Abstract
Objectives
The aim of this study was to describe the clinical characteristics and the frequency of malignant vs non-malignant causes for spontaneous gastrointestinal perforation in cats.
Methods
The medical records of cats diagnosed as having gastrointestinal perforation between August 2010 and July 2013 were reviewed. Diagnosis was confirmed by exploratory surgery. Patients with incomplete records, perforation due to external trauma, leakage at previous enterotomy or anastomotic sites, or foreign bodies were excluded. Each record was examined for different information pertaining to signalment, medical history, clinical and clinicopathological data, imaging findings, abdominal fluid examination, surgical findings, histopathological examination, treatment received after surgery and outcome.
Results
Thirteen cats were included. Five of these cats had concurrent illnesses, including viral upper respiratory tract disease, pancreatitis and chronic kidney disease. Two cats had previously received non-steroidal anti-inflammatory drugs and four had received corticosteroids. Clinical signs and clinicopathological abnormalities were not specific. Six of 13 patients were diagnosed during surgery with gastric perforations, four patients with duodenal perforations and three patients with jejunal perforations. Histopathological examination of the ulcerated wall was performed in 11/13 cats. Alimentary lymphoma was diagnosed in six cats. Non-neoplastic lesions (lymphocytic– plasmacytic inflammatory bowel disease, necrotic suppurative enteritis) were observed in the other five cats. The major limitation of the study was the small sample size.
Conclusions and relevance
Lymphoma may be a frequent cause of spontaneous perforation in cats. Therefore, histological examination of ulceration is essential in all cases. The direct and sole implication of anti-inflammatory administration in a gastrointestinal perforation is not clearly established in this study.
Introduction
Gastrointestinal perforation is a life-threatening situation in veterinary medicine, and may lead to septic peritonitis and death if left untreated. It is defined as being spontaneous when it arises in the absence of foreign body ingestion, gastric dilatation and volvulus, external trauma, leakage at previous gastrotomy or anastomotic sites, or iatrogenic trauma.1,2 In humans, the most common cause of perforation is peptic ulcer.3,4 Spontaneous perforation due to diverticulitis or coexisting malignancy are other, much less common, causes. 5 Previous administration of non-steroidal anti-inflammatory drugs (NSAIDs) and, to a lesser extent, corticosteroids has been shown to increase the risk of peptic ulcer complications (eg, perforation and haemorrhage) three- to fivefold. 6
In cats, gastrointestinal perforation has been uncommonly reported. Previous reports have consisted of individual or small case series, and little is known about the clinical characteristics, frequency and predisposing factors for spontaneous digestive perforation in cats.7–23 Causes similar to those seen in dogs seem to be involved, including NSAIDs,24,25 corticosteroids,25,26 hepatic disease, 27 uraemia,28,29 periods of high stress, shock, other causes of decreased gastric circulation,30–32 gastric hyperacidity due to systemic mastocytosis or gastrinoma, 33 other gastrointestinal neoplasia34 –39 or idiopathic inflammatory bowel disease (IBD).10,40 Rare cases of hypereosinophilic syndrome, 8 staphylococcal granuloma and toxicity following ingestion of Dieffenbachia species leaves have also been described.9,16 However, no cause for perforation was found in a significant proportion of reported cases.21,23,41
The objectives of this study were to describe the clinical characteristics and the frequency of malignant vs non-malignant causes for spontaneous gastrointestinal perforation in cats. The main hypothesis, based on our clinical experience, is that lymphoma is one of the most common causes of spontaneous perforation, and, by contrast, NSAID and corticosteroid administration are rarely implicated in this phenomenon.
Materials and methods
Animals
The medical records of cats diagnosed between August 2010 and July 2013 at the Veterinary Hospital Fregis as having gastrointestinal perforation were reviewed. To be included in the study, the diagnosis had to be confirmed by exploratory surgery. The decision to perform surgery was based on the presence of septic peritoneal fluid on abdominocentesis, pneumoperitoneum or suspicion of gastrointestinal perforation on abdominal radiographs and/or ultrasound. Patients with incomplete records, perforation due to external trauma, leakage of previous enterotomy or anastomotic sites, or foreign bodies were excluded.
