Introduction

The most common complications associated with pancreatic surgery include pancreatitis and local peritonitis.^1,2 Pseudocyst formation in organs other than the pancreas following pancreatic surgery or severe pancreatitis is an uncommon but well-studied complication in human medicine.^3–11 The development of pancreatic pseudocysts has been suggested to arise from damage to the main pancreatic duct or its branches.¹² Human studies have shown that in >50% of asymptomatic patients, these intramural gastrointestinal (GI) pseudocysts can resolve without treatment. In symptomatic patients, treatment options include percutaneous ultrasound-guided drainage, endoscopic drainage, or surgical intervention, depending on the location of the pseudocyst. To the authors’ knowledge, this phenomenon of pancreatic pseudocysts forming in organs outside of the pancreas has not been reported in veterinary medicine. The goal of this case report and comprehensive human literature review is to raise awareness of this disease process in dogs and provide information on the pathophysiology, diagnosis, management, prognosis, and long-term outcome.

Case Report

A 9.5 yr old female spayed Yorkshire terrier was presented for a 2 mo history of progressive weakness and persistent hypoglycemia on serial blood work. Previous history included chronic intermittent vomiting and lethargy for 1.5 yr. These episodes of vomiting and lethargy occurred every 3 to 4 mo and were suspected to be secondary to mild pancreatitis and/or gastroenteritis, because signs would resolve with medical management (gastroprotectants, antiemetics, fluid therapy, and a bland diet).

Initial diagnostics included complete blood cell count and urinalysis, which were within normal limits, and a chemistry panel, which showed a marked hypoglycemia of 39 mg/dL (reference 70–138 mg/dL) and hypophosphatemia of 1.4 mg/dL (reference 2.5–6.0 mg/dL). A resting cortisol and serum bile acids were within normal limits. A sample of serum insulin and blood glucose were measured concurrently for an amended insulin to glucose ratio. Serum insulin levels revealed elevated insulin of 49.0 μU/mL (reference 7.5–20 μU/mL) and hypoglycemia of 51 mg/dL (reference 70–138 mg/dL), consistent with an insulinoma.

Because of the high suspicion for an insulinoma, a triple-phase computed tomography (CT) scan of the thorax and abdomen was performed for diagnostic staging and surgical planning. The pancreatic triple-phase CT was performed according to a previously described method.¹³ A noncontrast scan was run before iodinated contrast administration. Iohexol^a was used as the contrast medium at a dose of 2 mL/kg and injected over 5 s. Scan delay was fixed and set as 15 s for the arterial phase, 30 s for the pancreatic phase, and 90 s for the delayed phase after contrast medium injection started. CT revealed a 5 mm diameter, hypoattenuating nodule with marginal ring enhancement in the body of the pancreas. The nodule was isoattenuating on the precontrast scan and consistently hypoattenuating with marginal ring enhancement in the arterial, pancreatic, and delayed phases. Although the nodule was hypoattenuating compared with surrounding pancreatic parenchyma, precontrast Hounsfield unit was 50, and >100 Hounsfield unit in the pancreatic phase. Thus, differentials for the current nodule on CT included probable insulinoma versus less likely nodular hyperplasia and adenoma. The right kidney had mild peridiverticular mineralization; otherwise, all other structures, including the liver and all regional peripancreatic and hepatic lymph nodes, were within normal limits.

Exploratory laparotomy was recommended based on serum insulin to glucose ratio and the pancreatic nodule visualized on CT. Exploratory laparotomy revealed a firm, 15 mm × 4 mm, well-encapsulated, pink, raised, multilobulated nodule in the body of the pancreas, as well as palpable thickening along the distal end of the right limb of the pancreas ∼2.4 cm aborad to the minor duodenal papilla. The proximal duodenum was palpably thickened, and there were two small (∼4 mm and 5 mm), tan, slightly raised nodules in the quadrate liver lobe that were not visualized on CT. All other abdominal organs were grossly normal. The mid-body pancreatic nodule was enucleated as previously described, using a combination of blunt dissection and electrocautery.^1,14 A vessel-sealing device^b was used to perform a separate partial pancreatectomy of the thickening at the distal end of the right limb of the pancreas, ∼2.4 cm aborad to the minor duodenal papilla. A full-thickness wedge biopsy of the proximal duodenum was performed and closed with 4-0 polydioxanone in a simple interrupted suture pattern. The two quadrate liver lobe nodules were each biopsied with a 5 mm dermal punch biopsy, and the defect was filled with gelatin sponge^c. Peripancreatic and hepatic lymph nodes were assessed as small and unremarkable.

