Effect of anaesthetic maintenance with iso ﬂ urane or propofol on ease of endoscopic duodenal intubation in dogs

Objective To compare the ease of endoscopic duodenal intubation (EDI) in dogs during maintenance of general anaesthesia with iso ﬂ urane or propofol infusion. Study design Prospective, randomized, partially blinded clinical trial. Animals A total of 22 dogs undergoing upper gastrointestinal tract endoscopy to include EDI were recruited. Methods


Introduction
Many duodenal or gastric conditions primarily affect the mucosal surface of the gastrointestinal tract (GIT) and are best evaluated by endoscopy (Zoran 2001). Endoscopic evaluation of the upper GIT enables visualization and biopsy of the mucosa without painful surgery in animals who may represent a greater surgical risk (Zoran 2001;Simpson & Hall 2005). Gastroendoscopy, requiring intubation of the duodenum, is indicated in veterinary patients presenting with diarrhoea and chronic vomiting (Simpson & Hall 2005).
Endoscopic evaluation of the duodenum, pylorus and cystic antrum may be unachievable even for experienced endoscopists (Happe & Ivd 1983;Hall 1994;Hall 2013). Positioning the animal in left lateral recumbency, avoiding overdistention of the stomach and minimizing time elapsed between commencing the procedure and attempting endoscopic duodenal intubation (EDI) are practical measures reported to aid EDI (Simpson & Hall 2005;Asorey et al. 2020). Active peristalsis and alterations in pyloric and antral tone may impede EDI (Donaldson et al. 1993). Many drugs affect the pyloric sphincter tone and upper GIT motility, and so pharmacological manipulation may also be employed to improve the ease of EDI (Hall 2013;McFadzean et al. 2017).
General anaesthesia is necessary to facilitate endoscopic evaluation of the upper GIT in the dog to ensure endoscopist and animal comfort and to protect expensive equipment (Zoran 2001). Previous studies have evaluated the effects of anaesthetic premedication on EDI (Zoran 2001;Hall 2013;McFadzean et al. 2017;Salla et al. 2020). EDI was more difficult when animals were premedicated with atropine or metoclopramide (Leib et al. 1990) and with morphine and atropine compared to atropine with or without acepromazine (Donaldson et al. 1993). Premedication with butorphanol resulted in easier EDI than that with methadone (McFadzean et al. 2017), but when combined with acepromazine, butorphanol and methadone led to equivalent ease of EDI (Salla et al. 2020). As such, the effect of anaesthetic premedication on EDI should be taken into account when selecting an anaesthetic protocol for upper GIT endoscopy (Zoran 2001;Hall 2013).
The ease of EDI during general anaesthesia induced and maintained with halothane, isoflurane and enflurane has been evaluated and compared, with no significant differences between volatile anaesthetic agents identified (Leib et al. 1990). Intravenous (IV) propofol and isoflurane delivered at 1.0e1.5% Â minimal alveolar concentration (MAC) have both been recommended to facilitate EDI and upper GIT examination (Leib et al. 1990;Donaldson et al. 1993;Sap & Hellebrekers 1993;Zoran 2001;McFadzean et al. 2017). The duration of upper GIT endoscopy including EDI was shortened when general anaesthesia was maintained with propofol boluses rather than with inhaled halothane and nitrous oxide (Sap & Hellebrekers 1993), as has been our clinical experience. Optimizing conditions for EDI may decrease time to achieve EDI, reduce mechanically induced GI trauma, increase clinic throughput, increase clinician satisfaction and facilitate the development of less experienced endoscopists.
To the best of our knowledge, no study has directly compared the effect of maintenance of general anaesthesia using inhaled isoflurane with an IV infusion of propofol, on the ease of EDI.
The purpose of this study was to compare the effect of maintenance of general anaesthesia with inhaled isoflurane or propofol infusion on the ease of EDI in dogs undergoing upper GIT endoscopy. We hypothesized that the maintenance of general anaesthesia with propofol infusion would enable easier EDI during upper GIT endoscopy than maintenance of general anaesthesia with isoflurane.

