Introduction

Septic peritonitis is a life-threatening disease caused by several conditions, including primary infection, abdominal organ abscessation and rupture, body wall perforation, or gastrointestinal (GI) or urogenital leakage.¹ Given the complex and dynamic nature of this disease, aggressive treatment is indicated for affected patients, consisting of initial stabilization, timely surgical intervention, and intensive postoperative care. Various prognostic indicators have been evaluated, including triage scoring systems, serum albumin and lactate levels, hypotension, initial antibiotic choices, the presence of a GI foreign body, and more.^2–6 Despite advances in medicine, reported mortality rates have been variable, ranging from 27 to 85%.¹ Data regarding long-term follow-up for patients surviving treatment for septic peritonitis are lacking. However, it is a reasonable assumption that barring significant comorbidities, these patients can do well if they survive the immediate postoperative period.

Neoplasia is one of the causes of septic peritonitis and can be responsible for GI perforation or, less commonly, organ abscessation.^2,7–11 Several studies reporting on septic peritonitis have contained cases caused by neoplasia, but none of the manuscripts draw specific conclusions about this subpopulation, likely because of the low case prevalence.^{2,3,6,7,9,12–16} To the authors’ current knowledge, there is no veterinary literature directly evaluating the outcome of patients with septic peritonitis caused by neoplasia.

There are several reasons why patients with neoplasia as a primary cause of septic peritonitis (NSP) may carry a different prognosis than patients with other primary causes (OSP). Normal immune function is required to successfully recover from septic peritonitis. It is possible that neoplasia could result in decreased healing ability following surgery due to increased tumor burden, elevated neoplasm cytokine levels and hypoalbuminemia secondary to cancer cachexia, and malignancy causing dysregulated cellular proliferation.¹⁷ Clinician and owner’s bias may also influence NSP patient outcomes, as treatment may be halted more readily owing to the potential of a guarded long-term prognosis.

There is a clinical need for data regarding outcome of NSP patients, given the acute life-threatening nature of septic peritonitis and the amount of emotional and financial investment of affected owners. The objective of this study was to evaluate for differences in survival to discharge and complication rates between NSP and OSP canine patients. The secondary objective was to describe the long-term outcome of NSP patients that survived to hospital discharge.

Materials and Methods

A single-institution retrospective cross-sectional cohort study was performed. As this was a retrospective study, Institutional Animal Care and Use approval and owner consent was not required. All patients in this study were clinically managed according to contemporary standards of care. An electronic medical record search was performed to identify dogs that underwent surgical treatment for septic peritonitis between January 1, 2010, and November 1, 2020. For inclusion, a diagnosis of septic peritonitis must have been made by visualization of intracellular bacteria by a boarded clinical pathologist on peritoneal fluid cytology, a positive bacterial culture of peritoneal effusion (obtained preoperatively or intraoperatively), a history of gross surgical lesions consistent with a septic nidus, such as a dog bite attack, or documented presence of intra-abdominal purulent discharge or GI leakage in the operative report. Additionally, the dog must have had a celiotomy as part of their treatment for septic peritonitis.

Dogs were separated into two groups, NSP and OSP, for data analysis. Dogs were included in the NSP group if neoplasia was diagnosed on histopathological tissue evaluation or if the operative report documented evidence of lesions suggestive of neoplasia (i.e., mass-like lesions) in the abdominal cavity and had confirmatory cytologic evaluation. All other cases were classified into the OSP group.

Data collected from the medical record included signalment at time of presentation, pertinent past medical history, preoperative laboratory work, and imaging study reports. Triage information for patients was collected and all statuses were retrospectively graded by one investigator (C.L.C.) using the Animal Trauma Triage (ATT) score and the Acute Patient Physiologic and Laboratory Evaluation (APPLE) score.^18,19 The APPLE_fast and APPLE_full scores were both calculated. APPLE_fast scores were calculated using parameters including glucose, albumin, lactate, platelet count, and mentation score. APPLE_full scores were calculated using parameters including creatinine, WBC, albumin, oxygen saturation, total bilirubin, mentation score, respiratory rate, age, fluid score, and lactate.^18,19 Operative reports and anesthesia records were reviewed for date of surgery, intraoperative complications, and diagnostics performed. The records for in-hospital postoperative care were assessed for postoperative complications and time to discharge. Patients were excluded from the study if they were not diagnosed with septic peritonitis as previously described or did not undergo surgical treatment for septic peritonitis.

