), or moderate bradycardia accompanied by hypertension and ventricular escape rhythm, were observed, atipamezole (Antisedan; Provet) was administered intramuscularly at five times the dexmedetomidine dose. The duration of the clinical procedure was recorded. At the end of the procedure, a 10 cm visual analogue scale (VAS), with 0 being labelled as worst possible and 10 as best possible, was used for an overall, subjective evaluation of the quality of anaesthesia. Thereafter, unless it had been necessary to antagonise the dexmedetomidine before the completion of the procedure, the cats were injected with IM atipamezole. The time to recovery, defined as the minutes elapsed from atipamezole injection to active interaction, was recorded. The rectal body temperature was measured at the end of anaesthesia.
The quality of recovery was assessed using a composite recovery score, ranging from 0–14 and based on the following descriptors: comfort, coordination, vocalisation, movement during sternal recumbency, locomotor activity and scratching and grooming, as described by Biermann et al. 11 Additionally, a 10 cm VAS, with 0 labelled as worst possible recovery and 10 labelled as best possible recovery, was used.
All the assessments were performed by the same anaesthetist (CA), who also injected the cats and was blind to the treatment.
Optimisation procedure
The optimisation procedure used in this study was a modification of the ‘direct search’ method, as described by Berenbaum, 17 and applied by Sveticic in a clinical trial involving human patients. 16 A detailed description of the methods and of the equations used is provided in the supplementary material.
The optimisation of the drug combination was conducted in a stepwise manner, and the treatment groups tested at each step were defined as a complex. The initial five drug combinations tested composed complex number 1. Within each complex, the best treatment groups in terms of quality of anaesthesia (greatest median composite anaesthesia scores, mean VAS anaesthesia scores and VAS recovery scores, and lowest medians of number of propofol boli and composite recovery scores) and minimal side effects (lowest median undesired effects scores) were identified. These groups were defined as ‘promising groups’, whereas the remaining ones were classified as ‘unsatisfactory groups’. The new complex, tested in 30 other cats (six cats per group), was composed of four of the previously tested treatments (all but the worst one) plus a fifth, generated by applying a mathematical model on the basis of the results of the previous step (see supplementary material). Data obtained from complex 2 were analysed as previously described and used to generate a new complex until an optimal drug doses combination was found. For alfaxalone, the maximal dose to be used was fixed at 2.5 mg/kg.
The optimisation procedure was considered completed when the scores obtained from the new drug combination were not better than those of the previously tested treatments, and when these results were consistent for three consecutive steps. The dose searching method was to conclude also in case the mathematical model generated a drug combination that had already been tested in at least one of the previous steps.
Statistics
Commercially available software (NCSS-2007 [NCSS] and SigmaStat and SigmaPlot 12 [Systat Software Inc]) were used. The normality of data was tested with the Shapiro–Wilk test. Descriptive statistics were used for comparison of treatment groups with respect to the following variables: composite anaesthesia score, VAS anaesthesia and recovery scores, composite recovery scores, medians of number of propofol boli and undesired effects scores. The cardiorespiratory variables (HR, RR and SAP) were analysed with repeated measures
P values <0.05 and Z values >1.96 were considered statistically significant.
Results
Data for body weight, age and rectal body temperature were normally distributed.
Four complexes, 120 cats and eight groups, five of which included in complex 1 and three new ones (N1, N2 and N3; Table 1), were necessary to establish the optimal drug combination, which was 0.014 mg/kg of dexmedetomidine, 2.5 mg/kg of alfaxalone and 0.3 mg/kg of butorphanol.
Doses for intramuscular alfaxalone (A), dexmedetomidine (D) and butorphanol (B), expressed in mg/kg, for eight treatment groups
The optimisation procedure concluded after completion of data collection for complex 4, as the fourth new treatment group obtained through the mathematical model was the same as N2, which had been previously tested.
The cats had a mean ± SD body weight of 4.2 ± 0.8 kg and a mean ± SD age of 4.4 ± 1.6 years; 61 were males (55 of which were castrated) and 59 were females (all of which were spayed). The eight treatment groups did not differ with respect to type of clinical procedure (P = 0.66; Table 2), duration of the procedure (P = 0.17; Figure 1), time to anaesthesia (P = 0.79; Table 3) and time to recovery (P = 0.29; Table 3). Hypertension and bradycardia were observed in 11% and 25% of the cats, respectively. However, no statistically significant difference in HR (P = 0.76), RR (P = 0.32), SAP (P = 0.35) and rectal body temperature (P = 1.2) was detected, neither between treatments nor between time points (Figures 2–4).
Number of cats per group with one of three types of baseline temperament, undergoing one of two types of minor surgical procedure
WC = wound curettage followed by bandage/dressing change; RX = control radiographic examination followed by external pins removal; TQ = tranquil and quiet; SS = stressed or scared; A = aggressive

Box and whisker plots of duration of the clinical procedure in 120 anaesthetised cats. Boxes represent the interquartile range (central 50% of values), while the horizontal lines within the boxes are the medians. The upper and lower whiskers represent the upper and lower ranges of values, respectively. The dots represent the outliers
Medians and ranges of time to anaesthesia, defined as the minutes elapsed from intramuscular (IM) injection to lateral recumbency; and time to recovery, defined as the minutes elapsed from atipamezole IM injection to active interaction