Histopathology of the pancreatic nodule was consistent with a pancreatic islet cell tumor (insulinoma), and the thickening of the distal end of the right limb of the pancreas was consistent with mild lobular exocrine hyperplasia. The duodenal histopathology was consistent with normal to mild lymphoplasmacytic enteritis with eosinophils. The liver histopathology showed only moderate panlobular hepatocellular hydropic change with hepatocellular bile stasis, mild to moderate lymphocytic portal hepatitis, and no evidence of metastatic insulinoma. The dog recovered uneventfully, and the blood glucose value was within normal limits (98 mg/dL) at extubation after the operation. The dog remained euglycemic on serial monitoring (q 4 hr) during the first 48 hr (90–120 mg/dL) and was discharged 2 days after the operation after the return of a normal appetite and stabilization of blood glucose levels.

Fourteen days after the operation, the dog was admitted because of anorexia, regurgitation, and vomiting. The dog was hypochloremic (109 mEq/L) and hyponatremic (136.3 mEq/L) with a metabolic alkalosis (pH 7.56, HCO₃₋ 41.2 mEq/L, pCO₂ 44 mm Hg), which was concerning for a proximal GI obstruction. An abdominal ultrasound was subsequently performed and revealed a thick-walled cystic lesion (6.3 cm × 4.5 cm), causing marked compression, ventral displacement of the gastric body and pylorus, and suspected gastric outflow obstruction as evidenced by a markedly distended fluid-filled stomach and empty small intestines. The origin of the lesion was difficult to discern with ultrasound but was suspected to be either in the body of the pancreas (i.e., pancreatic origin) or from the dorsal gastric wall (Figure 1A). Differentials included pancreatic pseudocyst, pancreatic or gastric wall abscess or cyst, or, less likely, neoplasia. Other findings included an irregularly thickened duodenal muscularis layer and evidence of concurrent pancreatitis.

View Figure

• Download as Powerpoint
• Download Figure

A second exploratory laparotomy was performed, and gastric outflow obstruction was noted secondary to a large firm cystic lesion along the dorsal gastric wall that was noted on the ultrasound. A large fluid-filled stomach with contents that could not be milked into the duodenum was noted. Although there was evidence of pancreatitis (erythema and edema) of the entire pancreas, the cystic lesion was not directly associated with the pancreas (Figure 1B). Fluid from the lesion (dark brown–green and gelatinous without odor) was aspirated for fluid analysis and cytology, aerobic and anaerobic culture and sensitivity, and for amylase and lipase levels to be compared with concurrently drawn intraoperative blood serum samples (Figure 1C). A flexible video endoscope was advanced intraoperatively into the esophagus. The esophageal mucosa and lower esophageal sphincter appeared normal. All portions of the stomach were examined, and the gastric mucosa appeared normal. The intramural location of the lesion along the dorsal wall of the body of the stomach approaching the pylorus was confirmed when illumination from the endoscope appeared more translucent because of the thinness of the stomach wall (Figure 1D). The scope was advanced through the pyloric sphincter, and the proximal duodenum was evaluated and appeared normal. The cystic lesion was drained, locally lavaged, biopsied, and omentalized. Pyloric outflow obstruction noted on initial examination of the abdomen was resolved because gastric contents moved freely into the duodenum with peristalsis. A jejunostomy tube (J-tube) was placed to bypass the compromised stomach and the pancreas in order to provide a source of enteral nutrition. The tube (8 Fr Argyle feeding tube^d) was placed percutaneously from the right caudal abdominal wall into the proximal aspect of the jejunum aborad to the duodenal flexure. A small stab incision was made into the jejunal lumen on the antimesenteric border; the tube was inserted ∼14 cm into the jejunum and secured with two purse-string sutures using 4-0 polydioxanone at the entry site. The jejunostomy site was anchored to the right caudal abdominal wall with a four-quadrant jejunopexy using 4-0 polydioxanone and secured to the skin with a purse-string and finger trap pattern using 3-0 nylon. A Jackson-Pratt drain was placed percutaneously through the left caudal abdominal wall to allow for continuous closed suction drainage after the operation. The Jackson-Pratt drain was anchored to the skin in a similar fashion as the J-tube.

