Atypical enteritis causing life-threatening pneumatosis intestinalis in a dog – radiographic and ultrasonographic findings Atypische enteritis die levensbedreigende pneumatosis intestinalis veroorzaakt bij een hond-radiografische en echografische bevindingen

A geriatric dog was presented for acute vomiting, anorexia and lethargy. Abdominal ultrasound was suggestive of the presence of gas within the small intestinal walls. Additional abdominal radiographs confirmed the ultrasonographic abnormalities, compatible with pneumatosis intestinalis. Explorative laparotomy revealed hemorrhagic lesions, thickened intestinal walls and serosal discoloration of the jejunum. Partial jejunectomy was performed and histopathology showed findings compatible with atypical bacterial enteritis. The dog recovered completely and did not show any clinical signs during a follow-up period of one year after surgery.

In human medicine, PI is subdivided in two groups primary PI, which is benign in nature, and secondary PI, which reflects a pathological, often life-threatening condition (Mitsuyoshi, 2015), requiring aggressive surgical and medical treatment (Hwang et al., 2016;Rajan et al., 2017). In primary PI, intramural gas creates a circular or cystic air pattern within the small intestinal wall, called pneumatosis cystoides intestinalis (Pear, 1998;Ho et al., 2007). In secondary PI, gas accumulates as a nodular or linear pattern within the small intestinal walls (Hwang et al., 2016;Rajan et al., 2017). In the majority of cases, the differentiation can be made based on ultrasound or radiography, although computed tomography is considered as the gold standard (Di Pietropaolo, 2020). Gas accumulation is more commonly seen in other organs, such as the wall of the stomach in gastric pneumatosis and emphysematous gastritis or in the wall of the large intestines in pneumatosis coli (PC) (Degner, 1992;Morris, 1992;Choi, 2009;Aste et al., 2005;Hwang, 2016).
To the authors' knowledge, primary PI and secondary PI each have been reported only once in dogs (Hedgespeth and Pérez, 2017;Song et al., 2013). In cats, life-threatening PI has been described in two case reports, respectively caused by idiopathic emphysematous hepatitis and a clostridium difficile infection (Hutchinson et al., 2018;Walczak, 2018). In this case report, the imaging findings are described of a dog with secondary PI who survived after aggressive surgical and medical treatment, which underlines the need for awareness of this condition.

Signalment, history and clinical findings
A twelve-year-old, spayed, female Labrador retriever was referred to the emergency service (Ghent University, Faculty of Veterinary Medicine, Small Animal Department, Merelbeke, Belgium) with com-plaints of anorexia, vomiting and lethargy for five days. The dog had no prior medical history. She received one subcutaneous injection of hyoscine butylbromide/ metamizol sodium (Buscopan compositum®, Boehringer Ingelheim, the Netherlands) and one intramuscular injection of prednisolone 0.6mg/kg (Prednisol 2,5%®, UDD diergeneesmiddel, the Netherands) prior to referral.
Upon presentation, abnormal physical examination findings included lethargy, mild tachycardia (140 bpm), prolonged capillary refill time and weak femoral pulse. The dog showed mild discomfort on abdominal palpation and was moderately dehydrated. Digital rectal examination demonstrated marked hematochezia. The patient was normotensive and normoglycemic.
The results of the blood examination obtained by the referring veterinarian prior to referral, revealed mild leukocytosis, mild anemia, mild hypoproteinemia, hypoalbuminemia, increased DGGR lipase and increased C-reactive protein (Tables 1 and 2).
Complete blood count was repeated at presentation and revealed moderate non-regenerative anemia and moderate leukocytosis (Table 2). Prothrombin time (PT)