Materials and methods
Ethical approval was obtained from the University of Liverpool Research Ethics Committee (VREC969). Dogs were recruited to the study following informed owner consent. Using previously published work (Donaldson et al. 1993;McFadzean et al. 2017), the EDI scoring system (Matz et al. 1991) and adjusting for use of the Mann-Whitney U test for comparison (Thrushfield 2018), a minimum of 10 dogs in each group were required to detect a difference of one point in EDI scoring system with 80% power and 95% confidence level.

Animals
A group of 22 dogs, all of whom were scheduled to undergo upper GIT endoscopy which included EDI, were prospectively recruited for the study by the attending medical clinician. Exclusion criteria included animals who were deemed unsuitable for the study anaesthesia protocol, dogs with body mass < 5 kg or dogs with a gastric foreign body who were undergoing gastric endoscopy only. Recruitment was ended when the data of 10 animals per group were available for analysis.

Study protocol
The study was designed as a prospective, randomized, partially blinded clinical trial in accordance with the consolidated standards of reporting trails (Table S1). The attending anaesthetist performed a clinical examination of the dog on the day of the scheduled procedure and confirmed their suitability for the study anaesthetic protocol. Animal age, sex, body mass, body condition score (BCS) and American Society of Anesthesiologists (ASA) physical status score (IeV) were recorded by the attending anaesthetist. Using an envelope system, dogs were assigned randomly to one of two groups: group ISO, general anaesthesia would be maintained with inhaled isoflurane (IsoFLo; Zoetis, UK) in 100% oxygen; and group PROP, anaesthesia would be maintained with a propofol (Propofol-Lipuro; Virbac, UK) infusion. Food, but not water, was withheld for 8e16 hours prior to induction of general anaesthesia. If required (in dogs undergoing colonoscopy examination following completion of upper GIT endoscopy), oral doses of a laxative agent (KleanPrep; Norgine, UK) and warm water enema were administered as directed by the attending medical clinician.
Senior lecturers and lecturers were deemed to be experienced at performing endoscopy. Residents and interns were deemed inexperienced. The same system was used to grade the experience of the attending anaesthetist.
Anaesthetic premedication included medetomidine (Sedator; Dechra, UK) 0.005 mg kg e1 and butorphanol (Dolorex; MSD Animal Health, UK) 0.2 mg kg e1 administered by intramuscular injection of the cervical muscles. The dog was returned to its assigned kennel, left undisturbed and monitored at a distance. After 20 minutes, a sedation score was assigned to the dog by the attending anaesthetist using a subjective sedation scale ranging 0e15 (no sedation to well sedated) by Gurney et al. (2009). A peripheral vein was cannulated using a cannula (Jelco; Smiths Medical, UK) of a gauge (23e18 gauge) appropriate for animal's size. Dogs were moved to the Effect of anaesthetic maintenance ME Tonge et al. endoscopy suite and pre-oxygenated via mask and a small animal circle (for dogs > 10 kg) or Mapleson-D (for dogs < 10 kg) breathing system. In both groups, general anaesthesia was induced by the attending anaesthetist with propofol (Propoflo Plus; Zoetis) via slow IV injection to enable endotracheal intubation. For group ISO, maintenance of general anaesthesia was with isoflurane vaporized in 100% oxygen delivered at 1.5% of fresh gas flow and adjusted by increments of 0.2% to maintain an adequate plane of anaesthesia. For group PROP, propofol was infused commencing at a rate of 0.2 mg kg e1 minute e1 and the infusion rate adjusted in increments of 0.05 mg kg e1 minute e1 to maintain an adequate plane of anaesthesia. For both groups, for the duration of the study period, a propofol bolus of 0.5 mg kg e1 IV was administered if the animal moved, swallowed or showed a sympathetic response to the procedure.
Following endotracheal intubation and after commencing administration of the ascribed anaesthetic maintenance agent, dogs were positioned in left lateral recumbency and instrumented to record heart rate, oscillometric arterial blood pressure, saturation of arterial haemoglobin with oxygen, respiratory rate, temperature and end-tidal carbon dioxide concentration via multiparameter monitor (Datex-Ohmeda; GE, UK) every 5 minutes during the study protocol period.