Histopathology of the gastric wall cystic lesion was consistent with chronic fibrosing steatitis with fat saponification. All changes were inflammatory and reactive in character, with no neoplastic cells, bacterial colonies, or fungal organisms identified. Amylase and pancreas-specific lipase (Precision PSL^e) of peripheral blood serum were within normal limits at 507 IU/L (290–1125 IU/L) and 53 U/L (4–140 U/L), respectively. Amylase and Precision PSL of the gastric cyst fluid were elevated at 1435 IU/L (290–1125 IU/L) and 8214 U/L (4–140 U/L), respectively. Cytological analysis of the gastric cyst fluid revealed viscous proteinaceous fluid with mild to moderate neutrophilic inflammation and cell degeneration with no evidence of neoplasia or infectious agents, consistent with inflamed cystic material. Amylase and lipase levels in the gastric cyst fluid were significantly higher than the serum levels, which confirmed the cyst to be a pancreatic pseudocyst. Aerobic and anaerobic cultures of the fluid were negative for bacterial growth. The dog was hospitalized for monitoring and supportive care for pancreatitis and was discharged 8 days after surgery once it was confirmed that both oral and J-tube feedings were well tolerated. The owner was instructed to offer oral food every 4 to 6 hr. If the dog did not eat, it was recommended to provide enteral nutrition via the J-tube with CliniCare^f liquid diet. A recheck abdominal ultrasound was performed 7 days following the second exploratory laparotomy for omentalization of the intramural gastric pseudocyst. The area of the stomach wall with the previous intramural gastric pseudocyst appeared to have a small amount of residual anechoic to hypoechoic material but improved pancreatitis and secondary steatitis/peritonitis.

The dog was presented again for anorexia and diarrhea 14 days following intramural gastric pseudocyst drainage and omentalization. The dog was subsequently hospitalized for medical management of presumed recurrence of pancreatitis. A recheck abdominal ultrasound revealed decreased size of the previous gastric pseudocyst (1.5 cm) and improved pancreatitis; however, there was evidence of an intramural cystic lesion in the mid to distal duodenum, orad to the duodenal flexure and J-tube site, with associated intraluminal compression and obstruction (Figure 2A).

View Figure

• Download as Powerpoint
• Download Figure

Abdominal exploratory surgery was performed the following day. An esophagostomy tube was placed preoperatively for enteral nutrition support. The abdomen was explored. The adhesions between loops of intestine were easily broken down via gentle digital dissection. Five centimeters orad to the duodenal flexure and jejunostomy site, the proximal and mid-duodenum were adhered to the right body wall (Figure 2B). There was a large (5.8 cm × 3.5 cm), intramural cystic lesion located at the mid to distal duodenum, similar to the gastric pseudocyst lesion. Communication with the duodenal lumen or the jejunostomy site was not noted in surgery. The cystic fluid aspirated was similar to the fluid previously aspirated from the gastric pseudocyst and was dark brown–green and gelatinous. A sample of the fluid was collected for aerobic and anaerobic culture and sensitivity, fluid analysis and cytology, and amylase and lipase levels to be compared with concurrently drawn intraoperative blood serum samples. Intraoperative endoscopy was performed to confirm the intramural location of the duodenal cystic lesion, and no communication of the cystic lesion with the duodenal mucosa was noted (Figure 2C). The previous jejunostomy site was broken down, and the J-tube was removed.

