Prolonged Survival in an Aged Labrador Retriever with a Metastatic Insulinoma

Introduction

Insulinomas, β cell tumors of the pancreas, are uncommon in dogs but are almost always malignant, with metastatic disease frequently identified at the time of diagnosis.¹ Insulinomas occur more commonly in middle-aged and geriatric dogs, with large-breed dogs being overrepresented. The most commonly reported breed affected by pancreatic insulinomas is the Labrador retriever.^2,3 Clinical signs, resulting from hypoglycemia, are particularly debilitating for dogs either prior to diagnosis or following relapse of surgical and/or medical therapy. Many patients exhibit weakness, collapse, disorientation, seizures, tremors, and polyphagia.^1,4–6 In one report, dogs treated exclusively with medical therapy following a diagnosis of insulinoma had a median survival time of 74 days, likely influenced by uncontrolled hypoglycemia.¹ In a study of dogs that underwent partial pancreatectomy for insulinoma treatment, dogs without surgically detected metastatic disease had a median normoglycemic disease-free interval of 14 mo, whereas the median normoglycemic disease-free interval was 1 mo in dogs with metastatic disease.⁴ Polton et al. (2007) reported a higher median survival time of 547 days and 217 days for dogs having presumed lymphatic and hepatic metastases, respectively; however, determination of advanced disease was overestimated by lack of histopathologic confirmation.⁶

Although mass resection with partial pancreatectomy is an effective option for dogs with insulinomas, there are complications that can arise from removing a portion of the pancreas. Previously reported postoperative complications after insulinoma resection include persistent hypoglycemia, diabetes mellitus, pancreatitis, and sepsis.^1,4,6

This report highlights an unusually prolonged, asymptomatic survival time in a dog following a partial pancreatectomy for a functionally active pancreatic insulinoma with histologically confirmed functional hepatic metastasis that also experienced an unusual postoperative complication of peritonitis and exocrine pancreatic insufficiency (EPI).

Case Report

A 12 yr old male Labrador retriever weighing 37.5 kg presented to the Colorado State University Veterinary Teaching Hospital for treatment of a suspected insulinoma. Pertinent history included multiple episodes of weakness and collapse over the preceding 2 wk. The dog also experienced several episodes of coughing and gagging. He had been evaluated by the referring veterinarian, who documented inappropriately elevated insulin (72.7 μg/mL; reference range, 5–20 μg/mL) in the face of hypoglycemia (29 mg/dL; reference range, 70–138 mg/dL). An abdominal ultrasound performed by the Gentle Doctor Animal Hospital’s Mobile Veterinary Diagnostics did not reveal any abnormalities. Specifically, no pancreatic mass was visualized, the liver appeared homogeneous, and all other structures appeared within normal limits.

At the time of referral, the patient was being managed medically. He was fed multiple small meals (q 4–6 hr) that were supplemented with glucose syrup, 20 mg prednisone per os (PO) q 24 hr, and was supplemented with calcium ascorbate and glucosamine for previously diagnosed degenerative joint disease. The patient was scheduled for an exploratory celiotomy for the removal of a presumptive insulin-secreting tumor.

Physical examination revealed an ulcerative hyperplastic lesion next to the nail bed on the second digit of the left forepaw and increased upper respiratory sounds ausculted over the trachea. Preanesthetic biochemical analysis was within normal limits except for hypoglycemia (47 mg/dL; reference range, 70–143 mg/dL). Complete blood cell count was unremarkable. The patient’s blood type was dog erythrocyte antigen 1.1 positive.

The patient was premedicated for general anesthesia with intramuscular acepromazine (0.008 mg/kg), atropine (0.02 mg/kg), and hydromorphone (0.05 mg/kg) approximately 60 min prior to induction. General anesthesia was induced with IV propofol (2.67 mg/kg) that was administered with a bolus of a crystalloid solution. IV doxapram (20 mg) was administered at the time of induction to aid in an examination of laryngeal movement. Laryngeal paralysis was not identified as the source of increased upper respiratory sounds.

Once the patient was intubated, he was maintained on 1–2% isoflurane. Blood glucose was measured following induction and was within normal limits (95 mg/dL; reference range, 75–130 mg/dL). Throughout general anesthesia, the patient received IV lactated Ringer’s solution with 2.5% dextrose. Blood glucose measured 15 min after induction was 95 mg/dL and was measured 2 hr after induction and was elevated (167 mg/dL). The patient continued to be monitored. Supplemental dextrose was discontinued 1 hr later once the patient was recovered from anesthesia. A fentanyl constant rate infusion (2.5–10 μg/kg/hr) was administered throughout anesthesia and adjusted pro re nata to manage intraoperative pain and reduce anesthetic gas requirement. Intermittent positive pressure ventilation was used to maintain respiration and oxygenation throughout surgery. The patient received two doses of perioperative IV cefazolin (22 mg/kg). One was administered 35 min prior to opening the abdomen and the second dose was administered 90 min later.

