Adult-Onset Nesidioblastosis Causing Hypoglycemia an Important Clinical Entity and Continuing Treatment Dilemma

PAPER Adult-Onset Nesidioblastosis Causing Hypoglycemia An Important Clinical Entity and Continuing Treatment Dilemma

Ronald M. Witteles, MD; Francis H. Straus II, MD; Sonia L. Sugg, MD; Mahalakshmana Rao Koka, MD; Eduardo A. Costa, MD; Edwin L. Kaplan, MD

Hypothesis: Nesidioblastosis is an important cause of insulin-dependent diabetes mellitus, pancreatic exo- adult hyperinsulinemic hypoglycemia, and control of this crine insufficiency, and need for reoperation. disorder can often be obtained with a 70% distal pancreatectomy. Results: Of 32 adult patients who underwent surgical ex- ploration for hyperinsulinemic hypoglycemia at our insti- Design: The records of all adult patients operated on tution, 27 (84%) were found to have 1 or more insulino- for hypoglycemia between 1974 and 1999 were mas, and 5 (16%) were diagnosed with nesidioblastosis. reviewed retrospectively. Patients with the pathologic Each patient with nesidioblastosis underwent a 70% distal diagnosis of nesidioblastosis were contacted for pancreatectomy. Follow-up duration for the 5 patients follow-up (1.5-21 years) and are presented. Patients’ ranged from 1.5 to 21 years, with 3 patients (60%) asymp- results were compared with those of 36 other individu- tomatic and taking no medications, and 2 patients (40%) als with this disorder who were previously reported in experiencing some recurrences of hypoglycemia. The 2 pa- the literature. tients with recurrences are now successfully treated with a calcium channel blocker, an approach, to our knowledge, Setting: The University of Chicago Medical Center (Chi- never before reported for adult-onset nesidioblastosis. cago, Ill), a tertiary care facility. Conclusions: Nesidioblastosis is an uncommon but clini- Patients: A consecutive sample of all patients operated cally important cause of hypoglycemia in the adult popu- on for hypoglycemia. lation, and must always be considered in a patient with a presumptive preoperative diagnosis of insulinoma. This Interventions: Seventy percent distal pancreatectomy study indicates that a 70% distal pancreatectomy is of- for all patients with nesidioblastosis, and maintenance ten successful in controlling hypoglycemia, and rarely therapy with verapamil hydrochloride for 2 patients. results in diabetes mellitus. However, the optimal treatment of this disorder remains to be determined. Main Outcome Measures: Achievement of normo- glycemia with and without medication, development of Arch Surg. 2001;136:656-663

EORGE F. LAIDLAW coined been reported in the literature; most have the term “nesidioblasto- short or no follow-up data. Though From the Departments of sis” in 1938, combining throughout the years there has been some Pathology (Dr Straus) and Surgery (Drs Witteles, Koka, the Greek words for is- confusion regarding the terminology, it is and Kaplan), University of let (nesidion) and builder now generally accepted that the terms “en- Chicago Pritzker School of G(blastos) to emphasize that in this condi- docrine cell dysplasia,” “islet cell hyper- Medicine, Chicago, Ill; and the tion cells differentiate and bud from the plasia,” “islet cell hypertrophy,” “micro- Department of Oncologic and pancreatic ductal epithelium to form new adenomatosis,” and “islet hypertrophy” are Endocrine Surgery, Medical islet tissue.1 While nesidioblastosis is rec- included in the diagnosis of nesidioblas- College of Wisconsin, ognized as the most common cause of neo- tosis.4 While most patients with an insu- Milwaukee (Dr Sugg). During natal hyperinsulinemia, it is believed to be linoma have a discrete lesion in one por- the time of this study, Dr Costa rare in adults, accounting for 0.5% to 7% tion of the pancreas, and multiple was a visiting scholar in of all cases of hyperinsulinemic hypogly- insulinomas are usually associated with Endocrine Surgery at the cemia. The first case report was by Sand- multiple endocrine neoplasia type 1 University of Chicago Pritzker 2 School of Medicine. Dr Witteles ler et al in 1975, and the largest case se- (MEN-1) syndrome, nesidioblastosis is al- 3 is presently a resident at ries was reported by Harness et al in 1981. most always characterized by a prolifera- Stanford University School of To our knowledge to date, only 36 total tion of abnormal B cells throughout the Medicine, Stanford, Calif. cases of adult-onset nesidioblastosis have entire pancreas.5 We report on the cases

(REPRINTED) ARCH SURG/ VOL 136, JUNE 2001 WWW.ARCHSURG.COM 656

PD The medical records of all adult patients operated on for hypoglycemia at the University of Chicago Medi- cal Center between 1974 and 1999 were reviewed. Of 32 cases identified, 27 (84%) were diagnosed with 1 or more insulinomas, and 5 (16%) were diagnosed with nesidioblastosis without an accompanying insulinoma. The patients with nesidioblastosis were contacted for follow-up, and appropriate laboratory testing was performed when possible. I Reports of adult-onset nesidioblastosis in the literature were found by a search of MEDLINE, and these patients were analyzed by symptoms, extent of I pancreatectomy, and outcome. Outcome was mea- sured by control of hypoglycemia and hypoglycemic symptoms with and without medication, need for reoperation, evidence of pancreatic exocrine insufficiency, and development of insulin-dependent Figure 2. Patient 4. A site of proliferating pancreatic ductules (PD), some of diabetes mellitus. which are lined by cuboidal cells with a similar morphology to islet cells in the 3 closely associated islets (I) (hematoxylin-eosin, original magnification ϫ 250).