To blind the endoscopist to the maintenance agent used, screens covered all anaesthetic and infusion equipment. The endoscopist did not enter the endoscopy suite until the maintenance phase of general anaesthesia had commenced. Upper GIT endoscopy was performed prior to colonoscopy examination in those dogs scheduled for both procedures. The endoscope [5.7 mm diameter (dog < 8 kg) or 7.9 mm diameter (dog > 8 kg)] (TL-100; Karl Storz, UK) was introduced into the stomach via the oesophagus. Swallowing or movement in response to the passage of the endoscope through the oesophagus into the stomach, the presence of gastroesophageal reflux fluid within the oesophagus, spontaneous opening of the lower oesophageal sphincter and the presence of gastric antral movement were recorded (yes or no). The endoscope was manipulated to enable visualization of the pyloric sphincter and the presence of duodenal reflux and spontaneous opening of the pyloric sphincter were recorded by the anaesthetist as reported by the endoscopist (yes or no). The ease of EDI was then scored (EDI score) by the endoscopist using a previously reported four-point scale (Matz et al. 1991): 1) immediate entry with minimal manoeuvring required; 2) rapid entry with moderate manoeuvring; 3) difficult entry with multiple attempts required; and 4) no entry after 2 minutes. The duration of time (seconds) needed to achieve EDI was measured using a mobile phone stopwatch by the attending anaesthetist. This time began when the endoscopist reported the start of this manoeuvre and stopped when the endoscopist positively confirmed the duodenum had been intubated. At the point of confirmed EDI, the study period was deemed to be complete and for the remainder of the endoscopic examination and recovery period anaesthesia management was determined by the attending anaesthetist.
On completion of the procedure, dogs were disconnected from the anaesthetic machine and breathing circuit and moved to a quiet recovery room. Recovery from general anaesthesia was observed and a recovery score assigned ranging from 1 (poor; marked excitement or struggling and needed sedation) to 4 (excellent, smooth recovery) using a previously described subjective scale (Ambros et al. 2008) by the attending anaesthetist.

Statistical analysis
Missing animal data (age, sex, body mass, BCS and ASA classification) were retrieved from anaesthetic and clinical records made during the same hospitalization period for the same condition. If time to achieve EDI was greater than 120 seconds, an EDI score of 4 was automatically applied at the time of data analysis.
Statistical analysis was performed using computer software (IMB SPSS Statistics for Windows, version 27; IBM Corp, NY, USA). Continuous variable data were tested for normality using the Shapiro-Wilk test. Descriptive statistics for normally distributed continuous variables are reported as mean ± standard deviation (SD). Descriptive statistics for categorical data are reported as frequencies, and non-normally distributed or ordinal data are reported as median (interquartile range). Statistical significance was set at p < 0.05. Differences between groups for normally distributed variables were determined using the t test, for non-normally distributed continuous variables using the Mann-Whitney U test and for categorical Table 1 Age, sex, body mass, body condition score (1e9) and American Society of Anesthesiologists physical status score (ASA IeV) of 21 dogs undergoing upper gastrointestinal tract endoscopy during general anaesthesia. Anaesthetic premedication consisted of medetomidine (0.005 mg kg e1 ) and butorphanol (0.2 mg kg e1 ) administered intramuscularly. General anaesthesia was induced with propofol, given to effect, and maintained with 1.5% (vaporizer setting) isoflurane in 100% oxygen (group ISO) or 0.2 mg kg e1 minute e1 propofol (group PROP). Data are shown as mean ± standard deviation or median (interquartile range).

Variable
Group ISO (n ¼ 10) Effect of anaesthetic maintenance ME Tonge et al.
variables using Fisher's exact test. The effect of anaesthetist and endoscopist experience on the main outcomes were controlled for using ordinal (EDI score) and linear (time to achieve EDI) regression analysis.

Results
A total of 22 dogs were recruited to the study (group ISO: n ¼ 11, group PROP: n ¼ 11). For one dog, general anaesthetic maintenance agent was changed from isoflurane to propofol infusion without explanation following induction of general anaesthesia, and this case was excluded from the subsequent analysis. There were no significant differences between the groups for demographic variables including age (p ¼ 0.35), sex (p ¼ 0.64), body mass (p ¼ 0.86), BCS (p ¼ 0.83) or ASA classification (p ¼ 0.48) ( Table 1).
