The risk of passive regurgitation during general anaesthesia in a population of referred dogs in the UK

Veterinary Anaesthesia and Analgesia, 2012, 39, 266 274 doi:10.1111/j.1467-2995.2011.00704.x RESEARCH PAPER The risk of passive regurgitation during general anaesthesia in a population of referred dogs in the UK Cecilia Lamata, Verity Loughton, Monie Jones, Hatim Alibhai, Elizabeth Armitage-Chan, Karen Walsh & David Brodbelt Department of Veterinary Clinical Sciences, Royal Veterinary College, North Mymms, UK Correspondence: David Brodbelt, Department of Veterinary Clinical Sciences, Royal Veterinary College, Hawkshead Lane, North Mymms, AL9 7TA UK. E-mail: dbrodbelt@rvc.ac.uk Abstract Objective To evaluate the risk of passive regurgitation during anaesthesia, and to identify major factors associated with this in dogs attending the Queen Mother Hospital for Animals (QMHA), the Royal Veterinary College. Study design A case-control study nested within the cohort of dogs undergoing anaesthesia with inhalation agents. Animal population All dogs undergoing general anaesthesia at the referral hospital between October 2006 and September 2008 (4271 cases). Methods All dogs anaesthetized at the QMHA during the study period were included. Regurgitating cases were defined as dogs for which reflux material was observed at the external nares or in the mouth, either during anaesthesia or before return to normal consciousness immediately after general anaesthesia. The risk of regurgitation was estimated and risk factors for regurgitation were evaluated with multivariable logistic regression (p < 0.05). Results The overall risk of regurgitation was 0.96% (41 cases out of 4271 anaesthetics, 95% confidence interval [95% CI] 0.67 1.25%). Exclusion of animals where pre-existing disease was considered a contributing factor to regurgitation (n = 14) resulted in a risk of passive regurgitation of 0.63% (27 cases of 4257 anaesthetics, 95% CI 0.40 0.87%). In the multivariable logistic regression model, procedure and patient weight were significantly associated with regurgitation. Dogs undergoing orthopaedic surgery were 26.7 times more likely to regurgitate compared to dogs undergoing only diagnostic procedures. Dogs weighing more than 40 kg were approximately five times more likely to regurgitate than those weighing <20 kg. Conclusions and clinical relevance This study highlights the rare but important occurrence of perioperative regurgitation and identifies that dogs undergoing orthopaedic procedures, and those weighing more than 40 kg, are particularly at risk. Further work is required to evaluate the reasons for these observations. Keywords anaesthesia, dog, regurgitation, risk factors. Introduction Regurgitation is defined as the passive discharge of gastric or oesophageal fluid from the mouth or nose. Complications include aspiration pneumonia, oesophagitis and oesophageal stricture. Regurgitation has been reported to occur in between 0.42% and 5.5% of anaesthetized dogs (Galatos & Raptopoulos 1995a,b; Wilson et al. 2005) but there are few published data regarding the causes. In comparison, several studies have investigated gastrooesophageal reflux (GOR). GOR is a more common but less visible complication of general anaesthesia, 266

and occurs when refluxate enters the oesophagus but does not ascend into the pharynx. It is diagnosed by measurement of the ph in the lower oesophagus. Reported incidence of GOR in clinical trials and other studies ranges from 16.3% to 55% (Galatos & Raptopoulos 1995a,b; Raptopoulos & Galatos 1997; Wilson et al. 2006a, 2007; Wilson & Evans 2007). Studies of risk factors have generally focused on GOR but factors associated with reflux are relevant to regurgitation as by definition GOR occurs in all cases of regurgitation. A number of studies have suggested type and duration of surgical procedure, length of fasting time and pre-anaesthetic drug selection as potential risk factors for GOR in small animals (Hashim & Waterman 1993; Galatos & Raptopoulos 1995a,b; Wilson et al. 2007). Galatos & Raptopoulos (1995b) reported that increased duration of anaesthesia and intra-abdominal surgery (compared to non-abdominal surgeries) were associated with significantly increased GOR incidence, whilst reduced fasting time was associated with reduced risk of regurgitation. Regarding