Figure 3. Patient 2. Several small islet cells (arrows) forming from and around proliferating small ductules (PD) (hematoxylin-eosin, original magnification ϫ 150).

Figure 1. Patient 5. Concentration of medium-sized islets (arrows) on small Figures 2 3 4 integrally associated ductules (hematoxylin-eosin, original magnification ϫ and proliferating pancreatic ductules ( , , , 100). and 5). Four of the 5 cases did exhibit foci of dilated pancreatic ducts or ductules, suggesting focal areas of duct obstruction (Figure 6), which might be a factor in stimu- of 5 individuals with adult-onset nesidioblastosis who we lating islet hyperplasia and nesidioblastosis. Many me- have treated at the University of Chicago Medical Cen- dium- and large-sized islets showed irregular non-oval ter (Chicago, Ill) and present a review of the literature outlines. Scattered islets of varying size contained large pertaining to this disease. hyperchromatic islet cell nuclei (dysplasia, Figure 7). Some of these were 6 to 8 times larger than smaller ad- RESULTS jacent nuclei. Immunohistochemical staining showed that the increased number of islet cells still contained the usual In each of the following 5 cases of hyperinsulinism, no ratios of insulin-, glucagon-, somatostatin-, and poly- islet cell tumor was found in multiple sections of the 60% peptide-secreting cells. to 80% distal pancreatectomy specimens examined. All cases showed a diffuse increase in islet tissue, forming CASE 1 islets that varied in size (Figure 1). Some pancreatic sections showed that nearly 50% of the cross-sectional area In 1979, a 22-year-old man was admitted to the Univer- was made up of islet tissue. Many small- and medium- sity of Chicago Medical Center, having experienced diz- sized islets were present in close association with small ziness, lightheadedness, and headaches for 3 months. His

(REPRINTED) ARCH SURG/ VOL 136, JUNE 2001 WWW.ARCHSURG.COM 657

©2001 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 Figure 4. Patient 3. Small ductule with a very closely associated (arrow) Figure 6. Patient 3. A dilated pancreatic duct filled with acellular mucoid medium-sized, irregularly shaped islet cell containing one large hyperchromatic material. Similar-appearing ducts were present in 4 of the 5 described islet cell nucleus (hematoxylin-eosin, original magnification ϫ 400). patients (hematoxylin-eosin, original magnification ϫ 400).

Figure 5. Patient 1. Another ductule with a very closely associated Figure 7. Patient 4. A large islet cell occupying most of the field containing medium-sized islet cell with considerable variation in islet cell nuclear size one very large and several larger than usual hyperchromatic islet cell nuclei and chromisity (hematoxylin-eosin, original magnification ϫ 450). (hematoxylin-eosin, original magnification ϫ 550).

blood glucose level was 2.3 mmol/L (42 mg/dL), and at at the time of operation. A 70% distal pancreatectomy that time, his insulin levels were inappropriately high. was performed, removing all pancreatic tissue to the left A 70% distal pancreatectomy was performed, removing of the superior mesenteric vessels. Pathological analysis all pancreatic tissue to the left of the superior mesen- was consistent with nesidioblastosis (Figure 3), and no teric vessels. Pathological examination results were con- insulinoma was found grossly or microscopically. sistent with nesidioblastosis (Figure 5), and no insuli- Postoperatively, the patient was initially asymptom- noma was found grossly or microscopically. atic and normoglycemic, and she was discharged with- At 21 years’ follow-up, the patient was found to be out any complications. Soon afterwards, she developed normoglycemic and generally asymptomatic, and with- recurrence of symptomatic hypoglycemic episodes. She out evidence of pancreatic exocrine insufficiency. At times was begun on a regimen of verapamil hydrochloride and of heavy alcohol use when he did not eat, he stated that has remained asymptomatic and normoglycemic, tak- he had symptoms of hypoglycemia. ing 80 mg by mouth, 3 times daily for the last 15 years without evidence of pancreatic exocrine insufficiency. CASE 2 CASE 3 In 1985, a 37-year-old woman was referred to the Uni- versity of Chicago Medical Center for evaluation of hy- In 1990, a 42-year-old woman had 11⁄2 years of diapho- poglycemia and symptoms of lethargy and anxiety. She resis, shaking, palpitations, confusion, dysphagia, dys- had a low blood glucose level (1.9 mmol/L [35 mg/dL]) arthria, diplopia, lightheadedness, and a 9.07-kg weight and an elevated serum insulin level (158 pmol/L). A 72- gain. Her symptoms usually occurred while awakening hour fast resulted in insulin values of 51 to 92 pmol/L and during exercise, and they were usually relieved by during periods of severe hypoglycemia. An exploratory eating. She had undergone a glucose tolerance test at an laparotomy was performed, and no insulinoma was found outside hospital that demonstrated a glucose level of 1.1

