Case Report
PROINSULIN-PREDOMINANT PANCREATIC NEUROENDOCRINE TUMOR-INDUCED HYPOGLYCEMIA AFTER ROUX-EN-Y GASTRIC BYPASS SURGERY
Khary Edwards, MD; Lori de La Portilla, DO; Faryal S. Mirza, MD; Pooja Luthra, MD
ABSTRACT
Abbreviations: MEN1 = multiple endocrine neoplasia type 1; MRI = magnetic resonance imaging; NIPHS = noninsulinoma pancreatogenous hypoglycemia syndrome; PNET = pancreatic neuroendocrine tumor; RYGB = Roux-en-Y- gastric bypass
INTRODUCTION
Objective: To present a case of recurrent hypoglyce- mia following Roux-en-Y gastric bypass (RYGB) surgery whose etiology was determined to be a proinsulin-predom- inant pancreatic neuroendocrine tumor (a proinsulinoma). Methods: A case report along with a brief discussion and review of the pertinent literature is presented. Results: The patient is a 62-year-old female who presented with symptomatic hypoglycemia 11 years after RYGB surgery. Initial workup revealed low insulin levels with elevated proinsulin levels. A 72-hour fast confirmed the presence of proinsulin-induced hypoglycemia second- ary to a pancreatic neuroendocrine tumor (PNET). She underwent distal pancreatectomy with splenectomy and a PNET tumor was successfully removed with resolution of her symptoms. Conclusion: Hypoglycemia after RYGB surgery is a well-established syndrome. While there are several etiolo- gies for this, PNETs (including proinsulinomas) should be considered in the differential diagnosis in this population. Proinsulinomas are an increasingly recognized cause of hypoglycemia. Proinsulin levels must always be included as part of the workup of hypoglycemia in an adult. (AACE Clinical Case Rep. 2019;5:e339-e343)
Submitted for publication March 28, 2019 Accepted for publication July 2, 2019 From the Division of Endocrinology and Metabolism, University of Connecticut School of Medicine, Farmington, Connecticut. Address correspondence to Dr. Pooja Luthra, Division of Endocrinology and Metabolism, University of Connecticut School of Medicine, 263 Farmington Avenue, Farmington, CT 06030. E-mail: luthra@uchc.edu. DOI:10.4158/ACCR-2019-0148 To purchase reprints of this article, please visit: www.aace.com/reprints. Copyright © 2019 AACE.
Pancreatic neuroendocrine tumors (PNETs) are islet cell-derived neoplasms of the pancreas that produce hormonally active peptides including insulin, gastrin, glucagon, and vasoactive intestinal peptide (1). They represent the most common neuroendocrine tumor (2). Insulinomas, neuroendocrine tumors which produce insu- lin, represent the most common type of functional PNETs and are the most common cause of hypoglycemia associat- ed with a tumor (3). Proinsulin is produced by the pancre- atic beta cells and converted into equimolar amounts of C peptide and insulin via enzymatic cleavage. The biological activity of proinsulin is 10% that of insulin (4). Recent advances in laboratory assays make the differ- entiation between proinsulin and insulin possible (5). This has led to the discovery that proinsulin-producing PNETs (proinsulinomas) that predominantly secrete proinsulin with normal or near-normal levels of insulin secretion can also cause hypoglycemia (5,6). In 2017, only 16 cases of proinsulin-secreting PNETs had been reported (6). While there have been cases of insulinomas in patients after Roux-en-Y gastric bypass (RYGB) surgery (7), there are no reports of proinsulinomas after RYGB surgery. We pres- ent a case of hypoglycemia secondary to a proinsulinoma after RYGB surgery.
