A retrospective study of perianesthetic and sedation deaths in dogs and cats submitted to Canadian veterinary diagnostic laboratories

Abstract

Background

Perianesthetic death or sedation death in companion animals is an infrequent but devastating complication. Few studies have investigated the pathology associated with these deaths.

Objective

To determine clinical features and postmortem findings for submissions to multiple Canadian diagnostic laboratories from perianesthetic/sedation deaths in dogs and cats.

Animals and procedure

Laboratory Information Management Systems were retrospectively reviewed for cases of perianesthetic/sedation death in dogs and cats. Inclusion criteria were: i) whole-body submissions and ii) death within 7 d after the procedure.

Results

Pathology reports determined the cause of death in 43% of dogs (73/168) and 34% of cats (50/147). Spay/neuter surgeries were the most common procedure for which animals were submitted (dogs: n = 72, 31%; cats: n = 111, 58%). The American Society of Anesthesiologists physical status in these animals was low (ASA status I or II) in 94% of dogs (68/72) and 93% of cats (103/111). Clinical history was considered complete in 60.3% of cases (242/401).

Conclusion and clinical relevance

These results had similar trends to those in previous studies that identified an important proportion of submissions for perianesthetic/sedation deaths lacked significant lesions to explain the cause of death. This study also identified spay/neuter procedures were involved in the largest proportion of submissions, despite their low pre-anesthetic/sedation risk.

INTRODUCTION

Perianesthetic death is any death occurring during or shortly after an anesthetic procedure. In dogs, the incidence of perianesthetic death is 0.10 to 0.23%, or 1 in 2000 to 1 in 588 anesthetic procedures (1–3). By comparison, the incidence of perianesthetic death in human medicine is 0.001 to 0.01% (1 in 100 000 to 1 in 10 000 anesthetic procedures) (4–7). To date, most research into perianesthetic death in veterinary medicine has investigated clinical risk factors (1,3,8). These studies have provided valuable information by identifying procedures that increase risk of death, leading to changes in practice standards. Interestingly, very few studies have investigated the pathology associated with cases of perianesthetic or sedation death. The largest study of perianesthetic deaths to date, the Confidential Enquiry into Perioperative Small Animal Fatalities (CEPSAF), had postmortem examination completed in only 10% of the cases (1).

Two studies specifically investigated pathologic findings associated with perianesthetic death (9,10). The first, published in 2011, reported postmortem findings of perianesthetic death in cats at 2 high-volume spay-neuter clinics in New York state (USA). Significant preexisting disease was identified in 33% of the cats (18/54); however, 63% of cats (34/54) lacked any significant gross or histologic lesions to explain cause of death (COD) (9). The second study investigated submissions of perianesthetic deaths at the Animal Health Laboratory in Guelph (Ontario) (10). All species were included in that study. Lesions of significant natural disease were identified in 59% of the animals (130/221), but no significant lesions to explain COD were present in 36% (80/221).

Studies investigating perianesthetic/sedation death in western and central Canadian laboratories have not been reported. This retrospective study investigated submissions from perianesthetic/sedation deaths in dogs and cats to 4 Canadian veterinary diagnostic laboratories. Information related to clinical history and postmortem findings and details on cases of perianesthetic death/sedation during spay/neuter procedures are summarized.