(REPRINTED) ARCH SURG/ VOL 136, JUNE 2001 WWW.ARCHSURG.COM 658

©2001 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 mmol/L (20 mg/dL), but no further workup was pur- 35 sued for the next year. In October 1990, the patient de- Splenic Hepatic Gastroduodenal SMA veloped severe hypoglycemia during a 72-hour fast (glu- 30 cose, 1.7 mmol/L [30 mg/dL]; insulin, 105 pmol/L). Magnetic resonance imaging, computed tomographic 25 (CT) scan images, and results of an abdominal ultraso- 20 U/mL

nography all demonstrated a nonspecific thickening in µ the tail of the pancreas. The patient underwent an ex- 15

ploratory laparotomy in December 1990. A 1-cm nod- Insulin, 10 ule found near the tail of the pancreas was excised and interpreted as consistent with diffuse islet cell hyperpla- 5 sia during frozen section examination. No other masses were palpated, but results of an intraoperative ultra- 0

sound demonstrated a 1-cm hypoechoic lesion in the head 0 20 40 60 of the pancreas. This region was explored surgically, but Time, s no discrete mass was found. A biopsy of this region also Figure 8. Calcium-stimulated arteriogram in patient 5. Note that in this yielded the diagnosis of islet cell hyperplasia. At this point, patient, the calcium-stimulated arteriogram resulted in a rise of insulin when a 70% distal pancreatectomy was performed. The patho- calcium was infused into both the splenic and gastroduodenal arteries, logic diagnosis was consistent with nesidioblastosis (Fig- suggesting abnormal islet cells throughout multiple vascular territories of the pancreas. SMA indicates superior mesenteric artery. To convert insulin ure 4 and Figure 6), and no insulinoma was found grossly values from microunits per milliliter to picomoles per liter, multiply the given or microscopically. value by 7.175. The patient experienced no complications postoperatively, and she was discharged normoglycemic and asymptomatic. The patient has remained asymptomatic mal levels without treatment throughout the following and normoglycemic (blood glucose, 3.3-5.4 mmol/L [60- several days. At 3 months’ follow-up, the patient had nor- 97 mg/dL]) after 9 years of follow-up; she is taking no mal fasting blood glucose and insulin levels (6 mmol/L medications and is without evidence of pancreatic exo- [112 mg/dL] and 69 pmol/L, respectively). At 3 years’ fol- crine insufficiency. low-up, he remained asymptomatic and normoglycemic (blood glucose, 5-6.7 mmol/L [90-120 mg/dL]), tak- CASE 4 ing no medications, and with no evidence of pancreatic exocrine insufficiency. In 1993, a 45-year-old man experienced diaphoresis and chills and was found to have a blood glucose level of 1.7 CASE 5 mmol/L (30 mg/dL) at an outside hospital. A 72-hour fast was performed, but the blood glucose level never fell be- In February 1999, a 48-year-old woman began experi- low 2.8 mmol/L (50 mg/dL), and no further workup was encing dizziness and occasional night sweats. In April attempted at that time. During the next 3 years, he in- 1999, she was hospitalized after having a seizure and was termittently experienced diaphoresis and forgetfulness, found to have a blood glucose level of 0.9 mmol/L (16 and gained 34 kg. In April 1996, the patient was found mg/dL). A 72-hour fast resulted in hypoglycemia, and the at the side of the road in his car and could not remem- patient had a glucose level of 1.7 mmol/L (31 mg/dL), ber how he had gotten there. He was taken to a hospital, an insulin level of 49 pmol/L, and a nonsuppressed C pep- where he was found to have a blood glucose level of 1.3 tide level. Test results for oral hypoglycemic agents were mmol/L (24 mg/dL), an insulin level greater than 1910 normal. A CT scan, an abdominal ultrasound, and an oc- pmol/L, a C peptide level of 9.92 nmol/L (normal, Ͻ1.3 treotide scan were performed in an attempt to localize a nmol/L), and a proinsulin level 360 pmol/L (normal, Ͻ21 possible insulinoma, but the results for all 3 tests were pmol/L). Test results for insulin antibodies and oral hy- normal. A calcium-stimulated arteriogram (Figure 8) poglycemic agents were normal. demonstrated a rise in insulin level when calcium was An abdominal CT scan was negative for a tumor, and injected into both the gastroduodenal and splenic an abdominal ultrasound was suggestive of a 1-cm hy- arteries. poechoic lesion at the junction of the head and body of The patient underwent an exploratory laparotomy the pancreas that was consistent with a possible insuli- in May 1999. The entire pancreas was mobilized, pal- noma. He underwent an operation in June 1996 for a pre- pated, and surgically explored, and no pancreatic masses sumed insulinoma, and though a thorough inspection of were palpable. An intraoperative ultrasound demon- the pancreas showed it to be diffusely nodular, palpa- strated a homogeneous pancreas with no masses suggestion and intraoperative ultrasound results failed to dem- tive of an insulinoma. On frozen section examination, a onstrate a tumor. A 70% distal pancreatectomy was per- biopsy specimen from the head of the pancreas was con- formed, removing all of the pancreatic tissue to the left sistent with diffuse islet cell hyperplasia. Suspecting ne- of the superior mesenteric vessels. Pathological features sidioblastosis, a 70% distal pancreatectomy was per- were consistent with nesidioblastosis (Figure 7), and no formed, removing all pancreatic tissues to the left of the insulinoma was found grossly or microscopically. superior mesenteric vessels. Pathologically, the re- Immediately postoperatively, the blood glucose level sected pancreas showed diffuse nesidioblastosis and rose to 9.2 mmol/L (165 mg/dL), and returned to nor- prominent islet hyperplasia, with islets constituting 25%