AACE CLINICAL CASE REPORTS Vol 5 No. 6 November/December 2019 e339
e340 Proinsulinoma after RYGB, AACE Clinical Case Rep. 2019;5(No. 6)
CASE REPORT
A 62-year-old female developed symptoms of hypo- glycemia (sweating, palpitations, shaking, and nausea) with episodes of witnessed loss of consciousness several months prior to presentation. Symptoms were prominent early in the morning, in fasting conditions, or at night and resolved after eating. She snacked frequently to avoid symptoms. She did not give a definite history of weight change related to her symptoms. She also had a history of a stable, 1.2-cm non-secreting pituitary macroadenoma and RYGB surgery 11 years prior to presentation with a total weight loss of 45 kg. There was no family history of hyper- calcemia, hypoglycemia, or endocrine tumors. She denied use of insulin or oral hypoglycemic agents. She had 1 episode of documented hypoglycemia at her primary care physician’s office, with loss of consciousness and a finger stick blood glucose of 28 mg/dL. She respond- ed quickly to glucose treatment. Fasting laboratory evalu- ation performed on a separate day showed normal serum glucose (99 mg/dL), creatinine, liver function, C peptide, and insulin levels. Proinsulin level was elevated (43.2 pmol/L; reference range is ≤8.0 pmol/L). Serum sulfonyl- urea screen was negative. Cosyntropin stimulation testing was performed to assess for possible adrenal insufficiency. Baseline serum cortisol level was 11 µg/dL. One hour after intramuscu- lar injection of 250 µg of cosyntropin, serum cortisol level appropriately increased to 23 µg/dL, excluding adrenal insufficiency. In addition, serum total and ionized calcium, 25-hydroxyvitamin D, and parathyroid hormone levels were normal. She was admitted for a supervised 72-hour fast. Her fasting serum blood glucose was 88 mg/dL, C peptide was 0.9 ng/mL, insulin was 3 µIU/mL, but proinsulin was elevated at 26.8 pmol/L (reference range is ≤8.0 pmol/L). At hour 72, she developed symptomatic hypoglycemia with venous blood glucose of 43 mg/dL. Laboratory test- ing revealed that insulin and C peptide levels were appro- priately low, sulfonylurea screen was negative, but proin- sulin level was elevated (Table 1). Her beta hydroxybutyr-
Table 1 72-Hour Fast Results
ate level was elevated at 1.4 mmol/L (reference range is 0.0 to 0.3 mmol/L). Low plasma insulin concentration with elevated proinsulin levels indicated a proinsulinoma or a predominantly proinsulin-secreting pancreatic tumor. She received glucagon at the end of the fast. Her blood glucose did not change and remained at 43 mg/dL at 10, 20, and 30 minutes after glucagon injection. Magnetic resonance imaging (MRI) of the abdomen (with and without contrast) was undertaken and demon- strated a 2.4-cm, mildly hypervascular lesion in the pancre- atic tail and a 0.9-cm, uncinate process lesion with mild T2 hyperintensity (Fig. 1). An indium (In-111) pentet- reotide scan performed was negative for a pentetreotide- avid neoplasm. She underwent an open distal pancreatectomy with splenectomy to remove the tumor without complication. Histopathology revealed a 2.5 × 2.4 × 2.2-cm, unifocal tumor from the pancreatic body (Fig. 2) that was immuno- reactive for insulin (Fig. 3). The antibody stain used was Novocastra liquid mouse monoclonal antibody (product code NCL-L-INSULIN, Leica Biosystems Inc., Buffalo Grove, IL). There was no lympho-vascular or perineural invasion. The tumor was histologic grade 1 with <2 mitosis per 2 mm2. Pathologic staging was pT2, pNX, pM N/A. Spleen and accessory spleen samples were negative for tumor involvement. Postoperatively, she did not have any further hypogly- cemic episodes and lost 7 kg of weight, as she was not snacking constantly. She has been symptom-free for at least 1 year of follow up. Fasting proinsulin level 1 month after surgery was 6.2 pmol/L (reference range is ≤8.0 pmol/L). She was also evaluated by neurosurgery for the nonfunctioning pituitary macroadenoma and was recom- mended conservative follow up as serial MRI did not show any change in size of the adenoma. Laboratory evaluation of pituitary function showed normal levels of insulin-like growth factor-1, adrenocor- ticotropic hormone, thyroid-stimulating hormone, free thyroxine, and alpha subunit. Her follicle-stimulating hormone and luteinizing hormone levels were in the appropriate reference ranges for postmenopausal women. Prolactin was mildly elevated at 23.79 ng/mL (reference range is 2.74 to 19.64 ng/mL). Genetic testing for multiple endocrine neoplasia type 1 (MEN1) by Invitae Corporation (San Francisco, CA) was negative for sequence changes, exonic deletions, or mutations in the genes CDKN1B and MEN1.