anaesthetic agent selection, diazepam premedication was associated with a significant reduction in GOR episodes compared to atropine and propionyl promazine premedication (Galatos & Raptopoulos 1995a) and morphine in combination with acepromazine significantly increased GOR risk compared to the use of acepromazine alone (Wilson et al. 2005). Propofol significantly decreased lower oesophageal sphincter pressure in dogs when compared with thiopental, suggesting a theoretical increased risk of GOR and regurgitation (Waterman & Hashim 1992; Raptopoulos & Galatos 1997). Although risk factors for GOR and regurgitation are likely to be related, additional work is required to confirm this and to evaluate the risks specific to passive regurgitation across a broad range of procedures and a wide population of dogs. Hence, the aims of this study were to document the risk of passive regurgitation in a large population of anaesthetized dogs attending a referral hospital and to identify major risk factors for regurgitation in these animals. It was hypothesised that patient age, weight, duration of fasting, procedure performed, and specific anaesthetic agents would be associated with increased risk of peri-anaesthetic regurgitation. Materials and methods A case-control study nested within the prospective cohort of dogs undergoing inhalational general anaesthesia at the Queen Mother Hospital for Animals (QMHA), the Royal Veterinary College was undertaken between 19th October 2006 and 2nd September 2008. All dogs undergoing general anaesthesia with endotracheal intubation and maintenance with an inhalation agent during the study period were observed for evidence of regurgitation. Cases of regurgitation were defined as those where the anaesthetist observed the presence of gastric contents in the mouth or nostrils, either during anaesthesia or before return to full consciousness after general anaesthesia (before or after endotracheal extubation). For every case identified, four control dogs (those in which regurgitation had not been observed during general anaesthesia or prior to return of consciousness) were selected randomly (Dohoo et al. 1982), from the same cohort, using a random number generator (Microsoft Excel 2007; Microsoft Corp., WA, USA). Dogs considered to have an increased risk of regurgitation due to the presence of certain pre-existing diseases (e.g. tetanus, hiatal hernia), obstruction of the upper gastrointestinal tract (i.e. oesophagus, stomach and small intestines), the administration of non-anaesthetic drugs which could affect gastrointestinal tract motility (e.g. metoclopramide) and any dog that had vomited or regurgitated within 24 hours prior to anaesthesia were excluded from the analysis of risk factors. Dogs were excluded from the study if the clinical notes were unavailable (due to inability to determine if they had a pre-existing condition or other reason that required exclusion) or if they did not recover from general anaesthesia due to death or euthanasia unrelated to regurgitation. Only one general anaesthetic was considered per dog in the analysis, and if a dog had undergone multiple anaesthetics during which it regurgitated on at least one occasion, then only the first anaesthetic where regurgitation was observed was evaluated. After dogs had been classified as regurgitating or not regurgitating, hospital records were used to collect the additional information used in this study. Information collected for each patient comprised of age, sex, breed, American Society of Anaesthesiologist (ASA) grade, weight, time the dog had been starved before anaesthesia, preoperative temperature, lowest intraoperative temperature, procedure performed and duration of general anaesthesia. All drugs given prior to anaesthesia and during anaesthesia were recorded. In addition, for cases of regurgitation, the colour, ph and time of reflux, and subsequent patient management were noted. Ó 2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 266 274 267