There were no significant differences between the two groups for the presence of gastroesophageal reflux fluid within the oesophagus (p ¼ 0.39), spontaneous opening of the lower oesophageal sphincter (p ¼ 0.67), gastric antral movement (p ¼ 0.06), the presence of gastroduodenal reflux (p ¼ 0.20) or spontaneous opening of the pyloric sphincter (p ¼ 0.39), assessed during endoscopy. Of the 11 dogs in group PROP, six responded to passage of the endoscope into the oesophagus compared with none in group ISO (p ¼ 0.01; Table 3).
Neither EDI score (Table 4) nor the time to achieve EDI were significantly different between the groups (p ¼ 0.39 and p ¼ 0.75, respectively; Table 5).
Of the six procedures with an EDI score of 4 (no entry after 2 minutes), two were in group ISO; time to achieve EDI (inexperienced endoscopist) was 315 seconds and (experienced endoscopist) 214 seconds; whereas four were in group PROP; time to achieve EDI (inexperienced endoscopist) 323 seconds and (experienced endoscopist) 200, 227 and 341 seconds (Table 6). When anaesthetist and endoscopist experience were controlled for using regression analysis, there was no significant association with either EDI score (anaesthetist experience p ¼ 0.45, endoscopist experience p ¼ 0.84) or time to achieve EDI (anaesthetist experience p ¼ 0.63, endoscopist experience p ¼ 0.62).

Discussion
In dogs undergoing upper GIT endoscopy, maintenance of general anaesthesia with inhaled isoflurane compared with propofol infusion did not affect the ease of EDI, as measured by EDI score or time to achieve EDI. All dogs were premedicated with medetomidine and butorphanol, and general anaesthesia was induced with propofol. However, significantly more dogs anaesthetized with a propofol infusion responded (swallowing, movement) to passage of the endoscope into the oesophagus than those in which anaesthesia was maintained with isoflurane (p ¼ 0.01).
Previous studies have shown the control of pyloric motility to be highly complex (Donaldson et al. 1993) with several hormones, neurotransmitters and receptors involved in its regulation (Donaldson et al. 1993;Salla et al. 2020). Gastroduodenal peristaltic activity results in changes in the size of the pyloric aperture, which can affect ease with which an endoscope passes into the duodenum (Donaldson et al. 1993) Excessive peristaltic activity of the gastric antrum may lead to retroflexion of the endoscope and ultimately failure to achieve EDI (Happe & Ivd 1983). Increased intestinal and duodenal tone may decrease the diameter of the pyloric lumen and thereby increase the difficulty of EDI (Donaldson et al. 1993). Gastric and small bowel motility may be influenced by general anaesthetic maintenance agent (Schurizek et al. 1989a;Schurizek et al. 1989b;Torjman et al. 2005;Boscan et al. 2014;Desmet et al. 2016) and so the choice of general anaesthetic maintenance agent may influence the ease of EDI.
Volatile anaesthetic agents and propofol cause smooth muscle relaxation and affect gastric and intestinal motility when used to maintain general anaesthesia in humans and other animals (Schurizek et al. 1989a;Schurizek et al. 1989b;Torjman et al. 2005;Dryn et al. 2018;Li et al. 2021).
Data relating to the effect of propofol on the GIT are mostly associated with human intensive care unit sedation, where short-term propofol administration prolonged oro-caecal transit time (Hammas et al. 1998). In mice, propofol inhibited gastric emptying and prolonged intestinal transit time Effect of anaesthetic maintenance ME Tonge et al. (Inada et al. 2004). In vitro propofol decreased spontaneous peristaltic activity of the human GIT, possibly owing to calcium channel blockade and reduction in muscarinic receptoreinduced contraction (Lee et al. 1999).