A complication was defined as an adverse event associated with surgical intervention; intraoperative complications occurred between the time of skin incision and closure, and postoperative complications occurred after skin closure.²⁰ A postoperative complication was classified as immediate if it was evident during the postoperative period up until time of hospital discharge. A postoperative complication was classified as short term if it started between hospital discharge and incision recheck at 14–21 days after surgery. A postoperative complication was classified as long term if it started or persisted after the time of suture removal. Intraoperative complications were graded using the Classification of Intraoperative Complication (CLASSIC) system.^20,21 Postoperative complications were graded using the contracted Accordion Severity Classification system^20,22 with some minor modifications. Dogs were excluded from grade 1 immediate complications if they were only treated for expected postoperative clinical signs including discomfort and GI upset such as regurgitating, vomiting, and diarrhea until discharge from the hospital. Dogs were excluded from grade 2 immediate and short-term complications if they were only treated with antibiotics that were already started preoperatively or intraoperatively as part of septic peritonitis management.

Records were also evaluated for histopathologic diagnoses and follow-up. NSP patient histopathological reports were reviewed for details including the location of the lesion, histopathological diagnosis, and surgical margins. Details regarding any adjuvant therapy administered were also obtained. When available, short- and long-term follow-up were recorded, as were the date and cause of euthanasia/death if applicable. Long-term follow-up for NSP patients was pursued with communication with referral veterinarians or owners when possible.

Results

Demographics and Triage Scores at Presentation

Eighty-eight dogs were presented for and surgically treated for septic peritonitis during the included time period. Two of the dogs had “mass-like lesions” diagnosed intraoperatively with no histopathology or cytology available and so were excluded. A total of 86 patients (12 NSP, 74 OSP) met the inclusion criteria. The demographic and triage details of both NSP and OSP groups are summarized in Table 1. There was no significant difference in triage scores between the NSP and OSP groups. The most common breed in the NSP group was mixed-breed dog (n = 4); other breeds were represented by only one dog for each breed. The most common breeds in the OSP group were mixed-breed dog (n = 17), golden retriever (n = 7), and Labrador retriever (n = 6).

TABLE 1 Patient Demographics and Triage Information on Presentation for 86 Dogs That Received Surgical Treatment for Septic Peritonitis

View Fullscreen

Preoperative Imaging

Table 2 summarizes the preoperative imaging studies performed for all patients. There were no cases in which there were concerns for pulmonary metastatic disease on thoracic radiographs, but two dogs (1 NSP, 1 OSP) had other abnormalities reported. The NSP dog’s report documented concern for mediastinal fat versus a thymic mass, and the OSP dog’s report documented concern for mediastinal lymphadenopathy (neoplastic versus reactive). No additional diagnostics were performed to further evaluate these abnormalities. Thirteen dogs had concern for neoplasia or metastases on abdominal imaging (4 NSP, 9 OSP). The lesions described in these reports included hepatic nodule or mass (n = 5), a splenic nodule or mass (n = 4), abdominal lymphadenopathy (n = 6), possible sarcomatosis (n = 1), and a pancreatic nodule (n = 1). The dog with possible sarcomatosis had a history of a jejunal leiomyosarcoma that was completely excised via resection and anastomosis 6 mo before presentation for septic peritonitis. Cytology or histopathology of these lesions were performed and were not consistent with neoplasia in 10 patients (1 NSP, 9 OSP). No additional perioperative diagnostics were performed on the lesions in the remainder of the dogs.