Abdominal ultrasound
Abdominal ultrasound was elected as primary imaging modality. The main abnormalities were related to the gastrointestinal tract. The stomach was moderately filled with a mixture of gas and fluid content; no peristaltic movements or contractions were noted during the investigation, indicating a possible functional ileus. The gastric wall thickness and layering remained normal. In the mid-to caudal abdomen, a large section of the jejunum demonstrated corrugation and loss of normal layering. The alterations in the wall made it difficult to reliably identify the inner layers of the affected jejunal wall, although the serosal surface and adjacent hypoechoic muscularis layer seemed intact. More centrally, the submucosal layer and mucosal layer, normally identifiable as a hyperechoic and hypoechoic layer respectively, could not be delineated properly. At this level, a strongly hyperechoic and irregular interface was noted. Distal from this, a strong inhomogeneous shadow artefact or 'dirty shadow' was created, indicating abnormal gas content in the more central mural layers. The presence of this artefact hampered assessment of the entire jejunal wall at this level ( Figure 1).
There was no evidence of perforation of the jejunal wall and the adjacent mesenteric fat was within normal limits. The jejunum immediately proximal and distal to this lesion had a normal appearance and diameter, thereby excluding complete mechanical obstruction. The other intestinal segments demonstrated normal wall layering, thickness and content.
A moderately enlarged liver with a homogeneous, mildly hyperechoic parenchyma was noticed. Dissecting between the liver lobes and surrounding the gall bladder, a small amount of anechoic free fluid was seen, and aspiration of the fluid was performed. Cytologic examination of this localized effusion yielded a non-septic modified transudate.
Other abdominal organs were within normal limits. As the intramural gas of the jejunum prohibited further evaluation, abdominal radiographs were elected as adjuvant imaging technique.

Abdominal radiography
On the abdominal radiographs, a linear intramural gas opacity was seen in a part of the jejunal loops, parallel with the serosal surface, suspected to be at the level of the more central layers (mucosa and submucosa). At several locations, this linear gas opacity had a mildly undulating appearance, possibly consistent with minimal peristaltism or simply following  the conformation of the jejunal loop, while in other segments, it appeared straighter. A heterogeneous soft tissue material was visible within the lumen, with few gas bubbles dispersed within this material (Figures 2,  3 and 4).
A second jejunal loop, located in the cranioventral abdomen, was gas dilated and seemed corrugated. This segment was markedly shorter than the previous-ly described abnormal segment. A large portion of the gastrointestinal system was either empty or gas filled showing no evidence of an intestinal obstruction.
A mild to moderate loss of serosal detail was noted in the mid-to cranial abdomen, which could be attributed to the mild peritoneal effusion.
In conclusion, the preliminary diagnosis was jejunal PI. As differential cause of the jejunal wall alterations, a foreign object/neoplasia or bacterial enteritis with gas forming bacteria such as clostridium were postulated. A computed tomography scan was considered an adjunctive diagnostic tool, but explorative laparotomy was elected in order to allow diagnosis via histopathologic samples and treatment simultaneously.

Enterectomy and histopathology
Anesthesia was induced with midazolam 0.3 mg/ kg (Midazolam®, Mylan, Belgium) and alfaxalone 3 mg/kg (Alfaxan®, Jurox, United Kingdom). Afterwards, the dog was intubated and anesthesia was maintained using isoflurane (Isoflo®, Zoetis, Belgium). A continuous infusion of Fentanyl 5 µg/kg/min (Fentadon®, Dechra, the Netherlands) and lidocaine 30 µg/kg/min (Lidor®, Richter Pharma, Austria) was used to manage pain. Because of hypotension, a continuous infusion of norepinephrine was started after the administration of boli of crystalloids. Eventually, the dog received fresh frozen plasma during surgery because of constant low arterial blood pressure likely due to low colloid-oncotic pressure.
While inspecting the abdominal content, a hemorrhagic lesion in the pylorus and duodenum and a thickened mesenteric intestinal wall in the proximal part of the jejunum were noticed. This part shaded into a 30 cm intestine segment with a thickened wall and abnormal color; 20 cm caudal to this part, another 7 cm long abnormal segment was present. There was a minimal amount of abdominal free fluid noticed. The in total 66 cm abnormal looking segment was removed by enterectomy and end-to-end anastomosis. Cranial to this segment, a 30 cm piece of intestine with a thickening on the mesenteric side, was observed and left in place.
Histopathology demonstrated a pronounced multifocal hemorrhagic to ulcerative neutrophilic enteritis with the presence of a large number of bacteria described as short, gram-positive rods.
Parts of the intestinal wall were sent for bacterial culture and revealed the presence of E. coli. No fecal culture was performed.

OUTCOME
After surgery, the dog was monitored in the intensive care unit and was supported with omeprazole IV 1 mg/kg BID (Pantomed®, Takeda, Belgium), maropitant IV 1 mg/kg SID (Cerenia®, Pfizer, Belgium), enrofloxacin IV 5 mg/kg SID (Baytril®, Bayer, Germany), amoxicillin clavulanic acid IV 20 mg/kg TID (Augmentin®, GSK Belgium, Belgium), methadone IV 0.2 mg/kg q4h (Comfortan®, Dechra, the Netherlands) and crystalloids (ringer lactate, Braun, Germany). After receiving the results of the culture, enrofloxacin was discontinued. The dog developed severe pancreatitis post-operatively, but gradually recovered and was discharged seven days after surgery. One year after surgery, the dog was clinically doing well and did not show any symptoms anymore based on the owners' experience.