Approximately 12 cm of intestine was resected, including the duodenum 1 cm orad to the origin of the pseudocyst, the duodenal flexure, and the jejunostomy site at the proximal jejunum (Figure 2D). A duodenojejunal anastomosis was performed with 4-0 polydioxanone in a simple interrupted suture pattern. The previous surgery site of the omentalized gastric pseudocyst was examined, and no grossly palpable fluid-containing structure or evidence of gastric outflow obstruction was identified.

Results of the paired enzyme analysis for the duodenal cyst fluid were similar to that of the gastric pseudocyst fluid. Amylase and PrecisionPSL of peripheral blood serum were within normal limits at 526 IU/L (290–1125 IU/L) and 49 IU/L (4–140 U/L), respectively. Amylase and PrecisionPSL of the duodenal cyst fluid were elevated at 1380 IU/L (290–1125 IU/L) and 7834 U/L (4–140 U/L), respectively. Cytological analysis of the duodenal cyst fluid revealed viscous proteinaceous fluid with moderate neutrophilic inflammation and cell degeneration with no evidence of neoplasia or infectious agents, consistent with inflamed cystic material. Aerobic and anaerobic cultures of the fluid were negative for bacterial growth. Histopathology of the duodenum was consistent with regional fibrinosuppurative peritonitis with dissection into the submucosa-muscularis junction, confirming pseudocyst formation containing fibrinosuppurative contents.

The dog recovered well and was discharged 2 days after the operation. Recheck abdominal ultrasound 3 wk later revealed further regression of the omentalized intramural gastric pseudocyst (0.33 cm diameter), and no recurrent intramural cystic lesions were noted throughout the GI tract (GIT) or other abdominal viscera. Long-term follow-up with recheck examinations revealed no recurrence of clinical signs, and abdominal ultrasounds performed postoperatively at 1 mo, 4 mo, 7 mo, and 19 mo showed no recurrence of gastric or duodenal pseudocyst formation or insulinoma metastases (Figure 3). The dog has remained euglycemic with no recurrence of clinical signs or evidence of insulinoma metastasis at the time of writing this report (22 mo following the duodenojejunal resection and anastomosis surgery) based on serial bloodwork, three-view thoracic radiographs, and abdominal ultrasounds.

View Figure

• Download as Powerpoint
• Download Figure

Discussion

Pancreatic pseudocysts are a rare clinical entity in humans and are even less common in dogs and cats.^15–23 A pancreatic pseudocyst is a completely encapsulated collection of sterile fluid containing pancreatic enzymes that accumulates within a well-defined wall of fibrous tissue with no epithelial lining.^16–19 This defining feature separates a pseudocyst from a true cyst that contains an epithelial lining. The fibrous inflammatory wall prevents leakage of the cystic contents and is formed from inflammatory changes in mesenteric, peritoneal, and serous linings.^18,23–25 Pancreatic pseudocysts, despite being a rare entity, are the most common cystic lesions found in the pancreas. In humans, this lesion accounts for 80–90% of cystic structures of the pancreas, and most cases of true cysts are neoplastic.^18,26 In both human and veterinary patients, pancreatic pseudocysts most commonly occur as a consequence of chronic pancreatitis.^{12,22,23,26,27} Although pancreatic pseudocyst incidence has been reported as 20–40% in human patients with chronic pancreatitis, there is also a 5–20% incidence in patients with acute pancreatitis.^16,27–31

The pathogenesis of pancreatitis is not well known; however, it is generally accepted to occur after activation of trypsinogen to trypsin within the acinar cells of the pancreas.^12,32–34 Trypsin then prematurely activates pancreatic proteases and phospholipase A₂, which then leaks into the interstitium causing edema and inflammation.^32–34 In cases of severe pancreatitis, the leakage of these pancreatic enzymes, in addition to recruitment of destructive inflammatory mediators, results in damage to vascular endothelium, impaired pancreatic microcirculation, and subsequent edema and hypoxia, resulting in further tissue destruction in pancreatic and peripancreatic tissues including adjacent organs.³⁴

