Diabetes Mellitus Following Resection of a Hepatocellular Carcinoma in a Dog Presented with Hypoglycemia
ABSTRACT
A 5 yr old male neutered Labradoodle presented for an episode of acute collapse. Point-of-care blood work showed hypoglycemia and abdominal ultrasonography revealed a liver mass arising from the caudate liver lobe. The dog underwent a partial liver lobectomy, and histopathology confirmed a fully resected hepatocellular carcinoma. Blood glucose levels normalized initially after surgery, but 1 wk later, the patient was diagnosed with diabetes mellitus based on the development of polyuria, polydipsia, hyperglycemia, and glucosuria. Appropriate treatment with insulin was initiated, and 1 yr following the diagnosis, the dog was still requiring administration of insulin twice daily. This case describes the uncommon development of diabetes mellitus in a dog following surgical resection of a hepatocellular carcinoma initially associated with hypoglycemia. Although very unusual, this should be considered as a potential complication of surgical treatment of such tumors, and affected patients may require long-term medical management.
Introduction
Hepatobiliary neoplasms are considered uncommon in dogs, representing only 0.6–1.3% of all canine neoplasms.1 Hepatocellular carcinoma (HCC), a malignant epithelial tumor primarily arising from hepatocytes, is the most frequent liver neoplasm diagnosed in dogs.2 Morphologically, such tumors may be classified as being massive, nodular, or diffuse. The massive form, defined as a large, solitary mass confined to a single liver lobe, is found in up to 53–83% of the cases.1–3 The biological behavior of the massive form of HCC is relatively nonaggressive, and therefore, the prognosis is generally good if complete surgical resection is possible.4
Laboratory changes in primary liver tumors are usually nonspecific and include leukocytosis, anemia, thrombocytosis, and raised levels of both alkaline phosphatase and alanine aminotransferase (ALT).5,6 Hypoglycemia has been reported as paraneoplastic syndrome in nonpancreatic tumors such as HCC, leiomyosarcoma, leiomyoma, mammal gland adenocarcinoma, and pulmonary adenocarcinoma.7–11 This type of hypoglycemia is termed non–islet cell tumor hypoglycemia (NICTH) and, initially, was thought to be associated with increased glucose utilization by neoplastic tissue and severe parenchymal destruction.12 More recently, it has been suggested that NICTH may be related to insulin-like growth factor 2 (IGF-2) overexpression and elevated serum IGF-2 levels.7,10 The mainstay of treatment of NICTH is tumor resection.9,11,13
This case report documents the development of diabetes mellitus (DM) 1 wk after surgical resection of a hepatocellular carcinoma in a dog that was initially presented with NICTH. To the authors’ knowledge, this has only been reported twice in the veterinary literature and the underlying pathophysiology remains unknown.8,9
Case Report
A 5 yr old male neutered Labradoodle presented with several weeks’ history of progressive lethargy and generalized weakness leading to an episode of acute collapse. No previous medical concerns were reported, and the patient had not received any recent or long-term medication. On presentation, the patient was depressed, and hypersalivation along with moderate discomfort on cranial abdominal palpation were the only abnormalities detected on physical examination. Neurologically, proprioceptive ataxia was noted. Body condition score was normal. Point-of-care blood work and ultrasonography revealed moderate hypoglycemia (50 mg/dL, reference range 65–110 mg/dL) and hepatic changes, respectively. Initial stabilization consisted of a 5% dextrose (Glucose 50%; Hameln) constant rate infusion (CRI) IV at 4 mL/kg/hr, along with maropitanta at 1 mg/kg IV q 24 hr and methadoneb at 0.2 mg/kg IV q 4 hr.
Diagnostic Investigation
In-house hematology and biochemistry revealed severe hypoglycemia (1.51 mmol/L, reference range 4.11–7.95), elevation of ALT activity over the readable range of the machine (>1000 U/L, reference range 10–125 U/L), elevation of alkaline phosphatase activity over the readable range of the machine (>4000 U/L, reference range 23–212), and elevation of γ-glutamyl transferase activity (22 U/L, reference range 0–11). Prothrombin time and activated partial thromboplastin time were within normal limits. Abdominal ultrasonography revealed a large (11 cm), pedunculated liver mass arising from the caudate lobe of the liver and a small volume of peritoneal effusion, which was not accessible for sampling. There were no other abnormalities identified on abdominal ultrasound. An abdominal computed tomography scan and thoracic imaging were offered at that point but were declined because of financial constraints.