DISCUSSION
Pneumatosis is described as the presence of air or gas in abnormal locations in the body (Studdert et al., 2012) and has an uncommon but well-documented  occurrence in humans. Pneumatosis most commonly involves the hepatic and portal venous system but can also involve other organs such as the bowel, in humans mostly the colon. Specifically, pneumatosis intestinalis is a rare, well-described radiologic finding in humans (Palamidessi, 2011). It has been reported as the presence of air or gas-containing cysts in the subserosal or submucosal layer of the small intestines (Degner, 1992;Aste et al., 2005;Song et al., 2013).
In human medicine, PI can be found in a wide range of clinical settings and has been reported in humans of all ages and both sexes (Heng et al., 1995;Takase et al., 2017). To the authors' knowledge, only two previous cases of PI in dogs (Song et al., 2013;Hedgespeth en Pérez, 2017), two cases in cats (Hutchinson et al., 2018;Walczak, 2018) and a case series of foals (Navas de Solis, 2012) have been described.
The pathogenesis of PI is still unclear, even in human medicine. Four hypotheses of etiopathology have been proposed suspecting either a mechanical, pulmonary, bacterial or mucosal damage origin (Aste et al., 2005;Degner, 1992;Hedgespeth and Pérez, 2017;Pear, 1998;Silveira et al., 2018). In the mechanical theory, an increase in the intraluminal pressure caused by mechanisms, such as intestinal obstruction, vomiting and diarrhea, ileus or endoscopy resulting in the dissection of gas into the intestinal wall through tiny defects or forcing gas through pre-existing ulcers into the intestinal mucosa is proposed. The pulmonary hypothesis has been formulated in patients with chronic obstructive pulmonary disease. Severe coughing may result in alveolar rupture and dissection of air along vascular channels in the mediastinum, tracking caudally to the retroperitoneum and then in the subserosa of the gastrointestinal tract (Ho et al., 2007;Degner, 1992;Pear, 1998;Choi et al., 2009). Alternatively, the bacterial model suggests that gas-producing bacteria enter the submucosal intestinal layer through areas of increased mucosal permeability or mucosal breaks, resulting in intramural gas formation (Degner, 1992;Pear, 1998;Aste et al., 2005;Choi et al., 2009). The mucosal damage theory on the other hand suggests that luminal gas under normal physiological pressure is forced into the gastrointestinal wall through injured mucosa (e.g. ulceration) (Silveira et al., 2018). However, PI is most likely caused by a combination of both mechanical and infectious factors (Pear, 1998;Gagliardi, 1996;Brown, 1999, Mitsuyoshi, 2015. Two types of PI have been described in human medicine, called primary or benign PI and secondary or malignant (or life-threatening) PI. Those patterns can be distinguished from each other based on abdominal radiographs or computed tomography (Pear, 1998, Choi et al., 2009. This differentiation has been better described in the stomach and colon (Thierry et al., 2019). Primary PI creates a circular or cystic air pattern within the small intestinal wall, called the curvilinear pattern, which fits with pneumatosis cystoides intestinalis (Pear, 1998;Ho et al., 2007). The findings in the life-threatening form of PI consist with a linear pattern of gas within the small intestinal walls (Hwang et al., 2016;Rajan et al., 2017). Also, sometimes this form appears together with the presence of hepatic portal or porto-mesenteric venous air, which can be noticed on computer tomography (Sud et al., 1996;Ho et al., 2007, Hwang et al., 2016. A feature that is considered an additional sign of a malignant process is the unchanged presence and location of intramural gas upon changing patient position. (Sud et al., 1996, Fischetti et al., 2004. Although those signs can give an indication of the form which the body is dealing with, radiographic or computed tomography appearance alone does not allow the differentiation between the two in all cases (Pear, 1998;Ho et al., 2007).
Secondary PI can be associated with a variety of illnesses ranging from a fairly benign process to lethal ones (Hutchinson et al., 2018). The etiology of gastro-intestinal pneumatosis in dogs is mostly associated with intestinal neoplasia, ulcers, mucosal tears, maliciously introduced foreign bodies, bacterial overgrowth, gastric torsion, trauma, pulmonary disease, diabetes mellitus and iatrogenic mucosal injury (Aste et al., 2005;Hwang et al., 2016). The patient of the present report had no history of previous trauma, surgery or suspected foreign body ingestion. The clinical signs started acutely, and explorative laparotomy and histopathology results showed severe enteritis with intramural bacteria as the underlying cause.