Patient’s value at hour 72 Reference range Beta-hydroxybutyrate 1.4 mmol/L 0.0-0.3 mmol/L 0.5 ng/mL 0.8-3.9 ng/mL C peptide
DISCUSSION
Glucose 43 mg/dL 2 µIU/mL ≤100 mg/dL 2.0-19.6 µIU/mL Insulin 26.4 pmol/L ≤8 pmol/L Proinsulin
Our patient presented with symptomatic hypogly- cemia in the setting of a prior history of RYGB surgery. Hypoglycemia (venous blood glucose <60 mg/dL) has been reported after RYGB surgery in up to 13% of patients in 1 study, with 0.7% of patients having severe hypogly-
Negative None Serum sulfonylurea screen
cemia (venous blood glucose <40 mg/dL) (8). Several mechanisms may contribute to hypoglycemia after RYGB surgery which have been reviewed by Salehi et al (9). Of these, dumping syndrome (postprandial hypoglycemia due to altered delivery of nutrition to the gut [10]) and noninsulinoma pancreatogenous hypoglycemia syndrome (NIPHS, a syndrome of postprandial hyperinsulinemic hypoglycemia due to beta cell hypertrophy [11]) are well reported. While these diagnoses were considered in our patient, she did not have the classic postprandial hypoglycemia associated with dumping syndrome and NIPHS, but instead had fasting hypoglycemia, more consistent with her PNET
diagnosis. The possibility of both syndromes coexisting was considered, but pathology of the resected pancreas did not display nesidioblastosis or beta cell hypertrophy to support NIPHS (11). In our patient, low plasma insulin level with low serum glucose and high proinsulin was indicative of a proinsulin- predominant PNET. Most patients with a negative 72-hour fast have beta-hydroxybutyrate levels >2.7 mmol/L because of suppressed insulin levels (12). In our case, the relatively low level of beta-hydroxybutyrate observed in the setting of normal insulin levels at the time of hypogly- cemia is likely due to the elevated proinsulin level and its insulin-like activity.
Proinsulinoma after RYGB, AACE Clinical Case Rep. 2019;5(No. 6) e341
Fig. 1. Magnetic resonance image of the abdomen and pelvis with and without contrast. There is a 2.4-cm, mildly hypervascular lesion in the pancreatic tail.
Fig. 2. Insulinoma with adjacent normal pancreas. This figure shows a well-circumscribed and encapsulated lesion comprised of monotonous-appearing cells on hematoxylin and eosin stain (×2).
e342 Proinsulinoma after RYGB, AACE Clinical Case Rep. 2019;5(No. 6)
A
B
Fig. 3. Immunohistochemical stain for insulin in the lesion and adjacent pancreas showing cyto- plasmic expression in the lesion cells (A; ×2) with no expression in the adjacent normal pancreas (B; ×20).
The patient did not become hypoglycemic until the end of the 72-hour fast. This has also been reported in the literature. In a study of 205 patients with insulinoma, 93% of patients had a positive fast within 48 hours, and 99% within 72 hours. Hence the recommendation is to do a stan- dard 72-hour fast to detect the very small percentage of patients who have a positive fast between 48 and 72 hours. It is likely that our patient falls into this group (13). Hyperinsulinemia usually completely suppresses glycogenolysis and glycogen stores are preserved in patients with insulinomas due to persistently elevated insu- lin levels. Therefore, when glucagon is given at the end of a prolonged fast, blood glucose levels quickly improve due to glycogenolysis. This was not observed in our patient, who did not have an increase in glucose with administra- tion of glucagon after hypoglycemia, at hour 72 of her fast. Although this is not the usual response with insuli- nomas, it is possible that because proinsulin only has 10% of the activity of insulin, she may have had a low level of glycogenolysis during the 72-hour fast, causing glycogen stores to be depleted by the end of her fast. This would also explain the delayed presentation of hypoglycemia at hour 72 of the fast. However, similar findings were not observed in other reported cases of proinsulin-secreting neuroendo- crine tumors (6). There have also been case reports of patients with insulinomas whose tumor insulin production was appro- priately suppressed during prolonged fasting, but have hyperinsulinemic hypoglycemia when given glucose or glucagon (14). In these cases, the patients did not develop neuroglycopenic symptoms during the 72-hour fast, and had appropriate fasting ketosis and elevated beta-hydroxy- butyrate production. However, this was not observed in our