TABLE 2 Summary of Preoperative Imaging Performed and Metastatic Concerns for Dogs in NSP and OSP Groups

View Fullscreen

Diagnoses

Table 3 summarizes details regarding diagnosis of septic peritonitis in the population. The 34 dogs (4 NSP, 30 OSP) that did not have a clinical pathologist’s interpretation of peritoneal fluid cytology with documentation of intracellular bacteria or positive bacterial growth on peritoneal fluid culture were diagnosed with septic peritonitis intraoperatively. Thirty-one of these dogs had a leaking GI lesion identified. One dog had a history of being attacked by another dog, and two dogs (1 NSP, 1 OSP) had documentation of intraabdominal purulent discharge in the operative report, one from abscessed uterine and ovarian pedicle stumps and the other from an abscessed hepatic mass. There was no culture documented for the dog with the uterine and ovarian pedicle stumps, and the dog with the hepatic mass initially had a culture submitted but subsequently canceled, presumably because the patient was euthanized the day after surgery. Ten of the 34 dogs had documentation of in-house diagnostics that were suggestive of septic peritonitis, including visualization of intracellular bacteria by a clinician or paired evaluation of peritoneal fluid and peripheral lactate and glucose concentrations.

TABLE 3 Summary of Diagnostics Performed on Peritoneal Fluid to Confirm Septic Peritonitis in NSP and OSP Groups

View Fullscreen

The causes of septic peritonitis are summarized in Table 4, and Table 5 outlines the available histopathological details for NSP cases. Eight of the 14 dogs that had a small intestinal (SI) perforation without presence of a foreign body had a history of nonsteroidal anti-inflammatory drug use within 30 days of presentation. Patient 11 had a documented ruptured jejunal mass that was found in surgery. The mass was not submitted for histopathological evaluation, but in-house cytology was performed by a clinician, which was consistent with a diagnosis of lymphoma.

TABLE 4 Causes of Septic Peritonitis and Number of Dogs Affected by Each Category, Divided into NSP and OSP Groups

View Fullscreen

TABLE 5 Available Histopathological Details of Neoplastic Lesions Diagnosed in the NSP Group

View Fullscreen

Postoperative Management and Complications

Postoperative IV antibiotics were used in all patients who survived through anesthetic recovery. Six dogs received ampicillin sulbactam only, 2 dogs received enrofloxacin only, and 43 dogs received a combination of both medications. Twenty-two dogs received a combination that included other antibiotics. Dogs were transitioned to oral antibiotics at the clinician’s discretion, and all dogs who were discharged from the hospital went home with continued antibiotic therapy except for 2 (one was discharged against medical advice because of financial constraints, and another was discharged to be euthanized with the primary care veterinarian). All in-hospital treatments were adjusted at the clinician’s discretion based on the patients’ clinical signs, any adverse events, and diagnostic results (for example, culture results or repeated laboratory work).

The postoperative immediate and short-term complications reported are summarized in Table 6. Eight dogs in the NSP group and 44 in the OSP group experienced complications. In the immediate postoperative time frame, 3 dogs (1 NSP, 2 OSP) had some documented abnormal neurologic signs. The NSP patient was noted to have an episode of head bobbing 1 day postoperatively that resolved with a single dose of midazolam. One of the OSP patients had a documented mild left-sided head tilt and mild proprioceptive deficits on the right thoracic and pelvic limbs that self-resolved with no intervention. A presumptive diagnosis of a vascular event was made, as the client did not elect to pursue further workup. The other OSP patient had a seizure, suspect secondary to enrofloxacin administration. Enrofloxacin was discontinued, and the patient received one dose of midazolam; they did not experience any further neurologic abnormalities.

TABLE 6 Summary of Postoperative Complications in Dogs (8/12 NSP, 44/74 OSP) That Were Presented for Septic Peritonitis and Received Surgical Treatment in NSP and OSP Groups

View Fullscreen

Seven (7/12, 58%) dogs in the NSP group survived to discharge, whereas 54 (54/74, 72%) dogs in the OSP group survived to discharge. Seventeen of the 54 (31%) OSP dogs were lost to follow-up after discharge from surgery. Forty-four dogs (7 NSP, 37 OSP) had a documented 2-week follow-up examination with a veterinarian after initial discharge. The median time of this examination was 15.5 days after surgery (range 9–35 days).