Culture of the intestinal wall revealed the presence of E. coli. Although E. coli is part of the normal gastro-intestinal flora, it is equally plausible that they did not play a causative role in this dog's illness. The reported bacteria in the wall on histopathology (rods) could fit with clostridium species, which have been reported in association with gastro-intestinal pneumatosis in both human and veterinary medicine. The gold standard of diagnosis would be the confirmation with a fecal sample. Aerobic and anaerobic microbial cultures of feces, fecal smears, fecal flotation and the detection of clostridium perfringens enterotoxin in feces should be considered in case of suspicion of PI secondary to bacterial enteritis (Aste et al., 2005).
Symptoms which are commonly seen in human patients with PI are abdominal pain, nausea, vomiting, hematemesis, fever, chills, tachycardia, tachypnea (Guillén, 2015). The patient in this report showed signs of anorexia, lethargy, vomiting and abdominal pain, similar to the signs described in the two previous canine and two feline case reports (Hedgespeth and Pérez, 2017;Hutchinson et al., 2018;Song et al., 2013;Walczak et al., 2018).
On complete blood count, leukocytosis and mildmoderate non-regenerative anemia were noticed. The anemia in combination with hypoproteinemia was most likely secondary to the severe intestinal bleeding that was noticed on digital rectal examination and during surgery. Anemia has also been reported in other cases of PI previously described in the literature (Hedgespeth and Pérez, 2018;Hutchinson et al., 2018;Song et al., 2013;Walczak et al., 2018).
In veterinary medicine, the first diagnostic tools to detect PI are usually abdominal radiography and ultrasonography. Computed tomography is considered the gold standard for establishing a definitive diagnosis in humans because it is the most sensitive imaging modality for the identification of small volumes of air within soft tissue (Di Pietropaolo, 2020;Ho et al., 2007). Computed tomography appears to be more sensitive than abdominal radiographs for the detection of hepatic portal and porto-mesenteric venous air, the presence of which can increase the suspicion of PI. The detection of hepatic portal or porto-mesenteric venous air can help distinguish benign causes of PI from lifethreatening causes of PI as the former has been shown to indicate transmural infarction (Ho et al., 2007).
Ultrasonographic findings of PI have been described as bright echoes within the layers of the small intestinal wall in humans (Aste et al., 2005), which were both seen in this case and in a dog and cat in other case reports considering PI (Hedgespeth and Pérez, 2018;Song et al., 2013;Walczak, 2018). However, differentiating intraluminal gas from intramural gas by ultrasonography may be difficult in some cases due to alteration of the normal appearance of the wall and severe acoustic interference of the gas (Song et al., 2013).
In humans, different therapies have been described for primary (benign) and secondary (life-threatening) PI. Treatment primarily includes resolution of the primary inciting cause. Therapy for life-threatening intestinal pneumatosis includes an aggressive medical and surgical treatment. Early explorative laparotomy is always indicated when there is a suspicion of necrotic bowel. Instead, aggressive medical management is warranted in non-surgical cases and includes hospitalization with intravenous fluids, antibiotics, analgesia and close monitoring. If the patient deteriorates during medical management, reimaging and exploratory laparotomy should be considered (Hutchinson et al., 2018). Therapies which have been described for benign gastro-intestinal pneumatosis are bowel rest, total parenteral nutrition and potentially broad-spectrum antibiotic treatment that includes anaerobic coverage (Aste et al., 2005;Fischetti et al., 2004Guillén, 2015Lang et al., 2011;Song et al., 2013).
Because of the uncommon presentation and the possibility of a foreign object or underlying neoplastic process, explorative laparotomy was elected in this case. Based on the severity of the jejunal lesions, enterectomy was performed and several samples were taken for histopathology.

CONCLUSION
In this care report, the imaging findings and treatment of a dog with secondary PI are reported. PI is an extremely rare disease in veterinary medicine and to the authors' knowledge, this is the first case report of a dog with secondary PI who survived after aggressive surgical and medical treatment.