Each record was examined for information pertaining to signalment; history; concurrent illnesses; treatments received within the previous 3 months; clinical signs; physical examination; clinicopathological data; radiographic, endoscopic and ultrasonographical findings; cytological examination and bacterial culture of abdominal fluid obtained by abdominocentesis; gross findings at surgery; location of perforation; histopathological examination; treatment received after surgery; and outcome.
A boarded radiologist reviewed all the radiographic and ultrasonographical records. A boarded pathologist reviewed all the histopathology slides.
The descriptive statistical analyses are expressed as a mean ± SD.
Results
Signalment
The medical records of 21 cats with gastrointestinal perforation were identified for review. Eight records were excluded for the following reasons: foreign bodies (n = 5) and incomplete medical history (n = 3). Thirteen cats were included in this study. Among these cats, different breeds were represented, including domestic shorthair (n = 11), Persian (n = 1) and Siberian (n = 1). Seven male (six neutered and one entire) and six female (five neutered and one entire) patients were identified. No gender predilection was noted. The age of cats with gastrointestinal perforation ranged from 9 months to 17 years.
History and clinical presentation
The feline immunodeficiency virus (FIV)/feline leukaemia virus (FeLV) status was known in 7/13 cats: all were negative. Five of 13 cats had concurrent illnesses, including viral upper respiratory tract disease (3/13), pancreatitis (1/13), chronic kidney disease (1/13), hypertrophic cardiomyopathy (1/13), dermatophytosis (1/13) and hyperthyroidism (1/13). Treatments received within the previous 3 months were as follows: antibiotics (6/13: metronidazole, enrofloxacin, marbofloxacin, amoxicillin, cephalexin, gentamicin); NSAIDs (2/13: meloxicam) or corticosteroids (4/13: prednisolone, dexamethasone); H2 receptor blockers (4/13: cimetidine, ranitidine); antiemetics (3/13: maropitant, metoclopramide); angiotensin-converting enzyme inhibitor (1/13: benazepril).
Common clinical signs of gastrointestinal perforation in this study are presented in Table 1. The most frequently reported signs included inappetence (11/13), vomiting (8/13) and lethargy (7/13), but signs also included weight loss (5/13), abdominal distension (4/13) and melaena (1/13). Further clinical signs observed during physical examination were dehydration (5/13), hyperthermia (3/13), abdominal pain (3/13), pallor (2/13), dyspnoea (2/13) and icterus (1/13).
Clinical signs in the 13 cats
Diagnostic, surgery and necropsy findings
Complete biochemical and haematological analysis was performed in 8/13 patients. The most common abnormalities were hypoalbuminaemia (7/8 cats; 18–29 g/l [reference range 30–40 g/l]) and a neutrophilic leukocytosis (6/8; 13.0–47.9 cells/µl [reference range 3.0–11.0 cells/µl]). Two cats had normocytic normochromic regenerative anaemia (haematocrit 13% and 18%, respectively [reference range 25–45%]). One cat treated for chronic kidney disease had an increase in uraemia (1.8 g/l [reference range 0.2–0.5 g/l]) and creatininaemia (2.8 mg/dl [reference range 0.6–1.6 mg/dl]). Two cats showed an increase in alanine aminotransferase activity (93 U/l and 132 U/l, respectively [reference range 8–80 U/l]).
Abdominal radiographs were assessed in 6/13 patients. Free abdominal gas was observed in 4/6 cats. Generalised ileus was noted in one cat, and focal ileus in two cats.
Abdominal ultrasound was performed in 10/13 patients. Main abnormalities included hyperechoic mesenteric fat (6/10 cats), focal wall thickening (6/10 cats), focal loss of wall layering (6/10 cats), ulceration (4/10 cats), presence of free peritoneal fluid (4/10 cats) or free gas (3/10 cats), fluid-filled stomach and segmental fluid dilation of intestinal loops (2/10 cats), and reduced motility of stomach and small intestine (2/10 cats).