Statistical analysis Data were entered into an Excel spreadsheet (Microsoft Excel 2007) and exported for analysis to SPSS statistics V.17 (PASW version: SPSS Inc., IL, USA) for windows. Risk factors for regurgitation were evaluated individually for their association with regurgitation in the univariable analysis and then a multivariable model was constructed. Categorical data were analysed with the Chi squared or Fisher s Exact tests. Continuous data were assessed graphically for normality and in the univariable analysis were analysed using the independentsamples T-test or the Mann Whitney U test as appropriate (Kirkwood & Sterne 2003). Additionally, continuous data were assessed categorically based on biologically relevant groupings and assessed as described above. Variables significant at the 20% level and biologically important factors were taken forward for evaluation in the multivariable logistic regression analysis using a manual forward selection approach (Dohoo et al. 2003). The multivariable model was checked for two-way interactions and model fit was assessed with the Hosmer Lemeshow test and delta beta and delta deviance diagnostic statistics (Hosmer & Lemeshow 2000). Statistical significance was set at the 5% level. Results During the study period 4271 inhalation anaesthetics were administered to dogs and 41 cases of regurgitation were identified, giving an overall risk of regurgitation of 0.96% (95% confidence interval 0.67 1.25%). Fourteen dogs were considered to have other contributing factors for regurgitation, or were excluded for other reasons (including the file was not available [three dogs], the dog had a disease or obstruction of the upper gastrointestinal tract [three dogs], the dog had vomited within 24 hours prior to premedication [three dogs] or the dog had another excludable condition or medication [five dogs]). After these were removed from analysis, the regurgitation risk was 0.63% (27 of 4257 dogs, 95% confidence interval 0.40 0.87%). Figure 1 shows the timing of regurgitation relative to anaesthetic induction. Two dogs vomited after premedication and then went on to regurgitate during general anaesthesia. In one dog, regurgitation was observed to have occurred at the time of Number of dogs 6 4 2 0 <=30 31 90 91 150 151 210 211 270 271+ Time of first regurgitation (minutes) Figure 1 The timing after induction of anaesthesia of first regurgitations in 27 cases of perioperative regurgitation. endotracheal intubation. This anaesthetic was the dog s second of the day (although the first where regurgitation was observed). One dog regurgitated on extubation and then multiple times during recovery. This was also the dog s second anaesthetic of the day. Another dog regurgitated on extubation and a further two shortly after extubation. In 20 dogs regurgitation was observed once during general anaesthesia, in three, twice and one, three times. For two dogs it was not noted when regurgitation occurred or how many times they regurgitated. Regarding possible contributory factors, two dogs regurgitated within 5 minutes of intravenous antibiotic administration, two dogs regurgitated within 5 minutes of being moved and one dog regurgitated within 5 minutes of the endotracheal tube being changed and morphine being given. In fourteen animals the refluxate ph was measured. Three dogs had a ph <5, eight dogs had a ph of 5, two dogs had a ph of 5.5, and one dog had a ph of 8. Two animals refluxate contained blood, and two were noted to contain food. Of the animals that had reflux, colour was recorded in 18 cases and the most common colour was brown (66.7%, 12 of 18 dogs). Other colours recorded were light or white (16.7%, three of 18), clear (5.6%, one of 18 dogs) and yellow (11.0%, two of 18 dogs). Post-regurgitation management included systemic treatment with ranitidine, metoclopramide or omeprazole and suctioning and flushing of the oro-pharynx with sterile saline. For evaluation of risk factors for regurgitation, 185 potential control dogs were identified and after assessment, 23 were excluded due to lack of 268 Ó 2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 266 274