Twenty-eight of the dogs that had a documented 2-week follow-up (5 [71%] NSP, 23 [62%] OSP) had no documented postoperative short-term complications. One (14%) NSP dog and 10 (27%) OSP dogs experienced postoperative short-term grade 1 complications. The NSP dog had occasional gag/coughing episodes when rising from sternal recumbency but was otherwise doing well and did not have any documented abnormal auscultation findings or oral lesions noted on physical examination. Five (13%) OSP dogs experienced GI upset including vomiting and regurgitation that resolved with oral antacids or were self-resolved by the time of the 2-week recheck examination. Two (5%) OSP dogs had a seroma at the incision site. One OSP dog was presented with a cough before the scheduled recheck examination and was diagnosed with suspect tracheitis after thoracic radiographs revealed no significant findings. The dog was prescribed a cough suppressant, and the cough was resolved on recheck examination 3 days later. Another OSP dog was presented with suspect pancreatitis before the scheduled recheck examination and was hospitalized overnight with supportive care for the presumed pancreatitis. At the recheck examination 10 days after surgery, the dog was doing well. The last OSP dog was re-presented for discharge from the incision site 12 days after surgery, and a cranial abdominal wall hernia was palpated on physical examination. The size of the hernia was not documented. The dog was hospitalized for daily bandage changes for 2 days to manage the discharge from the incision and monitor the hernia. After discharge from the hospital, the dog was examined again 17 days after surgery. There was no discharge documented from the incision site, and the hernia was documented to be palpably healing; the dog was sent home after a bandage change. The dog was presented again 21 days after surgery, and the hernia was documented to be resolved and the incision healed. The dog was managed with bandage changes, and the hernia and discharge were documented to be resolved at 21 days after surgery.

Two dogs (one NSP, one OSP) experienced postoperative short-term grade 2 complications. The NSP dog was presented 10 days after surgery for vomiting and pyrexia. The dog was hospitalized for supportive care for presumed pancreatitis for 4 days. Broad-spectrum antibiotics were also administered as a definitive cause for the pyrexia was not found. The owner was instructed to follow up with the primary care veterinarian after the dog was discharged, and the dog was subsequently lost to further follow-up. The OSP dog was presented with pollakiuria at the 2-week recheck examination. This dog was initially diagnosed with a ruptured retroperitoneal abscess of unknown primary origin resulting in septic peritonitis. A definitive diagnosis for the dog’s clinical signs was not reached, but abdominal ultrasound (AUS) identified a thickened bladder wall as well as mild effusion and fibrous tissue surrounding the dog’s urinary tract. Urinalysis also revealed isosthenuria. Urine and peritoneal effusion culture were both negative, and the dog was treated with subcutaneous amikacin and crystalloid fluids for 1 wk. The dog’s condition improved until it was diagnosed with septic peritonitis (suspect abscess recurrence) and was euthanized 49 days from the initial surgery.

One dog (OSP) experienced a grade 3 postoperative short-term complication. The dog was presented 7 days after surgery with persistent vomiting. An AUS was consistent with pancreatitis, and the dog had an esophageal feeding tube placed 10 days after surgery to aid in medical management. The dog was examined 20 days after surgery and was doing well; the feeding tube was removed 29 days after surgery. Two NSP (5%) dogs experienced a grade 4 complication. One of the dogs was presented 9 days from surgery after experiencing vomiting episodes at home that day and arrested approximately 10 hr after hospital admission for supportive care. There was no known immediate cause for the arrest; the dog was suddenly noted in the kennel to not be breathing. No postmortem examination was performed on that dog. The other dog was presented 17 days after surgery for persistent vomiting and was diagnosed with a distal duodenal ulcer on AUS. The owner elected euthanasia.

Long-Term Outcome for NSP Cases

Of the 12 dogs, 7 were discharged from the hospital after surgery. Four of these 7 dogs were later euthanized because of causes unrelated to their previous neoplastic diagnosis, and the median overall survival time was 604 days (range 250–1438 days). One dog had another emergent surgery for recurrent abscessation of a hepatocellular adenoma but developed disseminated intravascular coagulation immediately postoperatively and was euthanized 250 days after the initial surgery. Two dogs were alive at the time of data accrual; electronic communication with the owners confirmed that these dogs were doing well. The overall survival times for these dogs were 282 and 403 days. See Table 7 for details regarding long-term outcome for NSP patients.