patient. Proinsulinomas are rare occurrences, and if not for advances in laboratory assays that are now able to differ- entiate between insulin and proinsulin, these tumors would be identified as insulinomas. In our patient, the tumor was identified as an insulin-secreting tumor, since the antibody used for staining has a specificity for human insulin but cannot identify proinsulin (Fig. 3). Since proinsulin is the precursor molecule to insulin, one may expect higher insulin levels in tumors secreting proinsulin. Proinsulinomas however, appear to secrete intact proinsulin molecules that are not enzymatically cleaved, unlike insulin secreted from native pancreatic beta cells and insulinomas. This was demonstrated in a review of 16 cases with the majority of the cases having signifi- cantly elevated fasting proinsulin levels in the setting of normal C peptide and insulin levels (6). MEN1 was also a consideration in our patient as she presented with 2 of the common tumors associated with MEN1, both a PNET tumor and a pituitary adenoma. MEN1 can be diagnosed clinically in patients who pres- ent with 2 of the 3 classic tumors (parathyroid, pituitary, and pancreatic tumors) (15). However, she denied a fami- ly history of pancreatic, pituitary, or parathyroid tumors as well as hypercalcemia. MEN1 genetic testing was also negative. For bariatric surgery patients, an insulinoma (or proin- sulinoma) should be suspected if they present with symp- toms of neuroglycopenia (confusion, fatigue, weakness, syncope) or adrenergic symptoms of hypoglycemia (irrita- bility, perspiration, palpitations) while fasting or during the night. This is distinct from hyperinsulinemic hypoglyce- mia associated with the late dumping syndrome or NIPHS which occurs in the postprandial state, usually within 1
Proinsulinoma after RYGB, AACE Clinical Case Rep. 2019;5(No. 6) e343
DISCLOSURE
The authors have no multiplicity of interest to disclose.
REFERENCES
1. Ro C, Chai W, Yu VE, Yu R. Pancreatic neuroendocrine tumors: biology, diagnosis, and treatment. Chin J Cancer. 2013;32: 312-324.
2. Yao JC, Hassan M, Phan A, et al. One hundred years after “carci- noid”: epidemiology of and prognostic factors for neuroendo- crine tumors in 35,825 cases in the United States. J Clin Oncol. 2008;26:3063-3072.
3. Iglesias P, Díez JJ. Management of endocrine disease: a clini- cal update on tumor-induced hypoglycemia. Eur J Endocrinol. 2014;170:R147-R157.
4. Kitabchi AE. Proinsulin and C-peptide: a review. Metabolism.
1977;26:547-587.
5. Chia CW, Saudek CD. The diagnosis of fasting hypoglycemia due to an islet-cell tumor obscured by a highly specific insulin assay. J Clin Endocrinol Metab. 2003;88:1464-1467.
6. Murtha TD, Lupsa BC, Majumdar S, Jain D, Salem RR. A systematic review of proinsulin-secreting pancreatic neuroendo- crine tumors. J Gastrointest Surg. 2017;21:1335-1341.
7. Mulla CM, Storino A, Yee EU, et al. Insulinoma after bariatric surgery: diagnostic dilemma and therapeutic approaches. Obes Surg. 2016;26:874-881.
10.
8. Goldfine AB, Patti ME. How common is hypoglycemia after gastric bypass? Obesity (Silver Spring). 2016;24:1210-1211. 9. Salehi M, Vella A, McLaughlin T, Patti ME. Hypoglycemia after gastric bypass surgery: current concepts and controversies. J Clin Endocrinol Metab. 2018;103:2815-2826. van Beek AP, Emous M, Laville M, Tack J. Dumping syndrome after esophageal, gastric or bariatric surgery: pathophysiology, diagnosis, and management. Obes Rev. 2017;18:68-85.
11. Service GJ, Thompson GB, Service FJ, Andrews JC, Collazo- Clavell ML, Lloyd RV. Hyperinsulinemic hypoglycemia with nesidioblastosis after gastric-bypass surgery. New Engl J Med. 2005;353:249-254.