Exploratory celiotomy revealed a lobulated mass in the distal right limb of the pancreas measuring 2 cm in diameter. The mass extended ventrally into the mesentery. The mass was intimately involved with the duodenal blood supply and care was taken to leave the anastomosing region of the cranial and caudal pancreaticoduodenal artery intact. The pancreatic duct and minor duodenal papilla could not be visualized intraoperatively. There was moderate blood loss during the procedure, and 3-0 polyglyconate, regular hemoclips, and monopolar electrocautery were used for hemostasis. A partial pancreatectomy was completed via an encircling ligature of 3-0 polyglyconate on the right limb of the pancreas proximal to the tumor. Approximately one-third of the cranial right limb remained with the undisturbed pancreatic body and left limb, which appeared grossly normal. An enlarged pancreaticohepatic lymph node was noted and was extirpated for histopathology. The liver contained many diffuse, discrete, tan-colored masses ranging in size from 2 mm to 10 mm in diameter. Some lesions were raised and others were flush with the surface of the liver. Liver biopsies were taken from two representative areas of the liver and were submitted for histopathology. The skin lesion on the second digit of the left forepaw was also removed.

Histopathology of the pancreatic mass demonstrated a pseudoencapsulated cellular mass surrounded by exocrine pancreas. The mass was characterized by nests and packets of polygonal cells, with variable distinct cell borders and moderate amounts of eosinophilic cytoplasm, separated by fibrovascular septae. The cell nuclei were oval, indented, and contained finely stippled chromatin. There was prominent invasion of the fibrous pseudocapsule, and rare vascular invasion identified. The liver was expanded by multiple variably sized cellular masses with the same tissue and cytomorphologic features as the pancreatic mass. The histologic features were consistent with islet cell carcinoma and hepatic metastasis. There was no evidence of metastatic disease in the pancreaticohepatic lymph node; however, the node was expanded by accumulations of edema and red blood cells. The digital mass was diagnosed as focal adnexal hyperplasia.

The dog’s immediate postoperative recovery was uneventful. Dextrose was discontinued immediately postoperatively, and the patient was maintained on IV fluids with 20 mEq of potassium chloride/L. The fentanyl infusion was continued postoperatively for pain management. Multiple small meals of canned maintenance diet were offered, beginning 4 hr postoperatively. Serial blood glucose measurements in the postoperative period ranged from 132 mg/dL to 187 mg/dL.

Twenty-four hr postoperatively, the patient began vomiting, regurgitating, and developed inappetence in conjunction with moderate to severe abdominal pain that continued for 6 days. An abdominal ultrasound examination performed 5 days postsurgically revealed hyperechogenicity, consistent with inflammation, in the mesentery at the surgical site and small volumes of peritoneal fluid. A 2 mL sample of abdominal fluid was collected via abdominocentesis. The fluid was serosanguinous, turbid, and had a nucleated cell count of 37,000 cells/μL. There were 70,000 red blood cells/μL and the total protein was 3.9 g/dL. The fluid contained predominately nondegenerate neutrophils (69%) and monocytes (29%). Lymphocytes were also present (2%). In addition, abdominal fluid lipase was 6,150 IU/L. A presumptive diagnosis of pancreatitis and nonseptic suppurative peritonitis was made based on clinical signs, ultrasound findings, and abdominocentesis. Treatment included continued IV fluid therapy, analgesia (including fentanyl, carprofen, tramadol, and hydromorphone), antiemetics and gastrointestinal protectants (including maropitant, ondansetron, pantoprazole), a synthetic prostaglandin (misoprostol), and promotility drugs (including metoclopramide and lidocaine).^7,8

Seven days postsurgically, the patient was discharged from the hospital and was prescribed metoclopramide (0.3 mg/kg PO q 8 hr q 10 d). The owners were instructed to have the dog’s blood glucose monitored q 1 mo for the first 6 mo then q 3 mo thereafter. The owners were advised to watch carefully for clinical signs of hypoglycemia. Prednisone was not reinstituted, and the dog was fed multiple small meals throughout the day. A guarded prognosis was anticipated because of the multifocal hepatic metastatic disease.