available clinical notes (14 of 23 dogs) or insufficient information in the clinical notes to make an assessment (nine of 23 dogs). Thirty six were excluded due to their underlying medical condition and/or procedure. Of these latter excluded controls, five were excluded with upper GIT foreign bodies, four diagnosed with inflammatory bowel disease, two with an upper GI mass, five with vomiting within the preceding 24 hours, two with gastric dilatation volvulus (GDV), one died under anaesthesia, one was euthanased, one received omazeprole, two had previously regurgitated and their prior anaesthetic was recorded in the case group, one had been selected as a control for a preceding anaesthetic, one did not receive an inhalation anaesthetic and 11 presented with other upper GI disease. This resulted in 27 regurgitation cases being compared to 126 control anaesthetics. In the univariable analysis the sex distribution between case and control groups was not found to be significantly different (p = 0.376). The breeds which had the largest number of regurgitation cases were Labrador Retrievers, German Shepherd Dogs (GSD) and crossbreeds (Table 1). Labradors, GSDs and crossbreds made up between 8 9% of the controls each, yet between 11 26% of the cases (p = 0.005). Thirteen percent (17 of 126 dogs) of the controls and 7% of cases (two of 126, p = 0.529) were classified as brachycephalic. Case dogs were significantly heavier than controls (p = 0.006, Table 2). There was a non-significant trend to higher ASA grade in the control dogs (control dogs ASA grade median score 2 [range 1 Table 1 Breeds most frequently presenting for anaesthesia. The number (%) are reported for individual breeds if at least five were anaesthetized Breed Case dogs (number, %) Control dogs (number, %) Labrador 7 (26) 10 (8) Crossbreed 5 (18) 11 (9) German shepherd dog 3 (11) 10 (8) Jack Russell terrier 0 12 (10) Boxer 0 7 (6) Staffordshire Bull terrier 1 (4) 6 (5) Cavalier king Charles spaniel 1 (4) 4 (3) Doberman 1 (4) 4 (3) Golden retriever 0 5 (4) West highland white terrier 0 5 (4) Other Breeds 9 (33) 52 (40) Total 27 126 4], cases ASA median 1 [1 4], p = 0.056). There was no significant difference between cases and controls in the duration of pre-anaesthetic fasting (starved for more or <12 hours, p = 0.693). The use of full agonist opioids for premedication was significantly increased in cases of regurgitation (p = 0.040), case animals receiving a higher proportion of full agonist opioids administered compared to partial agonists and agonist-antagonists (Table 3). There were insufficient numbers to observe a dose effect. In relation to other drugs given, the administration of drugs by the epidural route was significantly associated with the presence of regurgitation (p = 0.039), morphine epidurals (with or without local anaesthetics) being more common in case dogs (Table 5). Administration of non-steroidal anti-inflammatory drugs (NSAIDs) was also associated with presence of regurgitation (p = 0.018). The major procedure group associated with regurgitation was orthopaedic procedures (Table 4). In the final multivariable logistic regression model (Table 6), only procedure type and patient weight remained significantly associated with regurgitation. Orthopaedic patients were over 26 times more likely to regurgitate compared to those undergoing diagnostic procedures alone (OR = 26.7, 95% CI 3.2 223.7, p = 0.004). Heavier dogs were also at increased odds of regurgitation (>40 kg relative to 0 20 kg, OR 5.0, 95% CI 1.2 20.3, Table 6). No interaction between the final model variables was detected. The model fit was considered good based on the Hosmer Lemeshow test (p = 0.797) and evaluation of delta-beta and delta-deviance diagnostic statistics did not detect any large values (only 1 delta-beta >1.0 and 1 delta deviance >2.5, Hosmer & Lemeshow 2000). To assess if the control group was representative of the population anaesthetized at the Queen Mother Hospital for Animals (QMHA) during the study period, age and whether the animal was neutered or not were compared for control dogs and the underlying cohort of dogs from which they were randomly chosen. Neuter status was not significantly different between the two groups (p = 0.072). Further, the mean age (months ± SD) of control dogs (72 ± 45 months) was similar to the overall study cohort (67 ± 44 months). The hospital clinical specialities (soft tissue surgery, orthopaedics, medicine etc.) to which the dogs were admitted were not significantly different between the two groups (p = 0.88). Ó 2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 266 274 269