(REPRINTED) ARCH SURG/ VOL 136, JUNE 2001 WWW.ARCHSURG.COM 659

©2001 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 of the total pancreatic tissue (Figure 1). No insulinomas tients with nesidioblastosis or an insulinoma have shown were found grossly or microscopically. inappropriately high insulin and C peptide levels. Our last In the immediate postoperative period, the patient patient had nondetectable proinsulin values, and this may had a benign course with mild hyperglycemia (quickly perhaps be a clue to the diagnosis of nesidioblastosis. returning to normal values), with no episodes of hypo- Conventional radiologic testing is not reliably help- glycemia (blood glucose, 3.5-8.3 mmol/L [63-149 mg/dL]). ful in differentiating an insulinoma from nesidioblasto- During the first postoperative year, the patient experi- sis. While the diagnosis of nesidioblastosis should be con- enced intermittent recurrences of hypoglycemic symp- sidered when imaging studies (CT, magnetic resonance toms, with blood glucose levels ranging from 2.5 to 3.3 imaging, ultrasonography, and angiography) do not lo- mmol/L (45-60 mg/dL) at the time of her symptomatic calize a discrete lesion of the pancreas, it should not be episodes. Of note, her symptoms often occurred post- relied on. Many insulinomas are small (Ͻ 1 cm), and even prandially after having consumed large, carbohydrate- in experienced hands, the sensitivity of these radiologic rich meals. During studies while the patient was eating 6 studies for an insulinoma is only 50% to 80%, and false- meals per day, she was asymptomatic, with blood glu- positive results occur as well. cose levels higher than 4.4 mmol/L (80 mg/dL). A fast- The most promising imaging technique for local- ing test was performed, during which the patient re- ization of an insulinoma is transgastric ultrasonography mained asymptomatic, with normal blood glucose levels of the pancreas. In the absence of detecting an insuli- for the first 24 hours. She became hypoglycemic (blood noma using this technique, the suspicion of a diagnosis glucose, 2.7 mmol/L [48 mg/dL]) 26 hours into the fast, of adult nesidioblastosis should be heightened. One should and she had neuroglycopenic symptoms at that time. Given then proceed to percutaneous transhepatic portal ve- her tolerance for a prolonged fast, she was discharged and nous sampling, or better yet, to a calcium-stimulated ar- was administered 80 mg of verapamil hydrochloride by teriogram of the pancreas.7,8,26-30 In this latter test, a se- mouth, to be taken 3 times daily, she was advised to con- lective arteriogram is performed of the gastroduodenal tinue eating 6 small low-carbohydrate meals daily. To date, artery, splenic artery, superior mesenteric artery, and he- she has remained normoglycemic and asymptomatic on patic artery. After dye is injected into a given artery, a this regimen. bolus of calcium is then injected. Insulin is sampled from a catheter placed in a hepatic vein. When an insulinoma COMMENT is present, calcium injection results in the release of insulin only if the specific artery that feeds the area of the Hypoglycemia of infancy is caused by nesidioblastosis of pancreas that contains that tumor is tested. On the other the pancreatic islet cells, and finding an insulinoma in hand, in the presence of nesidioblastosis, calcium injec- this age group is virtually unknown. On the other hand, tion into all of the pancreatic vessels may result in insu- in the adult, hypoglycemia characterized by a Whipple lin release from the entire pancreas, as the abnormal is- triad with hyperinsulinemia and increased C peptide lev- lets are usually found throughout the entire pancreas els is almost always owing to an insulinoma. These tu- (Figure 8).8,17,18,22,24 This finding would be highly sug- mors are single in most patients; however, in the case of gestive of nesidioblastosis in the absence of the MEN-1 patients with MEN-1 syndrome, multiple insulinomas of syndrome, in which multiple insulinomas are often pre- the pancreas are the rule. sent. In 1975, nesidioblastosis was first described in an While pancreatic resection represents the defini- adult hypoglycemic patient.2 The peptide testing was done tive treatment for an insulinoma, patients often require at the University of Chicago Medical Center. Since that medication to control their hypoglycemia preopera- time, 35 other cases of this entity have been reported in tively or postoperatively. Diazoxide is the most fre- the English-language literature. In this article, we add re- quent medication used,3,7,14,15,20,23,24 but its use is accom- ports of 5 new patients, who were treated at the Univer- panied by troublesome adverse effects, including fluid sity of Chicago Medical Center throughout the last 21 retention, hypotension, hypertrichosis, and bone mar- years (Table 1). While adult nesidioblastosis is still un- row suppression. Other medications that have been used common, it certainly is not rare. Previous studies have with varying success include somatostatin analogs20 and estimated that this entity accounts for 0.5% to 7% of all glucocorticoids. cases of adult hyperinsulinemic hypoglycemia.6-9 Adult There have been reports regarding the use of cal- nesidioblastosis occurred in 16% of patients who were cium channel–blocking agents to treat patients with insu- operated on for hypoglycemia at our institution during linoma,30-32 neonatal nesidioblastosis,33,34 and reactive hy- the last several decades. poglycemia.35 To our knowledge, cases 2 and 5 in this study The preoperative differentiation of adult nesidioblas- represent the first reported use of calcium channel– tosis from an insulinoma is difficult. Most patients, in- blocking agents in patients with adult nesidioblastosis. cluding our 5 individuals, had hypoglycemia that oc- The extent of surgical resection for adult-onset ne- curred with fasting or exercise, just as happens in a patient sidioblastosis remains controversial. As presented in with insulinoma. Thus, a fasting test result is usually ab- Table 2, most surgeons have performed a distal pan- normal. Service et al,25 on the other hand, reported the cases createctomy, thus removing the part of the pancreas to of 5 adults (Table 1, patients 32-36) who suffered from the left of the superior mesenteric vein. A few patients severe postprandial or “reactive” hypoglycemia, and who have been treated by 90% to 95% (near-total) pancre- were found to have nesidioblastosis of the pancreas when atectomy, while others were treated by small distal re- they were operated on. At the time of hypoglycemia, pa- sections. Near-total pancreatectomy would seem logical