MATERIALS AND METHODS

We conducted a retrospective study investigating the postmortem findings of perianesthetic death or death following sedation in dogs and cats, using data from 4 laboratories in western and central Canada. Inclusion criteria were i) whole-body submissions of a dog or cat for postmortem examination after a perianesthetic/sedation death, and ii) death must have occurred within 7 d after general anesthesia or sedation. Cases were excluded if only a portion of the animal was submitted, if death occurred ≥ 7 d after the anesthetic event, or if there was insufficient information to determine the time frame from anesthesia/sedation to death. The 4 pathology laboratories participating in the study were i) the Animal Health Centre in Abbottsford, British Columbia; ii) the Diagnostic Services Unit in Calgary, Alberta; iii) Prairie Diagnostic Services in Saskatoon, Saskatchewan; and iv) the Animal Health Laboratory in Guelph, Ontario. Each laboratory queried their database using keywords according to the laboratory’s method of storing postmortem reports to identify cases meeting our inclusion criteria. Database searches were conducted from the earliest available date to February 14, 2022. The dates were as follows: Animal Health Centre: February 2015 to July 2021; Diagnostic Services Unit: January 2012 to December 2021; Animal Health Laboratory: January 2016 to February 2022; and Prairie Diagnostic Services: June 2014 to January 2022. All laboratories agreed to participate in the study. Anonymity and confidentiality were maintained through redaction of personal information, as necessary.

Data requested included signalment (species, breed, sex, weight, age), clinical history on postmortem submission forms, gross pathology and histopathology findings, ancillary tests when available, and final interpretations from the case pathologist. One investigator (NR) reviewed all cases, using provided information (summary Microsoft Excel electronic spreadsheets or individual case reports) to extract all relevant information. Histories were considered complete according to the criteria reported by the Delay (2016) study (10), including information on the following: the health status of the animal before the procedure, the procedure for which the anesthesia/sedation was being performed, the timing of the complication, details of the complication, and whether cardiopulmonary resuscitation (CPR) was done. Cases in which the history noted additional documentation was attached were excluded in evaluating for completeness, as these documents were not available for review. The submitted history was used to assign an American Society of Anesthesiologists (ASA) physical status classification (11). The ASA physical status classification was discussed with a second investigator (DP, a Board-certified veterinary anesthesiologist) at the start of the study and any equivocal cases were independently reviewed by DP and consensus reached. The pathologist’s comments were reviewed to identify if COD was determined, to categorize COD, and to determine the most significant lesions that did or may have contributed to death.

Procedure type was defined as the primary reason for anesthesia and divided into 5 categories: spay/neuter, dental, abdominal, orthopedic, and other. Timing of death was determined based on history and classified as occurring during i) premedication, ii) induction, iii) maintenance, or iv) post-sedation/anesthesia. Premedication was defined as the interval from injection to achieve sedation until start of induction of general anesthesia. Induction was the period from delivery of the anesthetic induction agent to endotracheal tube placement. Maintenance was from placement of the endotracheal tube to start of recovery. Post-sedation/anesthesia began when the anesthetic vaporizer was turned off (or pharmacological antagonist administration, in cases of injectable anesthesia) or, in cases of sedation, when the procedure was finished. Intervals from the end of anesthesia were recorded where available. For procedures involving sedation only, intervals were divided into premedication (when animal received injection for sedation), maintenance (when procedure was being done) and recovery (when procedure was finished and/or animal was given a pharmacological antagonist).

Data investigating COD and the most significant lesions were limited to cases where both gross and histologic reports were available for review. Cases were first classified according to whether the COD was confirmed, uncertain, or unknown, based on the pathologist’s interpretation (Figure 1). For animals with a confirmed or uncertain COD, these were further categorized based on all information available (clinical history, postmortem report, and pathologist’s interpretation). These cases were subcategorized as i) preexisting disease, ii) undiagnosed condition, or iii) complication (Figure 1). Finally, lesions were divided into 5 major groups: respiratory, cardiovascular, digestive, other (included all other organs not listed in the other groups), and no significant lesions. Data supporting the results are available in an online repository (https://doi.org/10.7910/DVN/BXD2MI).

FIGURE 1.

Cause of death determination. Cases were classified as confirmed, uncertain, or unknown. Cases classified as confirmed or uncertained were further subcategorized as preexisting disease, undiagnosed condition, or complication.