Table 2 Univariable associations of continuous variables with regurgitation. Mean and standard deviation (SD) are reported unless otherwise stated (IQR represents interquartile range) Risk factor Cases mean (SD) Controls mean (SD) p value Weight (kg) 33.1 (18.5) n =27 Age (years) median (IQR) 4 (1 6) n =26 Total duration of anaesthesia (minutes) 227 (89) n =26 Duration of anaesthesia in minutes: Cases (prior to regurgitation) Controls (total anaesthetic time) 137 (87) n =25 Fasting time (hours) 15 (3) n =15 Body temperature before anaesthesia ( C) 38.1 (0.6) n =23 Lowest body temperature during anaesthesia ( C) 35.4 (1.3) n =21 24.2 (13.9) n = 126 median (IQR) 5 (2 8) n = 126 140 (74) n = 126 140 (74) n = 126 16 (4) n = 108 38.4 (0.7) n = 101 35.4 (2.0) n = 111 0.006 0.119 <0.001 0.845 0.480 0.030 0.925 Table 3 Univariable associations of premedication, induction and maintainence agents with the presence of regurgitation in dogs. Percentage represents proportion within case or control group Factor Cases (%) Controls (%) p Premedication (non-opioids) Acepromazine 19 (70.4) 60 (47.6) 0.104 Other* 2 (7.4) 23 (18.3) No premedication 6 (22.2) 43 (34.1) Premedication (opioids) Morphine 17 (63.0) 47 (37.3) 0.040 Methadone 6 (22.2) 37 (29.4) Meperidine (Pethidine) 3 (11.1) 9 (7.1) Butorphanol or Buprenorphine 1 (3.7) 29 (23.0) None 0 4 (3.2) Induction agent Thiopental combinations 11 (40.7) 29 (23) 0.232 Propofol combinations 16 (59.3) 88 (69.8) Alfaxalone combinations 0 5 (4) Other 0 4 (3.2) Anaesthesia Isoflurane 23 (85.2) 109 (86.5) 0.375 Maintenance agent Halothane 2 (7.4) 2 (1.6) Isoflurane + Halothane, Sevoflurane or Desflurane 2 (7.4) 13 (10.3) Unknown 0 2 (1.6) Nitrous oxide Used 10 (37) 33 (26.2) 0.367 Not used 17 (63) 93 (73.8) *Other non opioid premedication included medetomidine, diazepam, midazolam and ketamine. Discussion This study documents a rare but clinically significant complication of perioperative regurgitation (0.96% overall, and 0.63% when animals with certain pre-existing conditions were excluded). Given this frequency, and the finding that some dogs regurgitated on recovery, endotracheal intubation, with the cuff appropriately inflated until the animal regains full consciousness and laryngeal 270 Ó 2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 266 274

Table 4 Univariable association of the procedures performed and the presence of regurgitation in dogs. Percentage represents proportion within case or control group Factor Cases (%) Controls (%) p value Main Procedure performed Orthopaedic 15 (55.6) 24 (19) <0.001 Soft tissue 6 (22.2) 44 (34.9) Other surgery 5 (18.5) 14 (11.1) Diagnostic procedures only 1 (3.7) 44 (34.9 Diagnostic procedure as well as main procedure No 11 (40.7) 76 (60.3) 0.062 Yes 16 (59.3) 50 (39.7) More than one procedure No 8 (29.6) 56 (44.4) 0.23 Yes 19 (70.4) 70 (55.5) Table 5 Univariable associations of other drugs administered and ventilation method used during general anaesthesia and the presence of regurgitation in dogs. Percentage represents proportion within case or control group Factor Cases (%) Controls (%) p Intravenous antibiotics Yes 16 (59.3) 61 (48.4) 0.417 No 11 (40.7) 65 (51.6) NSAIDS Carprofen 16 (59.3) 40 (31.7) 0.008 Meloxicam 3 (11.1) 7 (5.6) Not administered 8 (29.6) 79 (62.7) Intraoperative opioids (prior to regurgitation) Morphine 11 (40.7) 27 (21.4) 0.102 Methadone 4 (14.8) 12 (9.5) Other 0 2 (1.6) None 12 (44.4) 85 (67.5) Epidural administered Morphine (±ropivicaine) 5 (18.5) 6 (4.8) 0.039 Ropivicaine only 0 1 (0.8) Not administered 22 (81.5) 119 (94.4) Anticholinergic administered Yes 3 (11.1) 16 (12.7) 1.00 No 24 (88.9) 110 (87.3) Continuous rate infusion Fentanyl 3 (11.1) 12 (9.5) 0.342 Other 1 (3.7) 1 (0.8) Not administered 23 (16.9) 113 (83.1) IPPV No IPPV 20 (74.1) 67 (53.2) 0.099 Spontaneous and IPPV 6 (22.2) 54 (42.9) IPPV only 1 (3.7) 5 (4) function, is a valuable tool to protect against aspiration of reflux, although leaving the tube in place this length of time may result in other complications. In the multivariable analysis, procedure (orthopaedic surgery) and patient weight (>40 kg) were associated with increased risk of regurgitation. Considering these risk factors, particular vigilance when anaesthetizing these patients, and rapid treatment if regurgitation is observed, could help reduce serious long term complications. The risk of regurgitation in dogs without pre-existing predisposing disease (0.63%) was similar to that reported in previously published work. Some work reported risks of 0.42 0.74% (Galatos & Raptopoulos 1995a,b), whilst other reports were of a substantially greater (risk of regurgitation 5.5%, Wilson et al. 2005). The variation in incidence across these studies was likely to reflect differences in procedures performed, premedication and anaesthetic drugs and doses used and the dog populations studied. Under general anaesthesia gastro-oesophageal reflux (GOR) occurs when gastric contents enters only the oesophagus and can cause the same Ó 2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 266 274 271