(REPRINTED) ARCH SURG/ VOL 136, JUNE 2001 WWW.ARCHSURG.COM 660

Age/Sex Symptoms Px, % Outcome (Drug) Follow-up, y Study Year Reference 59/M “Hypoglycemic stupor” 67 IDDM 1 1975 2 57/M None (abnormal lab value) 95 NG 3 1981 3 28/F “Symptoms of hypoglycemia” 33 HG/IDDM after 100% Px 5 1981 3 49/M Dizzy/sweats/blurry vision 60 NG 3 1981 3 37/F Syncope 50 NG (high insulin) 1 1981 3 20/M Convulsions/syncope/sweats 95 IDDM 0.25 1981 3 21/M Dizziness/HA/palpitations 70-75 NG . . . 1987 5 84/F Syncope/convulsions/sweats 80 IDDM 2 1995 6 66/M NA 80 Slight HG (diazoxide) . . . 1984 7 69/M Sweats/coma/coma 70 NG . . . 1997 8 32/M Confusion/sweats/syncope 50 NG 0.25 1986 9 48/M NA 95 NG . . . 1976 10 15/M NA 70-90 HG/NG after 2nd operation . . . 1980 11 58/F Dizziness/lethargy/sweats 75 NG 2 1981 12 56/F Weakness/palpitations/syncope Body/tail Continued syncope . . . 1983 13 29/F Blurry vision/clammy skin 75 HG refractory to drugs . . . 1983 13 53/M Confusion/vertigo/gait changes Partial NG . . . 1983 13 47/F Dizziness/lethargy 60 NG (diazoxide) 1.3 1983 14 26/M NA 75 HG (diazoxide) . . . 1984 15 36/M Confusion/amnesia/sweats Distal NG 0.3 1985 16 43/F Hunger 75 NG 2.5 1988 17 46/M Dizziness/syncope 80 HG/IDDM after 100% Px 6 1989 18 44/M Confusion/HA 90 IDDM 11 1989 18 29/F Syncope/seizure/HA/sweats 50 HG/NG after 95% Px 5 1989 19 63/F Dizziness/lethargy/confusion 95 NG 8 1989 19 42/F Seizure/confusion/weakness 50 HG/NG after 95% Px 2 1989 19 80/M Dizziness/weakness 70 HG/NG (octreotide) 7 1990 20 52/M NA NA NG 10 1995 21 50/M Syncope/dizziness/sweats 70† NG 1.25 1996 22 24/F Amnesia/fatigue/hunger 70 HG (diazoxide) 2.5 1997 23 66/F Syncope Distal NG (diazoxide) . . . 1998 24 37/M Seizures 70 NG 0-3 1999 25 16/M Sweats/unresponsiveness 70 NG 0-3 1999 25 72/M Blurry vision/sweats/dysarthria 70 NG 0-3 1999 25 72/M Sweats/diplopia/light-headed 70 NG 0-3 1999 25 78/F Flushing/blurry vision/confusion Limited HG 0.25 1999 25 22/M Dizziness/HA/lightheadedness 70 NG 21 2000 Current 37/F Lethargy/anxiety 70 HG/NG (verapamil hydrochloride) 15 2000 Current 42/F Sweats/palpitations/confusion 70 NG 9 2000 Current 45/M Sweats/chills/amnesia 70 NG 3.5 2000 Current 48/F Sweats/dizziness/HA/seizure 70 HG/NG (verapamil) 1.5 2000 Current

*Only 36 cases of adult-onset nesidioblastosis have been reported in the English-language literature prior to this study. Px indicates pancreatectomy; IDDM, insulin-dependent diabetes mellitus; NA, not available; NG, normoglycemic; HG, hypoglycemic; HA, headache. †Pancreatectomy was proximal, as part of a Whipple procedure.

because of the diffuse nature of the islet cell disease. This procedure has resulted in the resolution of hypoglyce- Table 2. Outcome vs Extent of Operation* mia; however, 40% of patients developed insulin- dependent diabetes mellitus. Exocrine pancreatic insuf- 90%- 60%- 95% Px 89% Px Ͻ60% Px ficiency occurs with some frequency as well. A 60% to (n=5) (n = 26) (n=6) 80% distal pancreatectomy results in a cure in about half of the patients, with no need for medication; an addi- Normoglycemic (no medication) 60 50 33 Normoglycemic (with medication) 0 19 0 tional 19% of patients were normoglycemic with medi- Hypoglycemic 0 23 67 cation. Of note is the fact that insulin-dependent diabe- IDDM 40 8 0 tes mellitus occurred in only 8% of this group. Smaller distal resections seem to be of only limited value thera- *All data are percentages. According to a summary of all reported cases in peutically. the literature (including the cases from this article), patients are more likely to be cured of hypoglycemia with larger pancreatic resections, but they are also In our 5 patients, each of whom was treated by dis- more likely to develop insulin-dependent diabetes mellitus (IDDM). Px indicates tal pancreatectomy, all are alive with follow-up dura- pancreatectomy. tions up to 21 years. Three are normoglycemic and using no medication, while 2 others are normoglycemic surgical approach to be satisfactory but not perfect. At following treatment with a calcium channel blocker. None the Mayo Clinic (Rochester, Minn), most patients with developed diabetes mellitus. Thus, we have found this adult nesidioblastosis have had resections of the distal

