Perioperative Outcome in Dogs Undergoing Emergency Abdominal Surgery: A Retrospective Study on 82 Cases (2018-2020)

Emergency abdominal surgery carries high morbidity and mortality rates in human medicine; however, there is less evidence characterizing the outcome of these surgeries as a single group in dogs. The aim of the study was to characterize the clinical course, associated complications and outcome of dogs undergoing emergency abdominal surgery. A retrospective study was conducted. Dogs undergoing emergency laparotomy were included in the study. Logistic regression analysis was performed to identify variables correlated with death and complications. Eighty-two dogs were included in the study. The most common reason for surgery was gastrointestinal foreign body. Overall 15-day mortality rate was 20.7% (17/82). Median (range) length of hospitalization was 3 (0.5 - 15) days. Of the 82 patients, 24 (29.3%) developed major complications and 66 (80.5%) minor complications. Perioperative factors significantly associated with death included tachycardia (P &lt; 0.001), hypothermia (P &lt; 0.001), lactate acidosis (P &lt; 0.001), shock index &gt; 1 (P &lt; 0.001), leukopenia (P &lt; 0.001), thrombocytopenia (P &lt; 0.001) at admission; intraoperative hypotension (P &lt; 0.001) and perioperative use of blood products (P &lt; 0.001). The results of this study suggest that mortality and morbidity rates after emergency abdominal surgery in dogs are high.


Introduction
Emergency abdominal surgery is a common procedure in both veterinary and human medicine. In human medicine, emergency surgery accounts for 11-15% of all total surgeries, however, it is responsible for almost 50% of all postoperative deaths and 30% of all surgical complications [1].
In 2015, Havens and colleagues [4] demonstrated an independent association between emergency general surgery and negative outcome in people. They found that patients had significantly more comorbidities in the emergency setting, but, even when controlling these factors, patients undergoing emergency surgery were still 39% more likely to die compared with non-emergency surgery [4]. Procedural urgency is also a known risk factor for anaesthetic-related death in small animals [15]. arterial pressure [SAP]), presence of septic peritonitis at admission and preoperative and postoperative clinical laboratory findings (haemato-biochemical and blood gas analysis). Preoperative comorbidities were retrospectively classified using a modification of the Charlson Comorbidity Index score used in Human Medicine [18] (Table 1). The criteria for diagnosis of septic peritonitis at admission were the evidence of peritoneal effusion (with abdominal ultrasound or at the time of surgery) and the identification of intracellular bacteria and/or positive bacterial culture results for a peritoneal fluid sample. Comorbidity Index (CI) 0 No comorbidity present 1 Any malignancy, compensated diabetes, mild cardiomyopathy 2 Moderate or severe cardiomyopathy, liver disease, renal insufficiency, decompensated diabetes 3 Metastatic tumour Other data recorded included indication for surgery, length of anaesthesia, length of surgery, requirement of blood product transfusions, use of vasopressors, use of synthetic colloids, complications, duration of hospitalization and 15-day mortality.
Complications were divided in intraoperative and postoperative. Intraoperative complications evaluated were hypotension (mean arterial blood pressure < 60 mm Hg as measured via a direct or indirect method), tachycardia (heart rate > 130 bpm), bradycardia (heart rate < 50 bpm) and hypothermia (temperature < 36 ºC). Postoperative complications were divided into minor and major ( Table 2).
Descriptive statistics were calculated and expressed as percentages. Data were analysed for distribution with a D'Agostino & Pearson test. Parametrical normally distributed data were expressed as mean and standard deviation (SD); non parametrical and not normally distributed variables were expressed as median (range). A Student Ttest or a Mann-Withney test were used to compare data between survivors and nonsurvivors. Statistical significance was set at P < 0.05. A logistic regression model was used to evaluate univariable association of perioperative variables with mortality and postoperative complications. Model-fit was evaluated using the Hosmer Lemeshow goodness-of-fit test statistic and the area under the ROC curve (Hosmer and Lemeshow, 2000; Dohoo e Stryhn, 2009). Results are reported as Odd Ratio (OR). Statistical significance was set at P < 0.05. Prism Version 6.0 (GraphPad Software Inc., CA, USA) and JMP software (SAS Institute Inc., Cary, NC, USA) were used to develop all statistical procedures.