RESULTS

A total of 425 postmortem cases, representing 232 dogs (54.6%) and 193 cats (45.4%), met the inclusion criteria (Table 1). More female than male dogs and cats were submitted. The number of postmortem perianesthetic death submissions increased between 2016 and 2021. However, when calculating the percentage these deaths represented out of all necropsies at each facility, the percentage remained relatively stable, with a median for all laboratories of 4.1% (range: 9.0 to 0.8%). Across all 4 laboratories, pathology reports that contributed to the data were provided by ~89 pathologists. Postmortem practices varied among laboratories; for example, heart weight was measured in between 14 and 75% of cases. However, it is unknown if this variation resulted from differences in practices among pathologists or laboratories.

TABLE 1.

Demographic data of dogs and cats with perianesthetic/sedation deaths that were submitted to 4 Canadian veterinary diagnostic laboratories for postmortem examination.

Dogs			Cats
Total (n)	232		Total (n)	193
Sex, n (%)	Male	102 (44.0)	Sex, n (%)	Male	69 (35.8)
	Female	129 (55.6)		Female	122 (63.2)
	NA	1 (0.4)		NA	2 (1.0)
Age (y)	Mean ± SD	4.8 ± 4.0	Age (y)	Mean ± SD	2.7 ± 3.7
n = 231	Median (range)	4.0 (0.13 to 15.0)	n = 188	Median (range)	1.0 (0.13 to 16)
Weight (kg)	Mean ± SD	16.4 ± 14.2	Weight (kg)	Mean ± SD	3.8 ± 1.8
n = 198	Median (range)	10.1 (0.78 to 60)	n = 159	Median (range)	3.2 (0.1 to 11.1)

NA — Not available.

Data on sex, age, and/or weight were not available in all submitted cases, resulting in differences in total numbers.

History and clinical impression data

History was provided in 401/425 (94%) submissions and was considered complete in 242/401 (60.3%) cases and incomplete in 159/401 (39.7%) cases. More than 80% of dogs (203/232) and cats (160/193) died without assistance (i.e., not euthanized), whereas ≤ 15% of dogs (23/232) and cats (28/193) were euthanized. More than 60% of the deaths occurred between premedication and the first 6 h after the procedure in both species (dogs: 129/213, cats: 113/173), with most deaths occurring during maintenance in dogs (49/213) and within the first 6 h of recovery in cats (49/173) (Figure 2).

FIGURE 2.

Timing of deaths in the perianesthetic period in dogs (black) and cats (gray). Deaths are reported after premedication, during induction (from when induction agent is given to intubation and has had monitoring values recorded), during maintenance (end of induction to end of procedure), and during intervals after termination of the procedure.

The submitting clinician commented on the suspected cause of perianesthetic death in ~25% of cases (dogs: 59/232, cats: 53/193). In dogs, the clinician reported an error either in surgery or anesthesia that may have contributed to the animal’s death in 18.6% (11/59) of cases. Errors reported included overdose, slipped ligature, owner-suspected negligence, hemorrhage, hypoxia, anesthetic complication, barotrauma from closed pop-off valve, owner-suspected incorrect diagnosis, and subcutaneous emphysema. The remaining 81.4% of deaths (48/59) were believed by the clinicians to be due to some unknown underlying disease or preexisting illness, with cardiovascular (15.3%, 9/59), respiratory (13.6%, 8/59) and embolism (10.2%, 6/59) the most common suspected lesions reported by submitting clinicians. In cats, underlying preexisting disease was suspected by submitters to be the COD in 88.7% (47/53) of the cases, with 11.3% (6/53) of the deaths due to errors. The errors listed included pulmonary over-inflation, excess air pressure during laparoscopic surgery, subcutaneous emphysema, hemoabdomen, hypercapnia from kinked exhaust tube, and barotrauma from closed pop-off valve. There were 34% (18/53) of cases attributed to disease of the cardiovascular system, 17% (9/53) associated with the respiratory system, and 7.5% of cases each of infectious (4/53) and anesthetic (4/53) reactions.