Box and whisker plots of heart rates (HR; beats per minute [bpm]) of 120 cats undergoing minor surgical procedures under injectable anaesthesia. See Figure 1 for further explanation

Box and whisker plots of respiratory rates (RR; breaths per minute [bpm]) of 120 cats undergoing minor surgical procedures under injectable anaesthesia. See Figure 1 for further explanation

Box and whisker plots of systolic arterial pressures (SAP; mmHg) of 120 cats undergoing minor surgical procedures under injectable anaesthesia. See Figure 1 for further explanation
All eight drug combinations resulted in a degree of unresponsiveness sufficient to allow for IV catheterisation 10 mins after injection, so that in all cats physical restraint was unnecessary. None of the cats required administration of glycopyrrolate or atipamezole before the end of the clinical procedure. None of the tested treatments resulted in clinically relevant side effects, and anaesthesia and recovery were uneventful for all of the cats enrolled in the study. The medians or means of VAS anaesthesia, composite anaesthesia score, number of propofol boli, undesired effects score, VAS recovery, and composite recovery score, are summarised in Table 4.
Variables for complexes 1, 2, 3 and 4. The worst treatment groups (to be excluded from the optimisation process) are coloured in red, while the best final treatments are in bold characters
Median
Mean
VAS A = VAS anaesthesia; CAS = composite anaesthesia score; PropB = number of propofol boli; UES = undesired effects score; VAS R = VAS recovery; CRS = composite recovery score; P = promising complex; U = unpromising complex
Discussion
The main finding of this study is that the optimal alfaxalone–dexmedetomidine–butorphanol combination established by applying the optimisation method resulted in reliable, rapid onset, and uneventful anaesthesia in cats undergoing minor surgical procedures.
The modified version of the optimisation method as used in this study was feasible, applicable to the clinical setting and compatible with the routine of a busy veterinary hospital.
The rationale for using a relatively high number of cats was that, for an objective and unbiased data collection, the investigator needed to be blind to the treatments at each step of the search process. This implied the repetition of the good treatments, so that each complex could be composed of five groups. A different study design, with the investigator responsible for data collection completely unaware of the study phases, would have allowed the avoidance of such repetition and made the trial less time consuming.
None of the cats had complications. However, mild hypertension and bradycardia were observed in some animals, a finding that limits the use of the anaesthetic protocol to healthy cats.
Providing a definition of bradycardia and hypertension in anaesthetised cats is challenging owing to the conflicting results in other studies. It was reported that cats have considerably lower heart rates when the measurements are taken in the home environment compared with the hospital, and that normal values for HR in quiet cats are 132 ± 19 bpm. 18 Considering that under anaesthesia even lower values could be regarded as normal, for this trial bradycardia was defined as HR <100. Regarding the SAP, according to Domanjko et al normal values range from 100–150 mmHg in anaesthetised cats. 19 Hence, in the present study values >150 mmHg were considered indicative of hypertension.
One limitation of the proposed anaesthetic protocol is that the addition of alfaxalone, which is available on the market only at a concentration of 10 mg/ml, greatly increases the volume of injectate in comparison with a classical dexmedetomidine–ketamine combination. This certainly makes the IM injection an unpleasant experience for cats, and implies that two injections are often necessary to split the volume into two anatomic sites. Additionally, a high-volume IM injection carries the potential for iatrogenic muscular injuries. In the light of these considerations, it was decided to fix the maximal alfaxalone dose at 2.5 mg/kg (corresponding to a volume of 0.25 ml/kg). Nevertheless, none of the cats enrolled in the study showed signs of pain at the injection sites or impairment of the injected limb motor function in the postanaesthetic period.
One important thing to consider when comparing different anaesthetic treatments is that the procedures to be performed in the anaesthetised patients should be similar in terms of duration and degree of nociceptive stimulation. Because of the clinical nature of the current trial, a standardisation of the surgical procedure was not possible. After an attentive analysis of the caseload of the institution in which the study was conducted, it was decided to enrol only cats undergoing one of the two most common minor procedures. Unarguably, the surgical curettage of a skin wound and the removal of a pin from a bone may be different in terms of types and intensity of nociceptive stimulation. However, the even distribution of the types of procedure, and the lack of significant difference in its duration, between treatments, should have prevented our findings from being biased.
Conclusions
The modified optimisation method, as described by Berenbaum, 17 was easy to apply and allowed us to establish a useful drug combination to be used in clinical patients.
Dexmedetomidine–alfaxalone–butorphanol combination, at doses of 0.014, 2.500 and 0.300 mg/kg, produces good quality injectable anaesthesia in cats, characterised by reliability, rapid onset and lack of complications, and can be recommended for minor surgical procedures.
Supplemental Material
Click here for Supplementary Material
Stepwise optimisation method to identify optimal clinical doses for intramuscular anaesthesia in cats. Detailed description of the modified ?direct search? method, as described by Berenbaum, reference 17 including the equations used to determine the number of cats to be enrolled in the trial, and for the dose-finding process
Footnotes
Acknowledgements
We would like to thank Dr Laura Züger for her fine contribution to data collection.
Supplementary material
Stepwise optimisation method to identify optimal clinical doses for intramuscular anaesthesia in cats. Detailed description of the modified ‘direct search’ method, as described by Berenbaum,17 including the equations used to determine the number of cats to be enrolled in the trial, and for the dose-finding process.
Conflict of interest
The authors do not have any potential conflicts of interest to declare.
Funding
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