Results
The initial medical record search identified 142 possible cases of emergency abdominal surgery in dogs. All but 92 were removed because of incomplete medical records. An additional 10 dogs that underwent caesareans sections were not included. Eighty-two dogs met the inclusion criteria during the study period. Of the 82 dogs enrolled 45 (55%) were females (35 sexually intact and 10 spayed) and 37 (45%) males (31 sexually intact and 6 castrated). The median age was 7.1 (0.4 -15.6) years and the median body weight was 18.5 (1.7-50) kg. Thirty breeds were included in the study. The most common breeds were mixed-breed dog ( Physical examination and laboratory findings recorded at arrival of all animals and the differences between survivors and non-survivors are expressed in Table 3 and Table  4, respectively. Perioperative variables significantly associated with mortality are represented in Table 5. The most frequent reason for surgery was non-linear gastrointestinal foreign body (n = 31 [39%]), followed by gastric dilatation and volvulus (GDV) (    . Development of intraoperative hypotension was associated with mortality (P < 0.001), postoperative development of SIRS/sepsis (P < 0.001), anorexia (P < 0.01) and postoperative hypotension requiring use of vasopressors (P > 0.001). No other intraoperative variable was associated with mortality or the development of postoperative complications.
Oesophagostomy and nasogastric feeding tubes were placed in 20/82 and 2/82 dogs, respectively, for management of patients' nutritional requirements after surgery and 11/82 dogs had a central venous catheter placed. The decision of feeding tube or central venous catheter placement was determined by the clinician at the time of surgery.
Eighteen dogs in this study received a blood product transfusion during the perioperative period. Thirteen (15,85%) dogs received fresh-frozen plasma and 5 (6.1%) dogs received packed red blood cells (RBCs). Postoperative vasopressors were used in 15/82 (18.3%) dogs. The requirement of blood product transfusion significantly correlated with mortality (P < 0.001) and postoperative development of SIRS (P < 0.01) and hypotension requiring vasopressor therapy (P > 0.001). The postoperative use of vasopressors was correlated to death (P < 0.001) and postoperative development of diarrhoea (P < 0.01), SIRS (P < 0.001), anorexia (P < 0.05) and major electrolyte imbalance (P < 0.05).  Sixteen (19.5%) dogs died or were euthanized during the perioperative period. One dog had a cardiopulmonary arrest 7 days after discharge because of complications related to concomitant disease (hepatic insufficiency). Cardiopulmonary arrest unresponsive to resuscitation efforts occurred in 1 dog during surgery and in 13 after surgery. The remaining 2 dogs were euthanized because of complications arising after surgery.

Preprints
A total of 66 (80.5%) dogs survived to discharge from the hospital and a total of 65 (79.3%) were alive at the follow-up visit 2 weeks after surgery. The overall 15-day mortality rate was 20.7% and median (range) duration of hospital stay for the dogs that survived to discharge was 3 (0.5-15) days.
The median (range) duration of the long term follow up was 90 (15-425) days for all dogs. One dog treated for gastric dilatation-volvulus underwent surgery 4 months after the first surgery because of intestinal obstruction created by adherence of the first surgery and was euthanized in the postoperative period because of complications related to concomitant diseases. One dog treated for intestinal foreign body was readmitted to the hospital 2 months after surgery in shock because another intestinal foreign body was causing obstruction and was euthanized. The remaining dogs were all alive at the time of the last follow up.

Discussion
To our knowledge, the present study was the first in which factors associated with death and complications were evaluated in dogs that underwent emergency abdominal surgery.