Table 6 Final multivariable logistic regression model of factors associated with regurgitation in dogs undergoing general anaesthesia. Odds ratios (OR) and 95% confidence intervals (95% CI) are reported Factor Cases (number, % total) Controls (number, % total) OR 95% CI p value Procedure Orthopaedic 15 (55.6) 24 (19) 26.7 3.2 223.7 0.004 Soft tissue 6 (22.2) 44 (34.9) 6.2 0.7 54.8 Other surgery 5 (18.5) 14 (11.1) 18.5 1.9 181.9 Diagnostic 1 (3.7) 44 (34.9) 1 (reference) procedures only Weight 0 20 kg 6 (22.2) 51 (40.5) 1 (reference) 0.045 20.4 40 kg 13 (48.1) 63 (50.0) 1.3 0.4 4.2 40.5 max 8 (29.6) 12 (9.5) 5.0 1.2 20.3 complications as described for regurgitation with the exception of aspiration pneumonia (Raptopoulos 1995). Reflux into the oesophagus can consist of gastric acid, bile or a mixture of both. Bile can work synergistically with gastric acid causing greater inflammatory injury to the mucosa than either bile or acid alone (Oh et al. 2006). GOR has been studied in greater depth in previous studies in part, due to the fact it occurs more commonly than regurgitation. The risks previously reported for GOR of 16.3 17.4% (Galatos & Raptopoulos 1995a,b) to 27 60% (Wilson et al. 2005, 2006a) were higher than those identified for regurgitation, and it is likely that in the current study a number of dogs experienced GOR without regurgitation. In the current study, regurgitation was observed at all stages during anaesthesia and into the recovery period. In contrast, where GOR was measured, this tended to occur sooner after induction of anaesthesia. Although Wilson et al. (2005) recorded the occurrence of GOR throughout anaesthesia, the median time to GOR from induction of anaesthesia was 10 minutes. Galatos & Raptopoulos (1995b) also documented that most GOR events occurred soon after induction of anaesthesia. The trend to earlier GOR in comparison to the timing of regurgitation observed in the current study may reflect that a number of GOR events occur during anaesthesia prior to regurgitation. Wilson et al. (2005) noted that dogs that had GOR soon after induction of anaesthesia were more likely to go on to then regurgitate, suggesting that GOR may well occur earlier than regurgitation. Further, the ph of regurgitant material documented here was consistent with the ph of reflux previously documented (Savvas et al. 2009). It is interesting to note in the latter study, longer starvation time was associated with lower reflux ph (Savvas et al. 2009). Risk factors previously associated with GOR are likely to be relevant to regurgitation also. Risk factors for regurgitation per se have been studied less than factors associated with GOR, again in part due to the relative rareness of regurgitation. Previous studies relating to GOR have reported duration of preoperative fasting, premedication with atropine, premedication with pure agonist opioids, age, type of surgical procedure and positioning to be associated with increased risk of GOR (Galatos & Raptopoulos 1995; Hashim & Waterman 1993; Wilson et al.2005, 2007). The risk factors identified for regurgitation in the current study were animal weight and type of procedure undertaken. Case dogs weighed significantly more than animals in the control group. No study in dogs so far has shown increasing weight to be associated with increased risk of regurgitation or GOR. Heavier dogs may have greater distortion of abdominal and thoracic cavities and this could affect gastric pressure and diaphragm function and alter the risk of regurgitation. Obesity has been linked with gastro-oesophageal reflux disease in human medicine (Anand & Katz 2010). The association with weight may also partly reflect breed, and in the univariable analysis there was a statistically significant trend for Labradors and GSDs to be classified as cases more than controls, though this breed association was not retained in the final multivariable model. Breeds most frequently classified as over 40 kg included these two breeds as well as a number of the giant breeds, suggesting a combination of obesity and larger size may have played a role. 272 Ó 2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 266 274