Approximately 3 mo postoperatively, weight loss, a voracious appetite, and voluminous stools were noted by the owner. Blood work performed by the referring veterinarian was within normal limits, including serum glucose. Trypsin-like immunoreactivity was low (1.9 μg/L; reference range, 5–35 μg/L), supporting a diagnosis of EPI. Parenteral pancreatic enzymes were administered, which resolved the clinical signs of EPI. Because of the surprisingly normal serum blood glucose measurements, the original surgical samples were resubmitted for further testing to confirm the previously diagnosed hepatic metastatic disease with functional endocrine activity.

Immunohistochemistry was performed on the pancreatic mass and liver biopsies using antibodies against chromagranin A^a, synaptophysin^b, and insulin^c to confirm the β cell histogenesis. Normal canine pancreas and omission of the primary antibody on the patient’s pancreatic mass and liver biopsy tissues were used as positive and negative controls, respectively. The pancreatic and liver masses demonstrated similar patterns of cellular staining, with the majority of neoplastic cells staining positive for all three antigens evaluated (Figure 1).

View Figure

• Download as Powerpoint
• Download Figure

Ten mo postoperatively, the dog was examined at the Colorado State University Veterinary Teaching Hospital. Thoracic radiographs, abdominal ultrasound, serum biochemical analysis, and a complete blood cell count were performed. There were no abnormalities noted on either the serial thoracic radiographs or abdominal ultrasound. Blood work revealed a serum insulin concentration just above the normal range (20.2 μIU/mL; reference range, 5–20 μIU/mL) and normal serum glucose (102 mg/dL). The remainder of the blood work was within normal limits.

Twenty-three mo postoperatively, the dog exhibited an episode of hypoglycemia (37 mg/dL), with clinical signs of weakness, disorientation, and lethargy. The serum blood glucose had been consistently monitored q 2 mo postoperatively and had always been within normal limits. The dog had not shown any signs of hypoglycemia. According to the owner, the dog had been acting like his younger self since they began administering the pancreatic enzyme supplements. The hypoglycemia was well controlled with the administration of prednisone at a dose of 10 mg PO q 12 hr. Four mo later (27 mo postoperatively), the dog experienced intermittent hyporexia, and 6 mo later (29 mo postoperatively), the dog became progressively dyspneic and acutely died approximately 48 hr later. A limited biochemical panel performed at the start of the dyspneic episode revealed elevated serum creatinine (2.3 mg/dL; reference range, 0.5–1.8 mg/dL), elevated alanine aminotransferase (210 IU/L; reference range, 10–100 IU/L), and elevated alkaline phosphatase (1,540 IU/L; reference range, 23–212 IU/L).

At necropsy, the presence of metastatic tumors within the liver, a mesenteric lymph node, and peripancreatic scar tissue were histologically confirmed. Subjectively, the metastatic liver lesions were assessed as comparable to the pathology observed intraoperatively. Additionally, hepatocellular degeneration, multifocal chronic bronchopneumonia, and membranous glomerulopathy were diagnosed postmortem.

Discussion

Insulinoma more commonly occurs in middle-aged or older patients and large-breed dogs. Labrador retrievers are overrepresented.^2,3 The signalment of the patient reported here is consistent with the classic signalment and presenting clinical signs of patients with an insulinoma. Like many patients, there was already evidence of metastatic disease at the time of surgery, which was only 2 wk after the initial clinical signs were first noted. Insulinoma with presumed functionally active hepatic metastasis was seen at the time of surgical exploration, and the histogenesis of the metastatic sites was confirmed with histopathologic analysis. Immunohistochemical staining for chromogranin A, synaptophysin, and insulin demonstrated the presence of metastatic disease within the liver as the cells in the liver biopsies had identical islet cell histogenesis and insulin expression as the functional primary tumor.^9,10

In humans diagnosed with an insulinoma, a standard of clinical practice includes radioimmunoassay to detect serum polypeptides such as chromogranin to ascertain islet cell origin.¹¹ Similarly, islet cell origin was demonstrated in this dog as immunohistochemical staining of chromogranin A accurately localizes the secretory granules of endocrine tumors.¹² Insulin expression was also confirmed in both the pancreatic and hepatic lesions via immunohistochemistry. The primary pancreatic tumor was functionally active; therefore, it was presumed the similarly staining metastatic hepatic lesions were also active, although immunohistochemistry cannot assess any functional potential. The hepatic metastases could have been functionally active, but exerting only subclinical effects following the reduced burden of neoplastic insulin-secreting tissue after the partial pancreatectomy. Human and canine clinical practice does include both management of symptoms resulting from abnormal production of functional hormone and the symptoms associated with tumor bulk.^3,11 The circulating insulin level measured 10 mo postoperatively was technically elevated, with concurrent normoglycemia. What was not known was how much neoplastic insulin-secreting tissue either releases functional insulin or contributes to systemic hypoglycemia.