(REPRINTED) ARCH SURG/ VOL 136, JUNE 2001 WWW.ARCHSURG.COM 661

©2001 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 pancreas to the right of the superior mesenteric vessels. in an adult with persistent hyperinsulinemic hypoglycemia. Am J Clin Pathol. 1989; This has resulted in an 80% cure rate with operation alone 91:336-340. 18. Fong TL, Warner NE, Kumar D. Pancreatic nesidioblastosis in adults. Diabetes (Jon van Heerden, MD, oral communication, 2000). At Care. 1989;12:108-114. the University of Michigan, Ann Arbor (Norman Thomp- 19. McHenry C, Newell K, Chejfec G, et al. Adult nesidioblastosis: an unusual cause son, MD, oral communication, 2000), each patient is of fasting hypoglycemia. Am Surg. 1989;55:366-369. tested preoperatively with diazoxide. Those who re- 20. Mozell EJ, Woltering EA, O’Dorisio TM, et al. Adult onset nesidioblastosis: response of glucose, insulin, and secondary peptides to therapy with sandostatin. spond favorably receive a distal pancreatectomy while oth- Am J Gastroenterol. 1990;85:181-188. ers who do not do well taking this medication undergo 21. Losada J, Sarria R, Val JFF, Ariztegui MAL. DNA Ploidy and PCNA index in pan- subtotal pancreatectomy. creatic lesions producing hyperinsulinemic hypoglycemia. J Surg Oncol. 1995; In summary, adult nesidioblastosis resulting in hy- 59:14-17. poglycemia is an important entity that should be under- 22. Kim HK, Shong YK, Han DJ, Cho Y, Lee IC, Kim GS. Nesidioblastosis in an adult with hyperinsulinemic hypoglycemia. Endocr J. 1996;43:163-167. stood by all surgeons who operate for an insulinoma. The 23. Fuller PH, Ehrlich AR, Susil B, Zeimer H. Insulin gene expression in adult-onset optimal treatment of this condition requires further study. nesidioblastosis. Clin Endocrinol (Oxf). 1997;47:245-250. Whether or not a nonoperative approach to this disease 24. Rinker RD, Friday K, Aydin F, Jaffe BM, Lambiase L. Adult nesidioblastosis: a will ever be practical remains to be determined. case report and review of the literature. Dig Dis Sci. 1998;43:1784-1790. 25. Service FJ, Natt N, Thompson GB, et al. Noninsulinoma pancreatogenous hypoglycemia: A novel syndrome of hyperinsulinemic hypoglycemia in adults inde- Supported in part by the Nathan and Frances Goldblatt So- pendent of mutations in Kir6.2 and SUR1 genes. J Clin Endocrinol Metab. 1999; ciety for Cancer Research, Chicago, Ill. 84:1582-1589. Presented at the 108th Scientific Session of the West- 26. Roche A, Raissonnier A, Gillon-Savouret MC. Pancreatic venous sampling and ern Surgical Association, Dana Point, Calif, November 14, arteriography in localizing insulinomas and gastrinomas: procedure and results in 55 cases. Radiology. 1982;145:621-627. 2000. 27. Doppman JL, Miller DL, Chang R, Shawker TH, Gorden P, Norton JA. Insulino- Corresponding author: Edwin Kaplan, MD, Depart- mas: localization with selective intra arterial injection of calcium. Radiology. 1991; ment of Surgery, the University of Chicago Medical Cen- 178:237-241. ter, 5841 S Maryland Ave, MC 5031, Chicago, IL 60637. 28. Miller DL. Endocrine angiography and venous sampling. Radiologic Clin North Am. 1993;31:1051-1067. 29. Doppman JL, Chang R, Fraker DL, et al. Localization of insulinomas to regions REFERENCES of pancreas by intraarterial stimulation with calcium. Ann Intern Med. 1995;123: 269-273. 30. O’Shea D, Rohrer-Theus AW, Lynn JA, Jackson JE, Bloom SR. Localization of 1. Laidlaw GF. Nesidioblastoma: the islet cell tumor of the pancreas. Am J Pathol. insulinomas by selective intra arterial calcium injection. J Clin Endocrinol Metab. 1938;14:125-139. 1996;81:1623-1627. 2. Sandler R, Horwitz DL, Rubenstein AH, Kuzuya H. Hypoglycemia and endoge- 31. De Marinis L, Barbarino A. Calcium antagonists and hormone release: effects of nous hyperinsulinism complicating diabetes mellitus. Am J Med. 1975;59:730- verapamil on insulin release in normal subjects and patients with islet-cell tu- 736. mor. Metabolism. 1980;29:599-604. 3. Harness JK, Geelhoed GW, Thompson NW, et al. Nesidioblastosis in adults: a 32. Imanaka S, Matsuda S, Ito K, Matsuoka T, Okada Y. Medical treatment for inop- surgical dilemma. Arch Surg. 1981;116:575-580. erable insulinoma: clinical usefulness of diphenylhydantoin and diltiazem. Jpn 4. Jaffe R, Hashida Y, Yunis EJ. Pancreatic pathology in hyperinsulinemic hypo- J Clin Oncol. 