Postmortem findings and cause of death

The numbers of reports that included both gross and histology data were 168/232 for dogs (72.4%) and 147/193 for cats (76.2%). Numbers were similar for each of the 3 determination categories: confirmed (dogs: 73/168, 43.5%; cats: 50/147, 34%), uncertain (dogs: 53/168, 31.5%; cats: 47/147, 32.0%), and unknown (dogs: 42/168, 25%; cats: 50/147, 34.0%).

In cases where the COD was “confirmed” or “uncertain,” ~50% of dogs (62/126) and cats (53/97) had an undiagnosed condition. Examples of undiagnosed conditions included congenital heart disease, feline infectious peritonitis, pneumonia, and neoplasia. Preexisting disease was identified in 21% of dogs (26/126) and 8% of cats (8/97). Examples of preexisting disease included brachycephalic obstructive airway disease, gastric dilatation-volvulus, pancreatitis, and neoplasia. Finally, 30% of dogs (38/126) and 37% of cats (36/97) were categorized under “complications.” Examples of complications included hemoabdomen, intestinal perforation, aspiration pneumonia, and tracheal tear.

The respiratory tract was the most common system to have lesions that contributed or was suspected to have contributed to death in both dogs (35/126, 27.8%) and cats (31/97, 31.9%). In dogs, the digestive system was the second most common system to have lesions (30/126, 23.8%), followed by the cardiovascular system (18/126, 14.3%). In cats, the cardiovascular system was the second most common system to have lesions (21/97, 21.6%) followed by the digestive system (11/97, 11.3%).

Lesions in dogs and cats in cases where a confirmed or uncertain COD was identified are summarized in Table S1 and Table S2 (available online from: www.canadianveterinarians.net).

Overall, for all procedures, 65.1% of dogs (151/232) and 75.6% of cats (146/193) were designated with an ASA physical status of I or II. Spay/neuter surgery was the most common procedure involved in perianesthetic death postmortem submission in dogs (72/232, 31%) and cats (111/193, 58%). The next most common surgical procedures, in descending order, were dental surgery (dogs: 53/232, 23%; cats: 31/193, 16%), abdominal surgery (dogs: 34/232, 15%; cats: 11/193, 6%), and orthopedic surgery (dogs: 17/232, 7%; cats: 8/193, 4%). More than 90% of dogs (68/72) and cats (103/111) undergoing a spay/neuter procedure had an ASA status of I or II. The ASA status by procedure and animal age by procedure are presented in Figure 3 and Table 2, respectively. In spay/neuter cases where COD was categorized, COD was unknown in 31.4% of dogs (16/51) and 37.0% of cats (30/81).

FIGURE 3.

American Society of Anesthesiologists (ASA) physical status classification assignments (I to V) for each procedure group in dogs (D) and cats (C). Dogs: n = 232; cats: n = 193.

NA — Not available.

TABLE 2.

Mean ages and standard deviations in dogs and cats by procedure type.

Procedure	Dog		Cat
	n	Age (y), mean ± standard deviation	n	Age (y), mean ± standard deviation
Spay/neuter	72	1.5 ± 1.8	106	0.9 ± 1.2
NA	17	2.4 ± 2.7	14	1.9 ± 2.3
Abdominal	33	5.29 ± 3.8	11	2.6 ± 2.5
Orthopedic	17	5.3 ± 3.2	8	0.8 ± 0.5
Dental	53	8.0 ± 3.5	31	8.1 ± 4.3
Other	39	6.6 ± 4.2	18	5.8 ± 4.1

NA — Procedure type not reported.

DISCUSSION

The main results in this study were that i) there were considerable discrepancies between suspected and actual COD; ii) 1/3 to 1/2 of all cases reported were for spay/neuter procedures, a population of animals with a low risk of perianesthetic/sedation death; and iii) reporting of clinical history was incomplete in many perianesthetic/sedation death cases with animals submitted for postmortem examination.

One of the main reasons clinicians submit cases of perianesthetic/sedation deaths for postmortem examinations is to determine COD. In our study, most (84.8%) clinicians suspected the animal had an underlying unknown condition. However, only ~1/3 of cases (43% dogs, 34% cats) had a COD confirmed by the pathologist, whereas most had either an uncertain COD or an unknown COD.