Cytopathological examination of abdominal fluid was performed in 3/13 patients before surgery. The amount of fluid was insufficient to be aspirated in the fourth cat diagnosed with free peritoneal fluid. Septic suppurative inflammation was present in these three cats. Aerobic and anaerobic bacteriological culture of the abdominal cavity or abdominal fluid was performed before or during surgery in 5/13 patients. Two of them were negative, one was positive for Enterobacter cloacae and Klebsiella pneumoniae, one was positive for Escherichia coli and one was positive for Streptococcus species.
Endoscopy was not performed in any of the 13 cats. Definitive diagnosis of gastrointestinal perforation was made at surgery. Six of 13 patients were diagnosed with gastric perforations, four patients with duodenal perforations and three patients with jejunal perforations. Detailed surgical findings regarding the peritoneal cavity were available for all patients. Generalised peritonitis was documented in three cats, and localised peritonitis was documented in 10 cats. Regional adenomegaly was observed in seven cats.
Histopathological examination of the ulcerated wall was performed in 11/13 cats. Alimentary lymphoma was diagnosed in six of these 11 cats. High-grade lymphomas, defined by a high mitotic rate, were diagnosed in four cases. Low-grade lymphomas, composed of small cells with a low mitotic rate, were diagnosed in two cases. Non-neoplastic lesions (lymphocytic–plasmacytic IBD, necrotic suppurative enteritis) were observed in the other five cats. The age of cats diagnosed with lymphoma ranged from 3–13 years, and the age of cats with non-neoplastic digestive disease ranged from 9 months to 17 years.
Treatment outcome
Various medical regimens were instituted prior to and after diagnosis of the perforation and commonly included intravenous antimicrobials, fluid therapy, gastrointestinal cytoprotectants and antiemetics. Abdominal surgery consisting of resection and anastomosis of the perforated segment was performed in all cats. Data are summarised in Table 2. Among the four cats diagnosed with high-grade lymphoma, two received oral chemotherapy with chlorambucil and corticosteroids, one received injectable polychemotherapy, including L-asparaginase, vincristine and corticosteroids, and one cat was euthanased upon finding out the histology results. The two cats diagnosed with low-grade lymphoma received oral chemotherapy with chlorambucil and corticosteroids.
Description of age, previous treatments, concomitant disease, feline immunodeficiency virus (FIV)/feline leukaemia virus (FeLV) status, tests that were undertaken prior to surgery, histological examination, perforation localisation, chemotherapy and survival time for each cat
CKD = chronic kidney disease; HCM = hypertrophic cardiomyopathy; NP = not performed
Three of 13 cats were lost to follow-up after surgery. The date of death was known in 8/13 cats. Three cats died or were euthanased in the postoperative recovery period (one cat died the day after surgery and two cats died 2 days after surgery) and no histological examination was performed. Of the seven surviving patients, two cats were still alive at the end of this study (>900 days) and five cats were euthanased 12–146 days after surgery.
Discussion
Digestive mucosal blood flow, bicarbonate buffer and mucous layer secretion play a large role in preventing autodigestion of the gastrointestinal mucosal lining in the normal patient. 42 Perforation develops when these cytoprotective mechanisms are overwhelmed. Spontaneous perforation mechanisms previously described include gastric hyperacidity; ischaemia; mucosal trauma or destruction by infiltrative disease; NSAID or corticosteroid administration inhibiting prostaglandin-stimulated mucosal blood flow and mucous layer secretion; and mastocytosis causing increased histamine release and inducing increased gastric acid production. 39
In this study, wall destruction by lymphoma was the most common histological finding (54%). Only a few similar cases have been previously described in cats.7,18,19 Lymphoma comprises nearly 30% of all feline tumours, and is the most common intestinal tumour, representing 55% of cases in an epidemiological survey of 1129 intestinal tumours in cats. 43 As with many cancers, incidence of alimentary lymphoma increases with age. But some earlier studies report younger median ages, most likely a result of a larger percentage of FeLV-positive cats in the study population. 44 In the present study, the FIV/FeLV status was known in 7/13 cats, among which were three of the six cats diagnosed with lymphoma, and all of them were negative. The youngest cat diagnosed with lymphoma was 3 years old and its FeLV status was unknown. This study demonstrates that lymphoma may be a frequent cause of spontaneous perforation in cats.