Dogs undergoing orthopaedic surgery were found to be at much greater risk of regurgitation than those undergoing diagnostic procedures and other surgeries. Previous work reported that intra-abdominal surgery increased the risk of GOR (Galatos & Raptopoulos 1995a), but no other type of surgery has been found to be associated with increased risk. That the association with soft tissue procedures was weaker in the current study is likely to reflect the spectrum of soft tissue cases seen at the QMHA. Though intra-abdominal surgery was performed, a large number of superficial surgical explorations, non-abdominal tumour resections and thoracic, head and neck surgeries were also undertaken. Orthopaedic surgery often involves pre-surgical diagnostic procedures under the same general anaesthetic. Changes of dogs position and transfer between hospital areas may result in changes in abdominal pressure, increased gastric pressure and increased risk of regurgitation. The position of the animal at the time of regurgitation would have been interesting to evaluate, as one study showed that dogs experiencing reflux in sternal position weighed significantly more than dogs that did not (Pratschke et al. 2001). The finding that dogs undergoing orthopaedic surgery were more likely to regurgitate may be a reflection of the type and dose of opioids used. Dogs undergoing orthopaedic procedures were premedicated with full opioid agonists in 87.5% of controls and 100% of cases compared to 73.5% of controls and 90.9% of cases for other procedures. In addition, over 80% of orthopaedic procedures received a premedication dose of morphine or methadone of 0.3 mg kg )1 or above, whilst only approximately 50% of non-orthopaedic procedures had a dose of 0.3 mg kg )1 or greater. None of these associations were statistically significant, but they support the potential contributory role of full opioid agonists in perioperative regurgitation that has been reported previously (Wilson et al. 2005). The lack of an association between any anaesthetic agent and regurgitation in the final multivariable model was likely to reflect the strong correlation between procedure type and anaesthetic drugs administered, with specific drugs routinely being administered for specific procedures (as discussed in part above). In the univariable analysis premedication with opioids, epidural anaesthesia and NSAID administration were all associated with regurgitation. Premedication with morphine has been shown to increase GOR risk previously (Wilson et al. 2005), though an association with epidural analgesia or administration of NSAIDs has not previously been reported. In the current study, the correlation between type of procedure and drugs administered made it difficult to separate these factors. It is uncertain whether the association reported with procedure relates to the procedure and animal positioning and handling itself, the drugs these procedures frequently received or some other unidentified factor. The study was limited by its reliance on preexisting records resulting in some missing data. However, the major risk factors were generally completely recorded. Ideally long-term follow-up would be available to assess the consequence of these regurgitating episodes; however such information was not available. Although the definition of regurgitation was considered for all patients anaesthetized during the study period, it is likely that some cases of regurgitation may not have been identified, particularly if regurgitation occurred after anaesthetic recovery. The current study therefore reports factors associated with regurgitation occurring during or soon after anaesthesia only. Further, in a case-control study, the selection of the control population is often a concern (Dohoo et al. 2003). However, due to the presence of the underlying sampling frame (all patients receiving an inhalation anaesthetic at the QMHA) and the randomisation of controls based on this sampling frame, it is unlikely