The development of a pancreatic pseudocyst has been suggested to arise from damage to the main pancreatic duct or its branches.¹² This disruption leads to leakage of pancreatic secretions and formation of localized fluid accumulation into a well-circumscribed pseudocyst without evidence of parenchymal necrosis.¹² Development of pseudocysts within the pancreas have been reported in dogs and cats secondary to pancreatitis and pancreatic trauma.^22,35 In humans, common underlying etiologies include alcoholism, trauma, and biliary disease.^18,23 There have also been several human reports documenting pancreatic pseudocysts following pancreatic surgery due to leakage from the ducts or ductal branch disruption, much like the dog in the current case report.^3,36–38

Although there are case reports of pseudocysts forming in the pancreas itself in dogs and cats, to our knowledge, there are no reports of pseudocysts of pancreatic origin forming in other organs in veterinary patients.^{24,25,35,39–43} The first reports describing development of pseudocysts of pancreatic origin in the wall of the GIT in humans were over 53 yr ago.^5,44 Since then, there have been numerous case reports and large case series describing these intramural pseudocysts in the stomach, duodenum, and colon.^{4–6,8–10,44–49} The exact mechanism of pseudocyst formation in the GIT is unknown; however, previous reports suggest that they form as a sequela to rupture of a pancreatic pseudocyst into the GIT wall, presence of a fistula between the pancreas and GIT, and inflammation of heterotopic pancreatic tissue within the GIT wall (i.e., a congenital anomaly in which pancreatic tissue is anatomically separate from the main gland).^{4–6,8–10,44–49} There have also been rare cases of these pseudocysts of pancreatic origin occurring in atypical locations such as the spleen, liver, mediastinum (pancreaticopleural fistulas), pelvic canal, and kidney.^6,50,51 These reports suggested that the organ affected was related to the path taken by the activated pancreatic enzymes.^6,50,51 Although it is unknown exactly why the dog in this report developed intramural pseudocysts, rupture of a pancreatic duct branch from surgery or severe postoperative pancreatitis following a partial pancreatectomy with enzymatic leakage near the stomach and duodenum could have led to formation.

The primary differentials for patients with large cystic lesions within the pancreas or peripancreatic organs and clinical signs of pancreatitis include pseudocysts, abscesses, and neoplasia with cystic component (i.e., necrotic center).^{19–21,24,52,53} In general, the most reliable diagnostic test reported for a pancreatic pseudocyst in any organ is comparison of pancreatic enzyme activity in the cystic fluid with peripheral serum or plasma.²⁵ Fluid from pseudocysts will typically have higher amylase and lipase levels compared with serum levels, as was seen with the dog in the current case report for both the gastric and duodenal pseudocysts.²⁶ Human studies evaluating fluid enzyme levels of various cystic lesions showed that pseudocyst fluid amylase levels were higher than neoplastic cysts, the majority of which were >5,000 IU/mL, whereas neoplastic cysts were typically <83 IU/mL.^26,54,55

The diagnosis of a pancreatic pseudocyst in humans is made based on a combination of clinical signs of pancreatitis and imaging techniques, including ultrasonography, CT, and endoscopic retrograde pancreatography (ERP).^18,24,55 ERP is an invasive imaging modality that provides radiological visualization of the pancreatic ducts via injection of contrast agent into the main pancreatic duct to aid in the diagnosis of pancreaticobiliary diseases. However, the use of ERP has declined because of the invasiveness of the procedure, risk of complications such as post-ERP pancreatitis, and the widespread availability of noninvasive cross-sectional imaging techniques (CT, MRI, and endoscopic ultrasound scan [EUS]).⁵⁵ EUS has been reported to have higher sensitivities for diagnosing both solid and cystic pancreatic lesions compared with cross-section imaging with MRI or CT.⁵⁵ Because of the rarity of pseudocysts forming in other organs in humans, and no reports in the veterinary literature, even advanced imaging modalities such as CT, MRI, and/or EUS may not be able to differentiate a pseudocyst from a cyst or abscess.⁵⁵ Thus, definitive diagnosis is often achieved with paired serum and fluid amylase and lipase levels, and therapy is achieved with drainage procedures via endoscopic, percutaneous, or surgical means.²⁶ In the current case report, abdominal ultrasound identified a cystic lesion causing compression of the dorsal wall of the stomach and also an intramural cystic lesion in the duodenal wall; however, it was not possible to differentiate pseudocyst from a cyst or abscess.