The present study demonstrated that emergency laparotomy is a high-risk procedure, with an overall 15-day mortality rate of 20.7% and total complication rate of 80.5% with 24.4% of major complications. Factors associated with death included presence of CI ≥ 2; BCS ≤ 2; tachycardia, hypothermia, lactic acidosis, thrombocytopenia and leukopenia at admission; intraoperative hypotension; development of major complications and perioperative use of blood product. Normal haemoglobin value at admission was positively associated with survival to discharge. In human medicine, the most common procedures performed in the emergency setting are intestinal surgeries [2,11,19]. We found similar results, but, whereas the most common indication for humans are intestinal obstruction caused by adhesions or neoplasia [2,10,11,19], gastrointestinal foreign body causing mechanical obstruction was the most common indication for surgery in the present study, which is in accordance with previous research in gastrointestinal emergencies in dogs [20,21].
Considering preoperative patient factors and their associations with mortality, the odds of non-survival increased with increasing CI and decreasing BCS. Previous research in people have found associations between malnutrition and presence of comorbidities with increasing postoperative complications and death [14,37]. It is also worth noting that most patient with CI 2 in the present study had a diagnosis of chronic enteropathy, and a relationship between high CI and low BCS that was not statistically evaluated cannot be excluded.
On the other hand, it is well established in people that mortality after emergency surgery increases with age [2,3,9,10,19]. This association was not found in the present study. This difference may be explained by differences in range of age between humans and dogs. Moreover, in contrast with studies in people, most common reason for surgery in the present study was gastrointestinal foreign body causing mechanical obstruction, condition commonly reported in young dogs [24][25][26].
Clinical signs in emergency abdominal surgery can vary significantly depending on the underlying condition. Most of these patients are in shock and require immediately treatment. In the present study tachycardia, hypothermia and SI > 1 at admission were found to be associated with death. These alterations can reflect decreased perfusion and the sympathetic compensatory response in patients with shock [38]. Hypothermia can be caused by decreased systemic perfusion and it is associated to reduced venous return [31]; and tachycardia can occur for multiple reasons, pain and/or compromised cardiovascular state. Their association with mortality may be related to severity and duration of hypoperfusion at presentation. This is in accordance with other studies; tachycardia and hypothermia have been identified as risk factors for death in dogs with hemoperitoneum [28] and GDV [31,39].
On the other hand, SAP values < 90 mmHg at admission were not correlated with failure to survive. This could indicate that most patient arrived in a compensatory shock stage and increased HR at this stage is more accurately predicting the severity of illness than SAP. Moreover, the method of SAP measurement (oscillometer or doppler) was not standardized, and that could have altered the results. On the other hand, median (range) SI value of survivors (0.94 [ 0.3 -2.8]) was significantly (P = 0.01) lower than value of nonsurvivors (1.2 [0. 6 -3.3], and SI > 1 was significant in predicting mortality in the multiple regression analysis. This is in accordance with previous researches that support that SI is more accurately predicting mortality in comparison with low SAP values [40], and SI have been associated with shock and mortality in both human [41,42] and dogs [43]. These results point out the need for haemodynamic stabilization before the surgery for these Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 27 July 2021 doi:10.20944/preprints202107.0607.v1 patients and that the SI, a powerful and easily calculated tool, should always be taken into account to predict more adequately the shock status of the patient. Dogs that failed to survive had a median (range) plasma lactate concentration of 5 (0.6 -16.6) mmol/L and a median venous pH of 7.23 (6.81 -7.46) compared with a median plasma lactate concentration of 1.9 (0.3 -14.8) mmol/L and a median venous pH of 7.39 (7.12 -7.55) in survivors. Lactate > 4.5 mmol/L and pH < 7.3 were associated with death in the multiple regression logistic model. Lactic acidosis is a common finding in critically ill patients and is used as a marker of patient perfusion and resuscitation response both in human and veterinary medicine [38,44]. High serum lactate concentrations at admission have been associated with failure to survive in dogs after gastrointestinal surgery [21], secondary septic peritonitis [45], GDV [30,[34][35][36] and trauma [46]. Nevertheless, the response of the lactate to intervention (fluid resuscitation) has been shown to provide a more accurate assessment of patient's prognosis in human [47,48] and veterinary [34,49] studies. In fact, in the present study 10/82 dogs had a lactate value at admission above 6 mmol/l, and 5/10 (50%) survive. Lactate clearance was not evaluated in the present study, so we are not able to say if it had any influence on survival rate. Even so, patients with high plasma lactate concentration at hospital admission should be aggressively treated and closely monitored perioperatively to improve outcomes.