selection bias of the controls was a major issue. The authors excluded a number of potential controls that had pre-existing conditions or had received specific non-anaesthetic medication considered to be at increased risk of regurgitation as it was the aim of the study to identify major patient and anaesthetic management factors associated with regurgitation. It is interesting to note of 36 controls excluded based on pre-existing disease, none regurgitated during the anaesthetic episode. Finally, this study had only sufficient power to detect major risk factors, and there may well have been other factors present that were not statistically associated with the outcome. In conclusion, this study highlighted a rare but clinically relevant frequency of regurgitation. Weight, and type of procedure were the main factors associated with regurgitation. Premedication with opioids, intraoperative NSAIDs and epidural use were also identified as factors in the univariable analysis but were highly correlated with the main Ó 2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 266 274 273

procedure type. Greater attention to possible regurgitation for animals undergoing orthopaedic procedures and larger dogs is recommended. Further work is merited to explore the associations between procedures and drugs administered with regurgitation. Acknowledgements The authors would like to thank the entire anaesthesia team of the QMHA for identifying and recording regurgitation cases as well as their surgical, medical and nursing colleagues for help with this study. References Anand G, Katz PO (2010) Gastroesophageal reflux disease and obesity. Gastroenterol Clin North Am 39, 39 46. Dohoo I, Martin W, Stryhn H (2003) Veterinary Epidemiologic Research. AVC Inc., Charlottetown. pp. 346 349. Dohoo I, Martin W, Stryhn H (2010) Veterinary Epidemiologic Research (2nd edn). AVC Inc., Charlottetown, 185 194. Galatos AD, Raptopoulos D (1995a) Gastro-oesophageal reflux during anaesthesia in the dog: the effect of preoperative fasting and premedication. Vet Rec 137, 479 483. Galatos AD, Raptopoulos D (1995b) Gastro-oesophageal reflux during anaesthesia in the dog: the effect of age, positioning and type of surgical procedure. Vet Rec 137, 513 516. Hashim MA, Waterman AE (1993) Effect of acepromazine, pethidine and atropine premedication on lower oesophageal sphincter pressure and barrier pressure in anaesthetized cats. Vet Rec 133, 158 160. Hosmer DW, Lemeshow S (2000) Applied Logistic Regression. John Wiley, New York, USA. pp. 91 142. Kirkwood BR, Sterne JAC (2003) Essentials of Medical Statistics (2nd edn). Blackwell Science, Oxford, UK, 165 173. Oh DS, Hagen JA, Fein M et al. (2006) The impact of reflux composition on mucosal injury and esophageal function. J Gastrointest Surg 10, 787 796. Pratschke KM, Bellenger CR, McAllister H et al. (2001) Barrier pressure at the gastroesophageal junction in anesthetized dogs. Am J Vet Res 62, 1068 1072. Raptopoulos D (1995) Post anaesthetic reflux oesophagitis in dogs and cats. J Vet Anaesth 22, 6 8. Raptopoulos D, Galatos AD (1997) Gastro-oesophageal reflux during anaesthesia induced with either thiopentone or propofol in the dog. J Vet Anaesth 24, 20 22. Savvas I, Rallis T, Raptopoulos D (2009) The effect of pre-anaesthetic fasting time and type of food on gastric content volume and acidity in dogs. Vet Anaesth Analg 36, 539 546. Waterman AE, Hashim MA (1992) Effects of thiopentone and propofol on lower oesophageal sphincter and barrier pressure in the dog. J Small Anim Pract 33, 530 533. Wilson DV, Evans AT (2007) The effect of topical treatment on esophageal ph during acid reflux in dogs. Vet Anaesth Analg 34, 339 343. Wilson DV, Evans AT, Miller R (2005) Effects of preanesthetic administration of morphine on gastroesophageal reflux and regurgitation during anesthesia in dogs. Am J Vet Res 66, 386 390. Wilson DV, Boruta DT, Evans AT (2006a) Influence of halothane, isoflurane and sevoflurane on gastroesophageal reflux during anaesthesia in dog. Am J Vet Res 67, 1821 1825. Wilson DV, Evans AT, Mauer WA (2007) Pre-anaesthetic meperidine: associated vomiting and gastroesophageal reflux during the subsequent anesthetic in dogs. Vet Anaesth Analg 34, 15 22. Received 20 January 2011; accepted 17 July 2011. 274 Ó 2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 266 274