Therapy
The patient was blood typed DEA 1.1 positive before surgery. The dog was premedicated with methadoneb 0.3 mg/kg IV, combined with medetomidine hydrochloridec 0.003 mg/kg IV. General anesthesia was induced with propofold given IV to effect, and anesthesia was maintained using isofluranee delivered in 100% oxygen. A 5% dextrose CRI was maintained during the procedure. A midline celiotomy approach was performed from the xiphoid process to the parapreputial area. Exploration of the abdomen revealed a mass originating from the papillary process of the caudate liver lobe. A partial liver lobectomy was performed using a 60 mm thoracoabdominal surgical stapler.f The rest of the exploratory celiotomy was grossly normal. The liver mass was submitted for histopathology along with the adjacent hepatic lymph node, which was removed for staging purposes.
Postoperatively, IV fluid therapy was maintained with Hartmann’s solutiong at 2 mL/kg/hr. Analgesia was initially provided with a ketamine CRIh at 2.5 mcg/kg/min and methadoneb at 0.2 mg/kg q 4 hr. Blood glucose levels normalized (9.5 mmol/L, reference range 3.5–6.1) immediately postoperatively and glucose supplementation was stopped. Blood glucose levels were 7–8 mmol/L the day after surgery. The dog was discharged 48 hr after surgery with a 10-day course of acetaminophen/codeinei 10 mg/kg per os q 8 hr.
Outcome
Histopathological examination of the liver mass revealed neoplastic hepatocytes predominantly arranged in a trabecular pattern whereby trabeculae were often up to 8–20 hepatocytes thick. The neoplastic hepatocytes were relatively well differentiated with moderate amounts of pale eosinophilic cytoplasm and a single round vesicular nuclei with 1–2 prominent nucleoli. Anisokaryosis was moderate. In some areas, there were multinucleated hepatocytes and scattered karyomegaly. Mitoses per 10 high-power field were equal to two. These findings were consistent with an HCC, which was completely excised. The architecture of the adjacent lymph node was intact and consistent with chronic blood drainage, without evidence for regional metastasis.
One week after discharge, the owner reported that the dog was drinking and urinating more than usual. A hematology and biochemistry in combination with urinalysis was performed at the user practice showing marked hyperglycemia (30.6 mmol/L) and glucosuria. The patient was therefore diagnosed with DM, and intermediate-acting insulinj at 0.5 U/kg q 12 hr was started.
Satisfactory control of DM was achieved 7 mo after diagnosis with 0.7 U/kg q 12 hr of intermediate-acting insulinj, and the DM remained well controlled after 1 yr of follow-up. At that time, the dog was reported to be bright and eating well.
Discussion
Among the large list of causes of hypoglycemia in adult dogs, sepsis, nonpancreatic neoplasia, insulinoma, hypoadrenocorticism, and liver dysfunction are the most common.14 According to a recent study, the prevalence of these diseases as an underlying condition causing hypoglycemia in dogs is as follows: insulinoma (69%), nonpancreatic tumor (14%), sepsis (7%), hypoadrenocorticism (6%) and hepatic failure (4%).14 In the case presented here, hypoglycemia was presumed to be associated with the liver mass identified on abdominal ultrasonography, and this was supported by resolution of the hypoglycemia after liver lobectomy. The authors acknowledge the limitation of this imaging modality for detection of other causes of hypoglycemia such as insulinoma, compared with computed tomography, and therefore, definitive exclusion of a pancreatic tumor was not possible.15 Insulin levels could have also been performed to rule out an insulinoma, although medical management of the severe hypoglycemia would have then been required for several days. Further investigations of hypoadrenocorticism were not performed, as electrolytes were within normal limits.
The median survival time for dogs with massive HCC treated surgically has been reported to be more than 1460 days when compared with dogs treated nonsurgically (270 days).4 Negative prognostic factors established in the same study included right-sided liver tumors, tumors involving either the right lateral lobe or caudate process of the caudate lobe, and increased activity of both ALT and aspartate aminotransferase.4 In this report, the fact that the mass originated from the papillary process of the caudate lobe, which extends toward the left side, crossing the midline, and is loosely covered by the lesser omentum, in combination with the mass being pedunculated made surgical resection more straightforward.16 ALT was unreadable on presentation, even though manual dilution was performed. ALT activity was not measured again postoperatively because of financial limitations. Unfortunately, aspartate aminotransferase was not part of the in-house biochemistry panel.
Ultrasound-guided fine-needle aspiration and cytologic evaluation are commonly performed to obtain a preoperative diagnosis of hepatic tumors. However, it is worth mentioning that the diagnostic accuracy of cytological analysis on FNAs of solitary hepatic masses has been reported to be low (22.9%), when compared with the final histological diagnosis.17 In the case presented here, abdominal ultrasonography did not reveal any other abnormalities apart from the liver mass, which was considered to be resectable and the most likely cause of the hypoglycemia. Therefore, fine-needle aspiration was not performed as it was considered that it would not alter the course of treatment. Extrahepatic metastasis is uncommon in dogs with HCC.2 However, further staging, with either thoracic radiographs or computed tomography—were discussed with the owner but declined because of financial constraints.