This was apparently the first study to divide COD into “confirmed” and “uncertain.” This division highlighted that identifying a lesion did not necessarily mean it contributed to anesthetic death. Furthermore, this approach highlighted that there are no clear standard guidelines to help the pathologist determine which lesions may be considered significant in cases of perianesthetic death. An example would be the significance of identifying cardiac lesions in cats. Our study identified cardiac lesions as the most significant lesion in 21 of the cases. However, in only 5 of these cases (23.8%) was the pathologist certain that this was the confirmed COD. Additional lesions in these cases that supported heart disease as the COD included the presence of edema and/or heart failure cells in lungs. A study investigating effects of heart disease on anesthetic complications during dental procedures in dogs did not identify a significant difference in anesthetic complications between dogs with heart disease (n = 100) and dogs without heart disease (n = 100) (12). Dogs in the heart disease group were either referred by their primary veterinarian or from a specialty cardiology service due to concerns that the heart disease was severe enough to cause anesthetic complications. None of the dogs exhibited signs of congestive heart failure before anesthesia and none died during the study period (12). These results provided further support that, even if an animal has lesions associated with the cardiovascular system, it is possible that the lesion was an incidental finding if there is no evidence of heart failure.

Approximately 1/3 of both dogs (25%) and cats (34%) lacked significant lesions that could have contributed to the death of the animals. Previous studies identified that 36% of anesthetic/sedation death cases in all species (10) and 63% of anesthetic death cases in cats (9) lacked significant lesions. When the results are taken together with those of the current study, there appears a clear pattern that an important proportion of cases do not have identifiable lesions. The reason for the lack of lesions remains unclear, but there are speculative explanations. First, complications may develop so acutely that there is not enough time for detected lesions to visibly form. Second, several complications and conditions may not result in gross or histologic lesions. Examples include acute cardiac dysrhythmias, medication errors, and adverse drug reactions. In addition, perianesthetic death submissions are challenging cases for veterinary pathologists and, despite recommendations (13), there are currently no consensus guidelines for this type of submission. The lack of consensus guidelines can make it challenging for a pathologist to know which body system or type of lesion should be focused on, complicating the determination of which identified lesions are considered significant.

Finally, we can potentially extrapolate methods in human medicine to our veterinary species. Studies in human medicine have consistently shown that many cases of anesthetic death are preventable and related to some form of human error (5,7,14), with ≥ 50% of anesthetic “incidents” (not just anesthetic deaths) reported to be preventable (15). In human medicine, depending on jurisdiction, some level of reporting is usually required for any anesthetic complication. In addition, cases of anesthetic death may be thoroughly reviewed by an independent multidisciplinary committee, which may include surgeons, anesthesiologists, and pathologists (16). Currently, such in-depth reviews of perianesthetic/sedation death cases are not completed in veterinary medicine. As previously mentioned, submission histories are often incomplete and diagnoses may be missed due to lack of information.

Spay/neuter procedures were most common for the submissions in our study, representing ≥ 50% of submissions in cats and 31% in dogs. The reason for the overrepresentation of this procedure is likely multifactorial. First, in North America, sterilization procedures are probably the most common surgeries performed in general practice, creating an increased pool of animals for which complications, including death, can occur (17). Second, owners may be more motivated to submit their pet dogs and cats for necropsy if the animals were considered healthy and at a minimal risk for complications before an elective surgery.

Most animals in the spay/neuter procedure group had low ASA physical status (I to II) (> 90% of both dogs and cats) and one of the lowest mean ages when compared to other procedure categories. There were assumptions when determining the ASA status in this manner; for example, animals undergoing spay/neuter procedures and dentistry were assigned a low ASA status (I to II) unless the clinical history reported any illnesses. This is because these procedures tend to be done as elective procedures, and thus these animals would be assumed not to have underlying health issues that would increase their risk under anesthesia. Despite this assumption, the overall trend in this dataset was that most animals undergoing spay/neuter procedures were considered at low anesthetic risk before the procedure.