Several human studies suggest that when there is high suspicion for an intramural gastric pseudocyst, and the patient is asymptomatic, it is preferable to avoid therapeutic intervention.^19,56 A human prospective cohort study showed that >50% of asymptomatic patients with pancreatic pseudocysts had resolution within 6 mo without any intervention.⁵⁷ Nevertheless, if signs relating to gastric outflow obstruction or severe pancreatitis develop, intervention is warranted, because untreated pseudocysts can lead to numerous complications including hemorrhage, infection, biliary complications (compression of the common bile duct), portal hypertension (compression or obstruction of the splenic vein/portal vein), and rupture, which can be fatal.^{9,11,15–17,51} The clinical decision for surgical intervention in the current case was primarily based on the acute worsening of clinical signs associated with gastric outflow obstruction. Because the authors were familiar with the human literature pertaining to the diagnosis of pancreatic pseudocysts with paired amylase and lipase samples before the surgery, we were able to make a diagnosis of a pancreatic origin pseudocyst in this dog with intraoperative fluid collection.

Treatment options reported for pancreatic pseudocysts in the pancreas of dogs include percutaneous drainage, surgical omentalization, surgical debridement, cystogastrostomy, and EUS-guided ethanol ablation.^25,39–43 Because of the rarity of GI intramural pseudocysts, there are no specific guidelines on treatment. A recent comprehensive metanalysis (1986–2015) in humans evaluating outcomes following treatment of pancreatic pseudocysts with surgery, endoscopic drainage, and percutaneous ultrasound-guided drainage revealed local recurrence rates between 1 and 22% with surgery, with a maximum time to local recurrence being 4.5 yr.⁵⁸ In the same study, endoscopic drainage had a recurrence rate between 0 and 40% with a maximum time to local recurrence of 2.6 yr. Percutaneous ultrasound-guided drainage had recurrence rates between 0 and 38% with a maximum time to local recurrence of 3 yr. Ten studies, with a total of 811 patients undergoing surgical intervention, had an overall success rate of 83% with <2% mortality. With percutaneous drainage, a total of 11 studies, with 401 patients, had an overall success rate of 67%, with some studies reporting 58% of patients requiring adjuvant surgical drainage. Thirty-four studies evaluating endoscopic cystogastrostomy or cystoduodenostomy drainage, with a total of 1463 patients, showed an 85% clinical success without the need for adjuvant surgical intervention.

In the case reported here, ultrasound-guided percutaneous drainage was not feasible given the dorsal location of the gastric lesion. Drainage with omentalization was elected as the entire dorsal gastric wall could not be resected. In addition, endoscopic cystogastrostomy drainage has not been previously reported in the veterinary literature and was thus not elected. Resection and anastomosis of the duodenal pseudocyst was elected based on location and ability to resect. However, omentalization could have been a reasonable surgical option as well. Twenty-two months following insulinoma removal, the dog was reported to be doing well with no evidence of local recurrence or metastasis.

Conclusion

Based on the findings of this report and the literature describing pseudocyst formation in the pancreas and peripancreatic organs in humans, it is important that clinicians consider this potential complication following pancreatic surgery or severe pancreatitis in dogs. The diagnosis of pancreatic pseudocyst can be confirmed via paired serum and fluid amylase and lipase levels. Based on human literature, asymptomatic dogs may not require treatment. However, in symptomatic patients, drainage should be considered depending on location of the pseudocyst. Therapeutic options include image-guided percutaneous drainage, endoscopic drainage, or surgery to either resect or omentalize the lesion. Long-term prognosis for humans following treatment is good, as was the outcome of the dog in this report.