1986;16:65-71. glycemia of infancy. Lab Invest. 1980;42:356-365. 33. Ulbrecht JS, Schmeltz R, Aarons JH, Greene DA. Insulinoma in a 94-year-old 5. Carlson T, Eckhauser ML, DeBaz B, Khiyami A, Park CH. Nesidioblastosis in an woman: long-term therapy with verapamil. Diabetes Care. 1986;9:186-188. adult: an illustrative case and collective review. Am J Gastroenterol. 1987;82: 34. Bas F, Darendeliler F, Demirkol D, Bundak R, Saka N, Gunoz H. Successful therapy 566-571. with calcium-channel blocker (nifedipine) in persistent neonatal hyperinsulin- 6. Walmsley D, Matheson NA, Ewen S, Himsworth RL, Bevan JS. Nesidioblastosis emic hypoglycemia of infancy. J Pediatr Endocrinol Metab. 1999;12:873-878. in an elderly patient. Diabet Med. 1995;12:542-545. 35. Lindley KJ, Dunne MJ, Kane C, et al. Ionic control of beta-cell function in nesid- 7. Harrison TS, Fajans SS, Floyd JC, et al. Prevalence of diffuse pancreatic beta is- ioblastosis: a possible therapeutic role for calcium channel blockade. Arch Dis let cell disease with hyperinsulinism: problems in recognition and management. Child. 1996;74:373-378. World J Surg. 1984;8:583-589. 36. Sanke T, Nanjo K, Kondo M, Nishi M, Moriyama Y, Miyamura K. Effect of cal- 8. Lee WL, Won JGS, Chiang JH, Hwang JI, Lee CH, Tsay SH. Selective intra- cium antagonists on reactive hypoglycemia associated with hyperinsulinemia. arterial calcium injection in the investigation of adult nesidioblastosis: a case re- Metabolism. 1986;35:924-927. port. Diabet Med. 1997;14:985-988. 9. Weinstock G, Margulies P, Kahn E, Susin M, Abrams G. Islet cell hyperplasia: an unusual cause of hypoglycemia in an adult. Metabolism. 1986;35:110-117. DISCUSSION 10. Bradley JR, Heileman JP. Hyperinsulinism due to pancreatic islet cell hyperplasia: report of a case in an adult. Ariz Med. 1976;33:543-545. Jon A. van Heerden, MD, Rochester, Minn: The University 11. Dahms BB, Landing BH, Blaskovics M, Roe TF. Nesidioblastosis and other islet of Chicago, and Dr Kaplan in particular, have long been rec- cell abnormalities in hyperinsulinemic hypoglycemia of childhood. Hum Pathol. 1980;11:641-649. ognized for their contributions and expertise in the field of en- 12. Nathan DM, Axelrod L, Proppe KH, Wald R, Hirsch HJ, Martin DB. Nesidioblas- dogenous hyperinsulinism. This presentation of a rare, yet im- tosis associated with insulin-mediated hypoglycemia in an adult. Diabetes Care. portant, facet of hyperinsulinism is another wonderful example 1981;4:383-388. of their many contributions. 13. Weidenheim KM, Hinchey WM, Campbell WG Jr. Hyperinsulinemic hypoglyce- Almost 20 years ago, in 1981 to be exact, 2 distinguished mia in adults with islet-cell hyperplasia and degranulation of exocrine cells of members of our association, Norman Thompson, who is with the pancreas. Am J Clin Pathol. 1983;79:14-24. us currently, and Jay Harness, who is not at this meeting, pre- 14. Keller A, Stone AM, Valderrama E, Kolodny H. Pancreatic nesidioblastosis in adults: sented to the members of the Western Surgical Association their Report of a patient with hyperinsulinemic hypoglycemia. Am J Surg. 1983;145: early experience with adult nesidioblastosis. Although there has 412-416. been a great deal of skepticism about the existence of an adult 15. Bauman WA, Merkle LN, Rachman RA, Mitsuto SM. Hypoglycemia in a diabetes nurse care coordinator. Diabetes Care. 1984;7:88-91. nesidioblastosis over the last 2 decades, sufficient data are cur- 16. Campbell IL, Harrison LC, Ley CJ, Colman PG, Ellis DW. Nesidioblastosis and rently available to refute any doubting Thomases. multifocal pancreatic islet cell hyperplasia in an adult: clinicopathologic features Our surgical group has now operated on 11 patients with and in vitro pancreatic studies. Am J Clin Pathol. 1985;84:534-541. adult nesidioblastosis; 1 for hypergastrinemia or the Zollinger- 17. Albers N, Lohr M, Bogner U, Loy V, Kloppel G. Nesidioblastosis of the pancreas Ellison syndrome, which required a Whipple operation and is