The presence of leukopenia and thrombocytopenia at admission were identified as risk factors for mortality (P < 0.001). WBC count < 4.5 x 103/L have been correlated with mortality in emergency abdominal surgery in human medicine [2]. The presence of leukopenia before surgery certainly evidences an immune imbalance or a severe inflammatory response and makes the dog more susceptible to any type of infection [50]. Thrombocytopenia may cause pulmonary and neurologic complications and severe blood loss, in human medicine patients with platelet counts < 200 x 103/L that underwent an invasive procedure are at risk of haemorrhage [51]. Fresh frozen plasma should be administered before surgery in severely leukopenic and thrombocytopenic patients.
Albumin value < 2.5 g/dL at admission was as well associated with mortality (P < 0.001) and perioperative use of blood products (P < 0.001). Hypoalbuminemia is often associated with critical illness and can result in complications like pulmonary oedema, enteral feeding intolerance, poor wound healing and hypercoagulability [52]. Hypoalbuminemia has been shown to be associated with increased morbidity and mortality both hospitalized people [53][54][55] and dogs [22,24,52,[56][57][58]. In human medicine one study demonstrated that the increase in the estimated odds of death ranged from 24% to 56% for every 0.25 g/dL reduction in albumin concentration [55]. Albumin concentration serves as a marker for disease and nutritional status and is essential for the maintaining of oncotic pressure [52]. It is worth to note that 36% of dogs that were hypoalbuminemic at admission developed postoperative hypotension, and that all dogs that developed postoperative hypotension and had postoperative albumin values measured had postoperative hypoalbuminemia. The use of albumin supplementation has been shown to improve outcomes in dogs [57,58] and its use in these vulnerable patients should be considered.
Intraoperative hypotension was diagnosed in 50 (60%) of the dogs in the present study and was significantly associated with postoperative development of hypotension, SIRS/sepsis and anorexia; use of vasopressors in the postoperative period; perioperative use of blood products and death. Similarly, Beck et al [29] found that hypotension detected at any time during the hospitalization period was associated with increased risk of death in patients with GDV. These results likely reflect the fact that dogs that developed intraoperative hypotension and required vasopressors therapy or blood product transfusions had more severe disease in general and were more likely to develop postoperative complications and fail to survive than other dogs. However, intraoperative hypotension episodes produce low tissue oxygenation that could certainly have worsened some conditions, like SIRS or anorexia. A relationship between intraoperative hypotension and development of postoperative anastomotic leakage after intestinal surgery has been previously demonstrated, and it has been attributed to local ischemia at the anastomosis site Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 27 July 2021 doi:10.20944/preprints202107.0607.v1 [22,23]. It could be hypothesized that intraoperative hypotensive episodes decrease enteric perfusion and enteric motility, which might have affected postoperative ileus and indirectly postoperative anorexia development. Moreover, absolute and relative intraoperative hypotension has been found to be responsible for acute myocardial and renal injuries in human medicine [59]. Unfortunately, we did not evaluate the incidence of acute myocardial injuries because of lack of specific data so we cannot draw any conclusions concerning this issue. Nevertheless, intraoperative normotension should represent an important therapeutic target to maximize clinical success and invasive blood pressure monitoring is advised for these patients.