Gastrointestinal perforation secondary to NSAID or corticosteroid treatment has been described in both cats and dogs.17,18,20,24–26,45–48 NSAIDs induce gastrointestinal tract perforation through direct contact injury to gastric epithelial cells and inhibition of cyclooxygenase.39,47,48 Despite experimental evidence of the adverse effects of NSAIDs in cats, NSAID-induced gastroduodenal perforation has only been reported in a few clinical cases.48,49 In our study, only two cats had received NSAIDs. These drugs were prescribed to treat hyperthermia and were then administered after the beginning of the clinical signs. Furthermore, the histological results of one of these two cats showed lymphoma. Histological examination was not performed in the second cat. The direct and single implication of NSAID administration in the gastrointestinal perforation is therefore not clearly established in this study.
Corticosteroids can also cause gastrointestinal perforation by decreasing epithelial cell turnover, which increases gastric mucosal susceptibility to acid injury. 47 Additionally, gastric mucus quantity and viscosity may decrease. Four cats in this study had received corticosteroids. Three of them were treated with steroids for chronic vomiting and weight loss for several months. Two of these three cats were finally diagnosed with alimentary lymphoma. The last one had no histological examination performed. The link between corticosteroid administration and perforation is not established in this study. Furthermore, the referring veterinarian prior to referral had treated some animals with corticosteroids or NSAIDs, and, for this reason, information on dosage and duration of treatment was not uniformly available. Therefore, it could not be determined whether complications were dose related.
In addition to anti-inflammatory drug administration, some concurrent diseases have been associated with a higher risk of gastrointestinal perforation in dogs and cats, including inflammatory bowel disease, non-neoplastic infiltrative gastrointestinal disease, gastrointestinal neoplasia, mastocytosis, hepatic disease, septicaemia, gastrinoma, pyloric outflow obstruction, and the stress of major surgery or illness.7–40 Concurrent diseases identified in this study included viral upper respiratory tract disease, hypertrophic cardiomyopathy, dermatophytosis and hyperthyroidism. Although viral upper respiratory tract disease has not been specifically identified as an ulcerogenic condition, it is often the target of NSAID therapy. But none of the three cats with upper respiratory tract disease had recently received NSAIDs. Cardiomyopathy and dermatophytosis, as well as their treatment, have never been described as predisposing factors for digestive perforation. In humans, hyperthyroidism may coexist with peptic ulcer in rare cases. 50 Nevertheless, the hyperthyroid patient presented in this study was diagnosed with a lymphoma by histopathological examination.
Two other cats were concurrently treated for suspected pancreatitis (based on specific pancreatic lipase immunoreactivity test and abdominal ultrasound) and chronic kidney disease. These diseases are known as predisposing factors for digestive perforation in humans, as well as in dogs and cats. However, the cat with pancreatitis and spontaneous perforation was diagnosed with lymphoma. Consequently, in this case we can speculate that the pancreatitis was a consequence of the septic peritonitis and not the cause of the perforation.