(REPRINTED) ARCH SURG/ VOL 136, JUNE 2001 WWW.ARCHSURG.COM 662

©2001 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 now cured and normogastrinemic 14 years later, and 10 for hy- ficult is distinguishing patients with NIPH from those with perinsulinism treated by distal pancreatectomy a` la the Kaplan insulinomas. This is particularly relevant in deciding which lo- group. Nine of these 11 patients have had an excellent re- calization studies to obtain preoperatively. Dr Thompson just sponse rate to their subtotal pancreatectomies. indicated that he uses endoscopic ultrasound, and we have re- I would like to ask Dr Kaplan and his coauthors to edu- cently enjoyed good success with this as well. However, I doubt cate all of us a little bit more on this rare entity by expanding that the quality of endoscopic ultrasound is universally good perhaps on some of the following queries: (1) Were all speci- enough to rely on across the entire country. If the authors can mens carefully looked at and “bread-loafed” to exclude the pres- suggest a reliable method to distinguish these patients preop- ence of a small occult insulinoma? We know that insulinomas eratively, either biochemically or by imaging techniques, I would can be very small, and are we sure that none were missed in be most interested. these 5 patients? (2) Is calcium stimulation arteriography now Richard C. Thirlby, MD, Seattle, Wash: In your last pa- performed at the University of Chicago if all preoperative lo- tient, did you obtain an octreotide scan, and do you have any calizing modalities are negative in the patient with definite en- experience with efficacy of octreotide in these patients? If it dogenous hyperinsulinism? (3) The extent of pancreatic re- works, the long-acting octreotide will be available quite soon section was somewhat arbitrary in the authors’ experience since on compassionate use. It might be a preferable alternative to a arteriography was utilized in a single patient only. How radi- three-times-a-day pill. cal in fact should we be? I don’t know what the answer is. Doesn’t Dr Kaplan: Dr. van Heerden, we agree that it is critical to it make more sense that nesidioblastosis should involve the en- examine the pathologic specimens very carefully to be certain tire pancreas, and is this perhaps the reason for failures in ba- that a small insulinoma is not missed. It is also true that ne- sically 40% of your patients and in approximately 20% of our sidioblastosis can occur in some patients who have an insuli- patients? I think we need to ask the question: is true localized noma. Thus, it is critical intraoperatively to be certain that an nesidioblastosis in fact an entity? It makes more sense teleo- insulinoma is not present. By the way, Dr Thompson, the term logically to me that nesidioblastosis should involve the entire nesidioblastosis includes what was called islet cell hyperplasia gland. (4) Should all patients with this diagnosis (ie, nesidio- in the past. blastosis) have a trial of calcium-channel antagonist therapy The best studies of calcium-stimulated arteriography have prior to, or instead of, pancreatic resections? been done by Dr John Doppman at the NIH [National Insti- Norman W. Thompson, MD, Ann Arbor, Mich: Dr van tutes of Health, Bethesda, Md]. He has shown that this test is Heerden recalled that we did present a paper with 6 patients excellent in regionalizing the area of the pancreas that con- with this entity almost 19 years ago. We have had 4 or 5 since, tains the insulinoma. More studies must be done on patients and our current way of evaluating patients with organic hypo- with nesidioblastosis to determine whether or not the entire glycemia is to do endoscopic ultrasound. If their results are nega- pancreas is stimulated to release insulin by this technique. If tive in demonstrating an insulinoma, that is the group that we so, this will add to the possibility that nesidioblastosis can be do selective pancreatic arterial stimulation with calcium. That diagnosed preoperatively. Dr Thompson’s suggestion that this is the way one can diagnose this disease preoperatively. It doesn’t test be done when transgastric ultrasonography is negative is a mean that every patient with hypoglycemia requires the study; good one. in fact, only about 5% of insulinoma suspects do. If results are How radical should the operation be? Drs van Heerden positive, then we pretreat these patients as we have for the last and Thompson have great experience in this field, and thus far, 15 years with diazoxide. If they respond to diazoxide, then the no perfect operation has been described. Perhaps Dr Thomp- 70% resection seems to be adequate. If they don’t respond to son’s idea of using diazoxide, or a calcium channel blocker first, diazoxide, then we do an 85% resection. During the last 15 years, and then judging the size of the operation by the response, may we have only had 1 recurrence after a 70% resection, and that have merit. was a patient, a young woman, who did require an 85% resec- Should all adult nesidioblastosis patients have a trial of a tion. We haven’t had any experience with calcium-blocking calcium channel blocker before an operation is considered? Cer- agents, but those drugs could be substituted, I think, for di- tainly if we can diagnose this entity preoperatively with cer- azoxide in pretesting. I think the diagnosis can routinely be made tainty, then a trial of medical therapy is appropriate. preoperatively. I don’t think we have to say that it’s really that Drs Service and Grant have described a new syndrome in difficult to make that diagnosis ahead of time and thus avoid which patients with nesidioblastosis had postprandial or reac- the possibility of being forced to consider a possible “blind dis- tive hypoglycemia, while insulinomas are present with fasting tal pancreatectomy at operation.” hypoglycemia. One of our patients had post prandial hypogly- My question is, how do you separate hyperplasia from ne- cemia, but the others had fasting hypoglycemia. At the Mayo sidioblastosis? As I looked at those slides, several looked like Clinic, those patients with reactive hypoglycemia are not be- islet cell hyperplasia rather than nesidioblastosis. ing operated on. That is our practice as well. Clive S. Grant, MD, Rochester: As Dr van Heerden al- A question was asked about octreotide scanning. Octreo- luded to previously, we have had experience, almost exclu- tide scanning is much better for the diagnosis of a gastrinoma sively over the last 5 years, with nesidioblastosis, but we think than it is for insulinoma. We tried an octreotide scan on our that it represents the pathologic counterpart of a new syn- last patient, and it was negative. drome reported by Dr John Service, of noninsulinoma pan- In summary, there is no question in my mind that adult ne- creatogenous hypoglycemia [NIPH]. Patients with NIPH al- sidioblastosis does exist as a distinct entity. I am cautiously en- most all have had a negative 72-hour fast, but when they are thusiastic that calcium channel blockers may be of value. If we symptomatic, they have hypoglycemia documented biochemi- can diagnose this entity preoperatively, and know that an insu- cally with concomitantly elevated insulin and C peptide lev- linoma is not present, I would be enthusiastic in trying medical els, and a negative screen for sulfonylurea. What remains dif- therapy first, and not operating on patients with this condition.

(REPRINTED) ARCH SURG/ VOL 136, JUNE 2001 WWW.ARCHSURG.COM 663