Despite the presence of presumed hormonally active metastatic disease, this patient enjoyed a longer than predicted disease remission and survival. One possible explanation for this patient’s prolonged survival was believed to result from the significant pancreatitis and peritonitis that developed postoperatively. Reports of improved survival in human patients experiencing postsurgical infections, such as improved patient survival following empyema development after surgical resection of respiratory tumors and improved survival in children experiencing infections of their limb-sparing surgery for osteosarcoma.^13,14 More recently, Lascelles et al. (2005) were able to show improved survival time in canine osteosarcoma patients that underwent limb-sparing surgery and developed postoperative infection.¹⁵ A mechanism for those patients experiencing prolonged survival after a bacterial infection results from nonspecific tumor inhibition dependent upon macrophages and natural killer cells of the innate immune system.¹⁶ There is currently no published data the authors are aware of establishing survival benefit of postoperative infection or peritonitis in dogs following celiotomy for the treatment of abdominal neoplasia. The patient reported here, although he did not develop septic peritonitis, had massive immunostimulation postoperatively as a result of suppurative peritonitis. It is possible that this stimulation of humoral immune reaction taking place within the peritoneal cavity was enough to affect the metastatic lesions in the liver. It is additionally possible that the cell-mediated immune reaction that was amplified as a result of the postoperative peritonitis, which also affected the metastatic disease. Alternatively, it is possible that some combination of the medications used in treating the postoperative peritonitis had an antitumor effect; however, this hypothesis is less likely as those medications are commonly used in the care of many cancer patients without evidence of tumor regression.

No cytotoxic chemotherapeutic drugs or other antineoplastic management strategies were used in this patient because of the associated risks and the lack of documented conclusive benefit.^5,17,18 Thus, another possible explanation for this dog’s prolonged disease-free interval is spontaneous partial regression of the metastatic lesions, possibly as a result of removing the large primary tumor. There are numerous studies that have documented improved survival times with removal of the primary insulinoma tumor, even in dogs with metastatic disease, but the return of hypoglycemia is expected and usually managed medically when clinical signs of hypoglycemia become evident. ^1–3,6 Complete spontaneous regression of both localized and metastatic neoplasia has been documented in human pancreatic tumors.^19,20 Although exceedingly rare, spontaneous regression of cancer has been quoted to occur in 1 out of 100,000 cases of cancer, but the actual incidence is unknown.¹⁹ Either ultrasound-guided liver biopsies or second-look celiotomy/laparoscopy would have been necessary to confirm spontaneous partial regression in this patient antemortem.

It is also equally plausible the metastatic lesions were curiously nonprogressive or very slowly progressive for 23 mo. The metastatic hepatic lesions were not detected on the first abdominal ultrasound examination and could have similarly been undetectable during the restaging that was performed 10 mo postoperatively. The dog described in this report experienced a subjectively unexpectedly prolonged survival. In a case series by Tobin et al. (1999), a median survival time of 381 days was reported for 26 dogs that underwent a partial pancreatectomy for insulinoma treatment, with 10 of those dogs also having hepatic metastasis.¹ A median survival time of 785 days was reported by Polton et al. (2007) for 19 dogs having a partial pancreatectomy for insulinona treatment.⁶ In that study, five dogs that underwent surgery were excluded from analysis based on impressions of inoperable disease, not all dogs had histopathologic staging of their disease, the primary tumor was not found in three dogs, and only one of the four dogs with presumed hepatic metastasis was biopsied. As such, advanced disease was overestimated, resulting in a median survival time greater than reported by Tobin et al. (1999).⁶ Caywood et al. (1988) reported an approximate median survival time of 6 mo for dogs with histopathologically confirmed hepatic metastasis, a cohort of dogs most representative of the patient described herein.⁴

In the current case, because of the confirmed presence of nonresectable metastatic disease in the dog’s liver, the postsurgical prognosis was guarded. The patient also suffered from remarkable postoperative complications that were life threatening. In many cases, the postoperative complications along with the guarded prognosis for return of clinical signs and survival would have influenced many dog owners to consider euthanasia. Instead, this patient enjoyed a longer survival and disease-free interval than predicted for a dog with metastatic insulinoma at sites distant to the pancreas.

Conclusion

This case reiterates the importance of exploratory celiotomy and removal of the primary pancreatic lesion in dogs with suspected insulinomas. This case may also be an impetus for future studies evaluating the effect of postoperative complications and peritonitis in patients with intra-abdominal neoplasia.