Postoperative major complications were encountered in 24.4% of dogs. This is also in accordance with human medicine where major complications rates range from 33 to 47 % [4,6,7]. Respiratory complications are the most frequent complications for emergency abdominal surgery patients in human medicine [4,8] and they have been found to be associated with failure to discharge in dogs with hemoperitoneum [28]. In the present study 6 dogs developed pulmonary complications, 2 of them pulmonary thromboembolic disease diagnosed at necropsy. Pulmonary complications were not individually analysed for associations with mortality due to low number of cases, but the development of major complications after surgery was associated with mortality (P < 0.001), as has been found in human medicine [11,13]. In human medicine it has also been demonstrated that the development of postoperative complications not only increases mortality in the immediate risk period, but also worsens long-term mortality [60].
Hypotension requiring use of vasopressor therapy was the most common postoperative major complication, present in 15 (18%) of the dogs in the present study. Of those 15 dogs, 10 (67%) failed to survive. The use of vasopressors in the postoperative period was associated with death and postoperative development of SIRS, anorexia, diarrhoea and major electrolyte imbalance. In total, hypotension was documented in 18/82 (22%) of patients in the postoperative period. As mentioned above, hypotension is a severe hemodynamic disturbance that cause poor tissue perfusion resulting in cellular suffering and potentially multi-organ dysfunction [29]. In 1 recent human study [61], a postoperative mean arterial pressure below a threshold of 75 mmHg was found to increase the risk of myocardial injury, with shorter durations at lower thresholds being likewise harmful, and postoperative hypotension was more correlated to myocardial injuries than intraoperative hypotension. Frequent blood pressure monitoring measurements are mandatory in the postoperative period, and a constant invasive blood pressure monitoring should be considered in critical patients.
In the postoperative period, SIRS was diagnosed in 12/82 (15%) dogs. Development of SIRS was associated with BCS ≤ 2; CI ≥ 2; tachycardia, SI > 1, and WBC < 6 x 109/L at admission; development of intraoperative hypotension; perioperative use of blood products and postoperative use of vasopressors. Of the 12 dogs that developed SIRS in the postoperative period, 10 failed to survive. The development of SIRS in the postoperative period can be considered as a worsening in the patient's recovery and assessing the evolution of the SIRS score daily in every patient should be considered. Unfortunately, one of the limitations of the study is that the SIRS score changes were not evaluated.
Minor complications were present in 80.5 % of dogs in this study. The most common complication was anorexia, present in 41 (50%) of dogs. Anorexia was significantly correlated with age, adult and old dogs were 3.7 and 13.7 times, respectively, more prone to develop anorexia in comparison to young dogs (<2 years). In human medicine, it is well known that increasing age is a risk factor for postoperative ileus [11,62]. Postoperative ileus could not be evaluated because the retrospective nature of the study, but it is possible that one of the reasons of anorexia high prevalence in elderly patients is related to postoperative ileus. Weight was also found to be related with the development of anorexia. Dogs weighing less than 10 kg were more likely to be anorexic compared with dogs > 10 kg. Other risk factors for the development of anorexia found in the present study were tachycardia, SI > 1 and pH < 7.3 at admission; intraoperative hypotension and Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 27 July 2021 doi:10.20944/preprints202107.0607.v1 postoperative use of vasopressors. All the conditions seem to be related to the severity of systemic injury, and critically ill patients seems to be more likely to develop anorexia in the postoperative period. Moreover, an experimental study on the pathophysiology of the perioperative causes of ileus identified the increase of catecholamines blood concentration as responsible of ileus in dogs [63]; this could explain the relationship of anorexia with the postoperative use of vasopressors. Nutrition is essential during recovery from illness or surgery. Early nutritional support has been shown to improve outcomes and shorten recovery time [64]. Caution should be made in those dogs with known risk factors for the development of postoperative anorexia, and feedings tubes placed at the time of surgery could be a good option so enteral nutrition can be started as soon as possible after surgery in critically ill dogs. Feeding tubes were placed in 22/82 dogs in the present study.