12. Service FJ, O’Brien PC. Increasing serum betahydroxybutyrate concentrations during the 72-hour fast: evidence against hyper- insulinemic hypoglycemia. J Clin Endocrinol Metab. 2005;90: 4555-4558.
13. Service FJ, Natt N. The prolonged fast. J Clin Endocrinol Metab.
2000;85:3973-3974.
to 3 hours of eating. Patients may experience weight gain from the need to constantly eat to prevent their symptoms. Fasting laboratory evaluation may show inappropriately elevated levels of insulin, C peptide, or proinsulin with low blood glucose. This should prompt further evaluation for the cause of hypoglycemia. Computed tomography or MRI of the abdomen are regarded as first-line imaging modalities for localizing PNETs tumors, including insulinomas and proinsulino- mas. Multi-detector computed tomography has sensitivity and specificity ranges from 63 to 82% and 83 to 100%, respectively, with a detection rate as high as 94.4%. MRI has sensitivity and specificity ranges between 85 to 100% and 75 to 100%, respectively, with a detection rate of 50 to 94% (16). While the overall sensitivity of indium (In-111) pentetreotide scanning for PNETs is 70 to 90%, the sensi- tivity for insulinomas is 50 to 70% (17). Endoscopic ultra- sound can also be used to determine the precise location and provide samples for histopathology. The sensitivity and specificity of endoscopic ultrasound can range from 83 to 93% and 83.7 to 95%, respectively, for detection of PNETs and has been used to detect proinsulinomas not identified on noninvasive imaging (18). If the suspicion for a PNET remains high, additional invasive techniques including selective arterial calcium stimulation can be undertaken. This technique has been shown to have a detection rate of 88 to 100% for insuli- nomas (19). However, the response to calcium stimula- tion in proinsulinomas differs from that of insulinomas. Proinsulinomas do not show a significant increase in proinsulin level with calcium stimulation in contrast to the increase in insulin level with insulinomas (19,20). This modality may not be as useful in identifying proinsulin- omas. There are currently no reports to suggest that any imaging modality can differentiate between insulinomas and proinsulinomas without histopathologic testing.
CONCLUSION
14. Wiesli P, Schmid C, Perren A, Pfammatter T, Spinas GA, Keller U. Hypoglycemia in response to glucose and glucagon in insulinoma patients with a negative prolonged fast: functional and morphological properties. J Endocrinol Invest. 2004;27:832-838. 15. Thakker RV, Newey PJ, Walls GV, et al. Clinical practice guide- lines for multiple endocrine neoplasia type 1 (MEN1). J Clin Endocrinol Metab. 2012;97:2990-3011.
16. Kartalis N, Mucelli RM, Sundin A. Recent developments in imaging of pancreatic neuroendocrine tumors. Ann Gastroenterol. 2015;28:193-202.
17. Tamm EP, Bhosale P, Lee JH, Rohren EM. State-of-the-art imaging of pancreatic neuroendocrine tumors. Surg Oncol Clin N Am. 2016;25:375-400.
18. Rodriguez A, Canto MI, Makary MA. Endoscopic localization and tattooing of a proinsulinoma for minimally invasive resection. Pancreas. 2011;40:474-477.
This case highlights the need to consider proinsulin- predominant PNETs and insulin-producing PNETs in the differential diagnoses for hypoglycemia in patients with history of RYGB surgery. Proinsulinomas remain rare occurrences, likely due to the lack of assays that can differ- entiate between insulin and proinsulin. With the advent of specific assays for proinsulin, we suspect that proin- sulinomas will be reported more frequently as a cause of hypoglycemia.
ACKNOWLEDGMENT
19. Won JG, Tseng HS, Yang AH, et al. Intra-arterial calcium stimu- lation test for detection of insulinomas: detection rate, responses of pancreatic peptides, and its relationship to differentiation of tumor cells. Metabolism. 2003;52:1320-1329.
20. Fadini GP, Maran A, Valerio A, et al. Hypoglycemic syndrome in a patient with proinsulin-only secreting pancreatic adenoma (proinsulinoma). Case Rep Med. 2011;2011:930904.
We thank the Department of Pathology at Hartford Hospital in Hartford, Connecticut for providing the images for Figures 2 and 3.