The most common clinical presentations in this study were non-specific signs of gastrointestinal or systemic illness (decreased appetite, vomiting, lethargy, weight loss, abdominal distension, dehydration, hyperthermia, abdominal pain, pallor, dyspnoea and icterus). This is in agreement with previous reports.1,19,51 These signs did not allow clinicians to make a final diagnosis of gastrointestinal perforation. Even if most patients initially presented with vital parameters such as temperatures, pulses and respiratory rates that were within reference ranges, abdominal pain was frequent, and concurrent administration of NSAIDs or corticosteroids in some animals may have masked clinical signs associated with acute peritonitis. 52
In addition, signs referable to digestive bleeding (haematemesis, melaene) were observed in only 1/13 cats, and decompensated shock was not noted in any of the cats. This finding may be owing to the lack of observation by owners or veterinary personnel, occult haemorrhage or a relatively low number of ulcers that bleed in this species. However, 2/8 cats in the present study had mild-to-severe normocytic normochromic regenerative anaemia. Anaemia was detected in up to 70% of cats with intestinal neoplasia in another study.44,53–57 Haemolysis or occult blood loss (eg, gastrointestinal haemorrhage) could account for the anaemia in these patients. Mild-to-moderate hypoalbuminaemia was also a common finding in this study (7/8 cats) and may be secondary to blood loss, chronic inflammation or malabsorption. Mild-to-moderate neutrophilic leukocytosis was observed in 6/8 cats. This haematological abnormality may be secondary to the focal or generalised peritonitis observed in these cats. But leukogram changes, including leukocytosis, are also common in dogs (25–70%) and cats (40%) with intestinal tumours without perforation and peritonitis.54,56,57
After considering history, clinical signs and blood results, there were no individual diagnostic tests that were consistently instrumental in making the diagnosis of gastrointestinal perforation prior to surgery in this study. Abdominal radiography, abdominocentesis with cytopathology and abdominal ultrasound were all helpful in increasing the clinical suspicion of perforation. This is in agreement with a previous study. 1
In this study, abdominal radiographs were assessed in 6/13 patients. Free abdominal gas was observed in 4/6 cats. Therefore, radiographs may be an important tool for diagnosing gastrointestinal perforation.
Sonographical suspicion of perforation was made in 3/10 cases. The finding of free gas with appropriate historical and clinical signs, in the absence of recent abdominal surgery or abdominocentesis, appeared in a previous study to be highly suspicious for ruptured gastrointestinal viscus. 58 In the absence of pneumoperitoneum, a combination of indirect sonographical findings (hyper-echoic mesenteric fat, peritoneal effusion, fluid-filled stomach and intestinal loops, reduced motility, focal wall thickening and loss of wall layering) may be associated with perforation. Early seal of the perforation by the omentum may prevent large amounts of fluid or free air accumulation within the peritoneal cavity and decrease the ability of ultrasound or radiography to indicate signs consistent with bowel rupture. 58 The large proportion of patients in this study (5/11 cats) in which the perforation was identified as having chronic characteristics may reflect a clinical inability to predict and diagnose bowel perforation adequately.
In previous reports, the location of the perforating gastric ulcers was almost exclusively duodenal, pyloroantral or fundic in cats with non-neoplastic gastric lesions. 19 In this study, 6/13 patients were diagnosed with gastric perforations, four patients with duodenal perforations and three patients with jejunal perforations. Lymphoma was equally associated with these three locations, which is not surprising considering the diffuse infiltration by nature of alimentary lymphoma observed in most cats.
The major limitation of the study was the small size of the sample population (n = 13). Because of this limitation, it is hard to draw firm conclusions with regard to causation, specificity and sensitivity of diagnostic techniques, efficacy of therapies and prognosis.
Conclusions
In the absence of pneumoperitoneum, clinical signs and clinicopathological abnormalities are not specific enough to allow final diagnosis of gastrointestinal perforation. Signs of collapse and shock are surprisingly rare in cases of perforation. Clinicians should use multimodality diagnostic procedures to avoid delay in diagnosis. Histological examination of a surgically resected resection portion is essential, as lymphoma should be suspected in all cats presented with spontaneous perforation.
Footnotes
Acknowledgements
We would particularly like to thank the veterinary laboratory Laboratoire d’Anatomie Pathologique Vétérinaire du Sud-Ouest (LAPVSO) for reviewing all histopathological slides.
Conflicts of interest
The authors do not have any potential conflicts of interest to declare.
Funding
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.