TABLE 7 Long-Term Outcome of Dogs with NSP Including Cause of Death and OST

View Fullscreen

Discussion

This study did not find a significant difference in complications or survival to discharge in dogs with neoplastic causes of septic peritonitis compared with dogs with nonneoplastic causes. There was a trend toward increased risk of death before discharge, with 74% of OSP patients surviving and only 58% of NSP patients surviving. This may not have been significant owing to a type 2 error, especially with neoplasia being a relatively rare cause of septic peritonitis. There may be multiple reasons for the small number of neoplastic cases in this study, including low incidence of gastrointestinal tumors in dogs and the possibility these cases may be euthanized without pursing surgery if evidence of neoplasia is suggested on preoperative imaging. The rarity of these cases has also been reflected in other studies. Cortellini et al. found only 6 out of 83 dogs (7%) with septic peritonitis to have a neoplastic cause,³ whereas Fink et al. evaluated 149 dogs with recurrent septic peritonitis, finding only 16 (11%) with a GI neoplastic cause.⁹

The most commonly diagnosed primary neoplastic causes of septic peritonitis were hepatocellular adenoma and GI lymphoma. Sixty-four percent of NSP patients were affected with gastrointestinal neoplasia, which made it the most common neoplastic cause of septic peritonitis in this population. This is congruent with previous literature documenting that the most frequent source of secondary septic peritonitis in dogs and cats is leakage of GI contents.^{1,6,11,23–28} It is important to note, however, that although lymphoma was the most common diagnosis in the GI (n = 3), two dogs each were diagnosed with GIST and sarcomas. It is possible that with a larger population, these statistics can change. Five dogs had septic peritonitis caused by rupture of an abscessed hepatic neoplastic mass, and of these dogs, the most common histopathological diagnosis was abscessation of hepatocellular adenoma. Hepatic abscessation, irrespective of the presence of neoplasia, has been documented to be commonly associated with bacterial agents,^29,30 likely due to bacterial colonization from hepatic filtration of the portal system vasculature. The necrosis that can occur in hepatic masses may predispose these tumors to forming abscesses, making this a logical finding. Generally speaking, lesions in the GI tract have little room to grow intraluminally before causing detectable signs in patients and leading to clinical investigation. In addition, tumor growth could result in a loss of intestinal wall integrity, leading to a high risk of abscessation and organ rupture. Contrarily, hepatic lesions, especially those located distal to the hilus, can grow to massive sizes before detection. It is unknown if patients with GI or hepatic neoplasia have any predisposition for developing septic peritonitis.

Several triage classification scores were used in this study, and although none of them were significantly different between NSP and OSP groups, all triage scores were significantly associated with survival to hospital discharge. This parallels the findings in the validation studies previously performed on these scoring systems to provide standardized data of reported clinical signs and complications at patient presentation.^18–20 Not enough triage data were available for every patient to receive an APPLE score; this was attributed to variables such as timing of presentation (i.e., overnight with limited diagnostics available) and clinician discretion of case management. When APPLE_full, APPLE_fast, and ATT scores were adjusted for, there was still no significant difference in survival to discharge between NSP and OSP patients. This finding supports that overall patient stability, rather than just the presence or absence of neoplasia, is more likely to influence the likelihood of short-term survival in patients with septic peritonitis.

The complex nature of septic peritonitis means that there may be many factors that could confound an association between neoplastic cause of septic peritonitis and survival. Besides triage scores, the myriad of variables including lesion location, comorbidities, and case management all contribute to the overall patient status and short-term prognosis. Although we adjusted for triage scores in a bivariable analysis, it is important to consider there may also be an inherent bias associated with a neoplasia diagnosis. Clinicians and owners may be more willing or likely to treat complications in OSP patients more aggressively, as they expect a good long-term outcome if the patient survives the short-term period. Conversely, clinicians and owners may more readily discontinue complication treatment in NSP patients, given the higher likelihood of disseminated disease in addition to a sometimes unknown or guarded long-term prognosis. It may be that this subjective bias contributed to the higher odds of NSP patients dying before discharge. However, this is a difficult factor to evaluate, especially in a retrospective study. Future studies may be able to identify other risk factors within NSP patients that contribute to the higher likelihood of death before discharge.