Maintenance of general anaesthesia with propofol boluses rather than a volatile anaesthetic agent has been shown to reduce the procedure time for upper GIT endoscopy, including EDI by 20e30 minutes (Sap & Hellebrekers 1993), as was our subjective, clinical impression. The results of the present study do not support this hypothesis; furthermore, our study showed no significant differences between the groups ISO and PROP for endoscopically derived observations associated with alteration in gastroduodenal and sphincter motility. Though not a statistically significant difference (p ¼ 0.06), antral movement was abolished in nine of 10 cases in group ISO compared with five of 11 cases in group PROP. This finding reflects that of studies during human general anaesthesia (Schurizek et al. 1989a;Schurizek et al. 1989b). Failure to achieve EDI has been attributed to an excessively flaccid gastric antrum (Happe & Ivd 1983). Different volatile anaesthetic agents may be considered to have similarly reductive effects on gastrointestinal motility (Boscan et al. 2014). Human gastric antral motility was almost abolished during enflurane and halothane anaesthesia (Schurizek et al. 1989a;Schurizek et al. 1989b), gastric emptying and intestinal motility in rats were reduced following a short period of isoflurane anaesthesia (Torjman et al. 2005) and in dogs sevoflurane anaesthesia reduced gastric and small intestinal propulsive motility (Boscan et al. 2014). The molecular mechanisms by which isoflurane administration causes alterations in GIT motility are yet to be fully elucidated but involve inhibition and disruption of muscarinic receptor G-protein coupling at several levels as well as alterations in plasma membrane transient receptor potential channels (Desmet et al. 2016;Dryn et al. 2018).
We believe the effect of anaesthetic maintenance agent on antral motility warrant further investigation and may have contributed to our impression of increased difficulty of EDI during isoflurane general anaesthesia.
Previous reports investigating the ease of EDI in dogs has focused on choice of anaesthetic premedication (Donaldson et al. 1993;Zoran 2001;Simpson & Hall 2005;Hall 2013;McFadzean et al. 2017;Salla et al. 2020). Butorphanol 0.2 mg kg e1 , the most commonly used dose at our institution, facilitates EDI and has been recommended for anaesthetic premedication prior to upper GIT endoscopy (McFadzean et al. 2017). Medetomidine at a dose of 0.005 mg kg e1 , rather than the recommended anaesthetic premedication dose of 0.01 mg kg e1 (Murrell & Hellebrekers 2005), was used to provide reliable preanaesthetic sedation and anxiolysis for a duration of approximately 1 hour. Its use in combination with propofol anaesthesia has also resulted in rapid and successful EDI (Sap & Hellebrekers 1993). Dexmedetomidine, the active enantiomer of medetomidine, significantly reduced gastrointestinal motility in mice (Li et al. 2021). The use of medetomidine for anaesthetic premedication may have contributed to ease of EDI through a reduction in gastrointestinal motility. There was no significant difference between groups in anaesthetic premedication sedation score (p ¼ 0.08); therefore, it seems improbable that premedication drugs affected the ease of EDI in the two groups.
Dogs in both groups received propofol for induction of general anaesthesia. As a clinical study, using client-owned animals, the induction of general anaesthesia using inhaled isoflurane was deemed ethically inappropriate and previous work comparing the use of isoflurane and propofol for maintenance of general anaesthesia used a similar protocol (Caines et al. 2014). We found no significant difference between either the induction volume of propofol (mL kg e1 ; p ¼ 0.08) or Effect of anaesthetic maintenance ME Tonge et al.