Within the spay/neuter cases, 30 to 40% of dogs and cats lacked significant lesions to explain the animals’ deaths. As a comparison, Gerdin et al (2011) investigated the COD in cats at spay-neuter clinics and reported 63% of the cats lacked lesions (9). Furthermore, the Delay (2016) study reported that 43% of dogs (45/105) and 34% of cats (31/90) had no significant lesions (10). The median age of these animals was 9 mo in dogs (range: 3 mo to 19 y) and 7 mo in cats (range: 2 mo to 13 y). Together, these studies indicated that a significant portion of younger animals undergoing elective procedures lack significant postmortem lesions. The underlying explanation for this observation is unclear. Young, apparently healthy animals are unlikely to have chronic disease predisposing to death and any associated lesion(s) identifiable during necropsy. It is also possible that COD in these animals related to perianesthetic/sedation complications that either were too acute for lesions to form or did not result in obvious lesions. In such cases, more detailed case information would be invaluable for accurately determining factors associated with COD, and consultation with anesthesiologists could be beneficial.

A detailed clinical history is an important part of any postmortem examination, particularly in cases of anesthetic death, as it may provide clues or direct the pathologist during their examination. None of the submitted histories in our study provided all recommended information (9). When using the criteria outlined in the Delay (2016) study (10), 39.7% (159/401) of histories in our study were considered incomplete. Lack of information on the submission form, including the type of procedure done, the physical status of the animal, details of the complications, and whether CPR was done, can substantially limit a thorough investigation. For example, myocardial and pericardial hemorrhage identified on postmortem examination can be associated with intracardiac injection; however, without a clinical history of this being performed during CPR, the pathologist might interpret this finding as a significant lesion. Without the additional history, the pathologist may falsely assume this to be a significant lesion and to be the confirmed COD.

Anesthetic/sedation drug protocols were not reported in almost 2/3 of the cases. Whereas interpretation of drug protocols is beyond the scope of a pathology report, providing this information may allow for consultation with an anesthesiologist, particularly in cases where COD is uncertain or unknown. Only a few studies have used a multidisciplinary approach when investigating anesthetic death (1). The confidential inquiry into small animal fatalities used a panel of specialists to determine COD; however, only 10% of the cases had a pathology report available for review (1). Clearly, the use of a multidisciplinary approach to investigating cases of perianesthetic/sedation deaths in veterinary medicine is currently underused.

Limitations of our study included the retrospective design that results in variability in data obtained and methods of searching. As an example, the data were based on pathology reports from 89 individuals, which could have contributed to variability in interpretation of the pathologic findings — a point that further supports the need to standardize the postmortem procedure for perianesthetic death submissions. In addition, there were histories excluded from analysis of the history completeness because additional documents that may have been attached were unavailable for review. The effect of this missing information on results skew was small, as the proportion of affected cases was relatively low (5.6%, 24/425 cases). Despite these limitations, our data supported previously reported trends and the idea that more studies, improved reporting, and better collaboration among disciplines are needed to investigate perianesthetic/sedation death and to help reduce its incidence in our veterinary patients.

In conclusion, the results of our study revealed that a large proportion of cases of postmortem submissions following perianesthetic/sedation deaths lacked lesions explaining COD. This finding agreed with previous work. Our results further identified challenges faced by pathologists when conducting postmortem examinations and attempting to determine COD in the face of incomplete clinical history and case information. A possible solution is to use a separate postmortem submission form for cases of perianesthetic death that advises submitting clinicians of what information is important for these cases. The predominance of submissions from spay/neuter surgeries is a concern since these animals were largely considered healthy and lacked significant lesions to explain the deaths. Finally, a standard approach to postmortem examination of perianesthetic/sedation death is not routinely taken, and use of a multidisciplinary investigative approach could help identify COD in cases where necropsy findings are uncertain or unknown.