Surgical site complications were present in 5 of 82 dogs (6%) and were significantly correlated with low BCS. Surgical site infections of 2.5 to 18.5 % are reported in small animal surgery, with variations associated with the type of surgery [65][66][67][68][69]. Undernutrition, emergency surgery and prolonged surgical times have been described as risks factors for surgical site infections in both human and veterinary medicine [67,68,[70][71][72][73]. In contrast, surgical time did not affect the development of surgical site complications in this study. In fact, surgical time was not related with any complication in the multiple logistic regression analysis, however, the maximum surgical time was 230 minutes. It may be concluded that if the patient is stable operating time is not a critical variable to considered.
Postoperative diarrhoea was associated with postoperative use of vasopressors. It could be hypothesized that severe hypotension requiring vasopressors can be a contributor to intestinal ischemia development that may promote morphological and functional changes that enhances intestinal permeability and can manifests clinically with diarrhoea [74]. Diarrhoea in the postoperative period was also associated with high haemoglobin value at admission and low BCS. In fact, postoperative vomiting was also associated with high haemoglobin value at admission. This may be explained by the fact that high haemoglobin value was related to the presence of dehydration at admission, and dehydrated patients were more likely to develop these complications.
Postoperative major electrolyte imbalance was associated with high CI and postoperative use of vasopressors. It is not surprising that patients that needed vasopressors for the maintenance of normal blood pressures were more predisposed to electrolyte disorders. The most common electrolyte derangement found in this study was hypokalaemia and the most common comorbidity encountered in patients classified with CI 2 was chronic enteropathy. Hypokalaemia have been found to be associated with gastrointestinal disease in previous research [75] and could be one of the reasons between the association of high CI and major electrolyte derangement in the present study. Nevertheless, electrolytes concentrations should be closely monitored in these patients.
Perioperative use of blood products was associated with postoperative development of SIRS, hypotension requiring vasopressors therapy and death. We did not differ between different types of blood products, like packed RBCs or fresh frozen plasma so it is difficult to make any conclusion about these results. We neither evaluated separately massive transfusion, known risk factor for postoperative complications and death in previous research [28]. Further study is warranted to investigate the impact on prognosis and needs for transfusions in dogs undergoing emergency abdominal surgery.
The median hospitalization time was relatively short (3 days; range 0.5 -15 days). It would appear that even if these patients usually require intensive monitoring, the median length of hospital stay is short.
This study has a number of limitations, the most important of which are a small sample size and the retrospective design where collection and interpretation of the data may be subject to information bias. Due to the retrospective and multicentric nature of the study, the laboratory test to be performed was not standardized and there is a lack of specific data, like postoperative albumin value in some patients. Types of treatment and Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 27 July 2021 doi:10.20944/preprints202107.0607.v1 resuscitation of the animals could have been different between clinician of the different hospitals and may have distorted the outcome and clouded the findings. Surgeries were also performed by multiple surgeons with different degrees of experience, which could have affected the results and the development of some complications. Larger multicentre prospective studies evaluating outcomes in dogs undergoing emergency abdominal surgery should be explored.

Conclusions
In conclusion, mortality and morbidity rates after emergency abdominal surgery in dogs from this study are comparable to those reported in human medicine. Several variables that correlated with failure to survive and postoperative complications development were identified in our patient population. We believe this study provided useful information for recognition of dogs at higher risk than other before emergency abdominal surgery and that may help clinicians identify those patients which can benefit of more intensive monitoring and intervention, as well as to inform owners about prognosis. This information could be used to create protocols for optimizing recovery after surgery in this dog population. Stabilize these parameters before surgery and verify the stabilization of the patient through end points, provide intensive intraoperative and postoperative monitoring and preventing some condition, like postoperative anorexia, with the use of feeding tubes, could be useful tools to improve outcomes. Further investigation in prospective studies is warranted to fully address the role of some interventions in improving outcomes.