propofol bolus dose (mL kg e1 ; p ¼ 0.61) between the groups. Gastric emptying and therefore motility are affected to a similar degree following induction and maintenance of general anaesthesia with propofol alone or IV propofol followed by maintenance with a volatile anaesthetic agent (Bennett et al. 1994). We believe that the propofol bolus doses used in group ISO are unlikely to have led to a failure to detect a difference in the ease of EDI between groups. Significantly more dogs responded to passage of the endoscope through the oesophagus into the stomach in group PROP than in group ISO (p ¼ 0.01). A light plane of general anaesthesia (Simpson & Hall 2005) or, following anaesthetic premedication, isoflurane at 1 Â MAC (McFadzean et al. 2017) has been reported as adequate to maintain general anaesthesia and to facilitate upper GIT endoscopy. MAC isoflurane in unpremedicated dogs is approximately 1.2e1.3% (Steffey & Howland 1977;Ewing et al. 1993;Barletta et al. 2016). Anaesthetic premedication with medetomidine reduces MAC isoflurane by up to 50% (Ewing et al. 1993). While a propofol infusion rate of 0.2 mg kg e1 minute e1 produced a light plane of general anaesthesia in dogs (Aguiar et al. 2001) and accordingly was selected as the initial postinduction propofol infusion rate for this work. However, following anaesthetic premedication with medetomidine at 0.005 mg kg e1 , a propofol infusion rate of 0.4 mg kg e1 minute e1 reportedly led to difficulty maintaining a stable plane of general anaesthesia (Hammond & England 1994). In unpremedicated dogs, a propofol infusion rate of approximately 0.3 mg kg e1 minute e1 compared with an end-tidal isoflurane concentration of 0.73% was required to maintain general anaesthesia adequate for magnetic resonance imaging (Caines et al. 2014). There was no significant difference in anaesthetic premedication sedation score between groups ISO and PROP (p ¼ 0.08), so it seems improbable that different background sedation levels between groups caused the reported difference in animal response to passage of the endoscope. The attending anaesthetist scored the level of sedation following anaesthetic premedication (Gurney et al. 2009); however, this scoring system was adapted from a previously published system (Kuusela et al. 2000) with intra-and interrater variability. We believe that our initial isoflurane administration rate was relatively higher than that of propofol infusion and may have led to an increased plane of anaesthesia at the start of the procedure resulting in the differences in response rate.
Previous work suggested that differing anaesthetist's experience in the use of propofol infusion and volatile anaesthetic agents may affect their administration rates (Caines et al. 2014) and subsequently the conditions during upper GIT endoscopy and EDI. Endoscopist's experience likewise influences the ease of EDI, with more experienced endoscopists finding the procedure less difficult (Matz et al. 1991). As a busy teaching institution, we were unable to ensure that all procedures were performed by the same endoscopist and anaesthetist, and this represents a limitation of this study. However, we used regression analysis to control for anaesthetist and endoscopist experience and did not identify a significant association between anaesthetist or endoscopist experience and time for EDI.
Another limitation is the clinical, and hence heterogeneous, nature of our study population. The inherently smaller volume of the gastric antrum of smaller dogs may increase the difficulty of EDI (Matz et al. 1991); therefore, dogs < 5 kg in weight were not included in this study. The dose of volatile anaesthetic agent required to maintain general anaesthesia (Yamashita et al. 2009) and its effect on GIT smooth muscle contraction and ease of EDI varies with patient age. We found no significant differences between groups for age, sex, body mass, BCS and ASA physical status score and therefore believe such factors did not affect our findings.
In conclusion, when isoflurane and propofol were compared, the use of both anaesthetic maintenance agents allowed successful EDI in all cases. Neither agent led to significantly easier Table 4 Ease of duodenal intubation score (1e4) in 21 dogs undergoing upper gastrointestinal tract endoscopy under general anaesthesia maintained with either isoflurane in oxygen (group ISO) or propofol infusion (group PROP). For study, details see Table 1  Ease of duodenal intubation score: 1 ¼ immediate entry with minimal manoeuvring required, 2 ¼ rapid entry with moderate manoeuvring, 3 ¼ difficult entry with multiple attempts required and 4 ¼ no entry after 2 minutes (Matz et al. 1991); n: number. Ease of duodenal intubation score: 1 ¼ immediate entry with minimal manoeuvring required, 2 ¼ rapid entry with moderate manoeuvring, 3 ¼ difficult entry with multiple attempts required and 4 ¼ no entry after 2 minutes (Matz et al. 1991).
Effect of anaesthetic maintenance ME Tonge et al.
EDI or reduced time to achieve EDI. However, significantly fewer dogs responded to the passage of the endoscope through the oesophagus into the stomach during general anaesthesia maintained with isoflurane compared to propofol. Practitioners may wish to consider this finding when selecting a propofol infusion rate for upper GIT endoscopy and other similarly invasive procedures. The abolition of antral movement, which may increase the difficulty of EDI, was recorded in more cases where general anaesthesia was maintained with isoflurane compared to propofol, and this finding may warrant future investigation.