There are multiples studies supporting either partial or complete liver lobectomy as the mainstay of treatment for massive HCC, as recurrence or local metastasis are unlikely with reported occurrences of 0–13% and 0–37%, respectively.3,4,6 Surgical resection should resolve the hypoglycemia in cases with NICTH, which has been reported to improve just after mass resection in both human and veterinary medicine.9–11,13 NICTH has been found to be associated with low insulin levels and increased serum concentration and overexpression of IGF-2.7,10 In a report of a dog with HCC diagnosed histologically, an acute-onset seizure was correlated with hypoglycemia (1.9 mmol/L). Low levels of paired serum insulin (3.0 mU/mL; reference range, 10.0–25.0) and glucose (2.4 mmol/L; reference range, 4.1–5.9) ruled out an insulin-producing neoplasm before a definitive diagnosis was achieved.7 In the same patient, overexpression of IGF-2 in the tumor was also proved by immunohistology staining. In another dog with NICTH associated with a pulmonary adenocarcinoma, low levels of glucose and insulin (47 mg/dL; reference range 75–128 and <0.02 ng/mL; reference range 0.27 to 0.65, respectively) were detected along with evidence of IGF-2 in the neoplastic tissue.18 The authors acknowledge the limitation that neither paired insulin-glucose concentration nor IGF-2 expression was assessed in this case.
In human medicine, IGF-2-induced NICTH has been described in HCC and among other types of tumors.19 Interestingly, major liver resection (hemihepatectomy) due to hepatic neoplasia has been associated with moderate disturbance in glucose homeostasis and, occasionally, an early exacerbation of insulin resistance, which has been proposed as a risk factor for developing overt diabetes in the future.20 Development of prediabetic stage has been recently reported in a 60 yr old man after resection of a large solitary benign fibrous pleural tumor initially presented with hypoglycemia and associated with high serum IGF-II concentrations.21 The current hypothesis for the underlying mechanism is permanent damage to the pancreatic β cells after prolonged abnormal stimulation by IGF-II.21
Two previous cases have been described of dogs with NICTH associated with massive HCC that developed DM after liver lobectomy.8,9 In the first case, reported in 1985, the patient became diabetic 2 days following surgery. Glucose levels were successfully controlled with daily use of Neutral Protamine Hagedorn.8 In the second case, hyperglycemia was detected immediately after tumor resection (203.8 mg/dL; reference range 60–111) and polyphagia and polydipsia did not resolve with dietary management. Glucosuria was also detected on urinalysis. As a result, the dog was started on intermediate-acting insulin therapy twice daily 7 days after surgery.9 In our patient, a 5% dextrose CRI was maintained during the procedure. Blood glucose before and after mass resection were 4.5 mmol/L and 9.5 mmol/L (reference range 3.5–6-1), respectively. Glucose supplementation was stopped postoperatively, and blood glucose was closely monitored over the first 24 hr postoperatively and remained between 7 and 8 mmol/L (reference range 3.5–6.1). The owner first reported polyuria and polydipsia 7 days after surgery. The diagnosis of DM was made at the primary practice based on the presence of hyperglycemia (30.6 mmol/L) and glucosuria. One year postoperatively, the dog remained insulin dependent, which appears consistent with previously reported cases. The case described by Leifer et al. (1985) was still receiving insulin 6 wk after diagnosis and subsequently died of pyelonephritis. The case reported by Manuba et al. (2006) lived for 2 yr after surgery and was still receiving insulin therapy at that time. The pathophysiology behind the development of DM in these cases remains unknown. Development of diabetes mellitus has not been reported in dogs after resection of HCC without pre-existing hypoglycemia.
Measurements of paired serum insulin and glucose and serum IGF-II should be considered as part of the diagnostic investigations in those cases where hypoglycemia is associated with a massive solitary hepatic tumor to reach confirmation of NICTH. Close monitoring of serum glucose levels and glucosuria are highly recommended during the postoperative period. In human medicine, glucagon-like peptide-1 levels have been reported to be decreased in a case similar to the one presented here.21 Additionally, even though recent studies suggest that DM in dogs is not entirely an autoimmune disease, insulin autoantibodies, glutamic decarboxylase antibodies, and insulinoma-associated protein-2 antibodies have been described in diabetic dogs.22 Assessment of the above-mentioned parameters may therefore be considered when dysglycemia occurs following HCC resection in order to achieve a better understanding of the glucose metabolism in these patients.
Conclusion
Although it appears rare, development of DM should be considered as a possible complication after resection of a hepatocellular carcinoma associated with paraneoplastic hypoglycemia. All reported cases developed DM in the first week postoperatively, and all cases required long-term administration of insulin.
The authors confirm the agreement from the pet owner in using the medical records of the patient with scientific purpose. The authors also state the patient described here was managed according to contemporary standards of care.
Contributor Notes