The presence of neoplasia was not a significant factor to development of intraoperative or immediate postoperative complications. The predominant intraoperative surgical complication described in this population was intraoperative death. This was not surprising given that surgical treatment of septic peritonitis is not an elective procedure and usually performed in debilitated animals. Immediate postoperative complications were varied and likely largely related to the location of the lesion and the degree of systemic compromise experienced by the patient. These findings emphasize the likely negligible degree in which neoplasia itself contributed to the complication rate. An SI perforation secondary to nonsteroidal anti-inflammatory drug use versus to GIST rupture would both cause similar clinical signs in a dog, not because of the primary causes but because they both lead to the same sequelae. This further alludes to the potential influence of clinician and owner bias on survival to discharge.

In this study, the dogs in the NSP group that survived to discharge generally did well in the long term. Four dogs that died following discharge but before manuscript preparation were euthanized because of unrelated causes, and two dogs were doing well at the time of data accrual. Of significance, the dogs that were alive were diagnosed with hepatocellular adenoma (n = 1) and GIST (n = 1), which tend to be more locally aggressive lesions and not readily disseminate. One dog was euthanized 250 days after initial surgery because of recurrence of septic abdomen secondary to abscessation of hepatocellular adenoma. It is unclear why there was recurrence of hepatic abscessation in this dog, as this has not been reported before. The dogs that were euthanized later were similarly diagnosed with locally aggressive lesions, whereas the dogs that did not survive to discharge had diagnoses including hemangiosarcoma and high-grade GI lymphoma. These tumors are typically associated with disseminated disease and, subsequently, poorer prognosis and shorter survival times.^31,32 As such, some unavoidable bias was introduced in the population that survived long term. It is reasonable to assume that the patients with malignant diagnoses with guarded to poor prognosis would do worse in the short term. Interestingly, three of the six dogs that did not have recurrence of their lesion in the long-term period had incomplete margins on histopathological examination of their original lesion. Two of these cases were GIST, one in the cecum and the other at the ileocecocolic junction. Although previous literature has documented superior long-term prognoses for surgically completely resected cecal GIST versus SI GIST, it is unknown if this pattern is reflected on incompletely resected GIST lesions.³³ The small case number in this study precluded any potential analysis that would be statistically significant. Regardless, this provides evidence that patients with neoplastic causes of septic peritonitis who survive the immediate perioperative period can do well long term.

The limitations of this study were largely attributed to the retrospective nature, as well as the relatively low number of NSP patients given this is a rare cause of septic peritonitis. There were several aspects of the study design that introduced inherent bias in the NSP population. This study specifically evaluated outcomes of NSP patients that underwent surgical treatment—it may be that cases that had metastatic disease or other comorbidities may have been screened out by euthanasia instead of proceeding to surgery. If these cases were surgically treated, the survival rate in the NSP patients may have been even lower. Also, one of the NSP cases included in the cohort was not definitively diagnosed as neoplasia via histopathology but rather by cytology. There were also cases that were excluded because of the lack of a definitive histopathological neoplastic diagnosis. Had these cases been included, the outcome of the NSP patients would have been biased unfavorably; it is reasonable to assume that most owners do not pursue further diagnostics once the decision to euthanize has been made. Lastly, there were two cases of septic peritonitis that were included on the basis of visualized purulent discharge intraoperatively. These were included because of the established propensity for bacterial infection associated with abscessation of these organs^29,30,34; however, inclusion of these cases could have negatively biased the outcome of both NSP and OSP populations.

Conclusion

In conclusion, the most commonly diagnosed primary neoplastic causes of septic peritonitis were hepatocellular adenoma and GI lymphoma. Neoplasia in itself was not a significant prognostic indicator for survival to discharge after surgical treatment of septic peritonitis, and patients in the neoplasia cohort who survived to discharge did well long term. The diagnosis of a primary, localized, neoplastic lesion alone should not deter clinicians and owners from pursuing treatment for septic peritonitis.