Abstract
The accuracy of factory-calibrated continuous glucose monitoring (fCGM) within hypoglycemic ranges, especially under the status of chronic hyperinsulinemic hypoglycemia like insulinomas, remains an issue. Even so, fCGM is known to be useful for detecting hypoglycemia unawareness in insulinoma cases. A 25-year-old woman presenting with sudden unconsciousness was diagnosed with insulinoma; fCGM facilitated diagnosis by continuous monitoring for hypoglycemia. Before surgery, she was treated with continuous and frequent bolus infusions of 50% glucose via central venous catheter. To evaluate the accuracy of fCGM values in this case, a comparison between fCGM and capillary blood glucose (CBG) values was also performed. According to the simultaneously measured values, those of fCGM were largely in accordance with those of CBG. Moreover, compared with the previously reported case not having glucose infusions via central venous catheter, both the mean absolute relative differences (MARDs) and the absolute differences (Δ glucose) between fCGM and CBG values were larger in the present case, although no significant differences of MARDs and Δ glucose between the two cases were observed in several different conditions including fasting, post-meal, hypoglycemia, and others. Therefore, we should note possible increased differences between fCGM and CBG values in cases using frequent intravenous glucose infusions as well as case-dependent differing levels of consistency between them.
INSULINOMA is a rare neuroendocrine tumor, but accounts for most of the functioning endocrine tumors in epidemiological studies across Japan [1]. Insulinoma presents as hyperinsulinemic hypoglycemia, which can cause various symptoms and severely impair patient quality of life [2, 3]. A delayed diagnosis of insulinoma remains a clinical issue [4]. One of the main causes of delay is hypoglycemia unawareness, which can obscure a patient’s notice of hypoglycemia and worsen the outcomes [3-5]. In this context, continuous glucose monitoring (CGM) has been demonstrated to be a useful tool to detect hypoglycemia in cases of insulinoma with hypoglycemia unawareness [5-7]. However, most currently available CGM systems require calibration by frequent self-monitoring of capillary blood glucose (CBG) levels, and allow only a retrospective view of past hypoglycemic events, excepting a device equipped with a real-time CGM system [8].
The factory-calibrated continuous glucose monitoring (fCGM) systems Freestyle® Libre Pro and Freestyle® Libre (Abbot Diabetes Care, Oxon, UK) recently have become clinically available in Japan. These fCGM systems require no CBG-based calibration due to their factory calibration, and the Freestyle® Libre allows real-time access to the patient’s current glucose levels [9, 10]. We previously reported a case in which these features of the Freestyle® Libre contributed to avoiding episodes of hypoglycemia by detecting unnoticed hypoglycemia, thus improving the patient’s health-related quality of life [11]. However, the accuracy of fCGM systems in the hypoglycemic range under chronic hyperinsulinemic hypoglycemia such as that found with insulinoma has not been fully examined [12, 13].
Here we report a case of insulinoma in which fCGM facilitated diagnosis by continuously monitoring hypoglycemia. In addition, we encountered a potential pitfall in using fCGM in insulinoma cases by analysis of simultaneous measurements of fCGM and CBG values.
Case presentation
A 25-year-old woman was referred to our hospital for treatment of severe hypoglycemia. About two weeks before her hospitalization, she suddenly became unconscious at night, and her family took her to a nearby hospital. Her point-of-care glucose level was 38 mg/dL. She rapidly recovered consciousness after glucose administration. Her serum insulin and C-peptide levels (5.7 μU/mL and 1.12 ng/mL, respectively) were relatively high despite hypoglycemia (38 mg/dL). Her family history and hyperinsulinemic hypoglycemia suggested the possibility of insulinoma. She had been diagnosed as a carrier of a multiple neoplasia type1 (MEN1) gene mutation. She also had family history of MEN1, including her biological father. No MEN1-associated abnormalities had been noticed in her regular follow-up, except for a relatively low fasting glucose level (60 mg/dL), which had been recorded two years before her hospitalization.
Her height was 159.6 cm; bodyweight, 59.5 kg; axillary temperature 36.9°C; pulse, 78 bpm; and blood pressure, 102/61 mmHg. There were no remarkable findings on physical examination. She reported neither increased appetite nor bodyweight gain during the past several years. She had no history of anti-diabetic agent use, and none of the laboratory data suggested other endocrine diseases (Table 1). Anti-insulin antibody was negative. She was subjected to a 72-hour fasting test, which was stopped 3.25 hours after beginning because of her low point-of-care glucose level (49 mg/dL) for her safety due to the possibility of hypoglycemia unawareness, even though she remained asymptomatic. Fasting test revealed hyperinsulinemic hypoglycemia (plasma glucose, 57 mg/dL; serum insulin, 8.2 μU/mL; serum C-peptide, 1.54 ng/mL). Glucagon administration after the fasting test revealed a significantly increased plasma glucose level (86 mg/dL). fCGM was also performed using the Freestyle Libre Pro, which revealed the existence of hypoglycemia unawareness despite 10–20 mL/hour continuous infusion of 50% glucose via central venous catheter, with bolus infusions for low point-of-care glucose levels (Fig. 1a) and oral diazoxide administration (up to 1,200 mg/day). Abdominal dynamic computed tomography (CT) scans showed an early-enhanced tumor 16 mm in diameter at the head of the pancreas (Fig. 2a, b). Endoscopic ultrasonography (EUS) demonstrated a tumor, which the CT scan had detected, and some other small tumors at the head and body of the pancreas (Fig. 2c). EUS-guided fine needle aspiration cytology showed that the tumor at the pancreatic head was positive for insulin staining. Taking these findings together, we made a clinical diagnosis of insulinoma.
Table 1
Laboratory data of the patient at hospitalization.
| Complete blood count |
Biochemistry |
Selected hormones and others |
| WBC |
4,400/μL |
AST |
21 IU/L |
Insulin |
19.9 mU/L |
| RBC |
403 × 104/μL |
ALT |
33 IU/L |
C-peptide |
3.55 ng/mL |
| Hb |
12.3 g/dL |
ALP |
196 IU/L |
|
|
| Plt |
21.0 × 104/μL |
LDH |
135 IU/L |
TSH |
1.830 IU/L |
|
|
T-Bil |
0.4 mg/dL |
Free T4 |
1.010 ng/dL |
|
|
Amy |
85 IU/L |
ACTH |
21.9 pg/mL |
|
|
TP |
6.1 g/mL |
Cortisol |
12.3 μg/dL |
|
|
ALB |
3.5 g/mL |
GH |
0.48 ng/mL |
|
|
Cr |
0.60 mg/dL |
IGF-1 |
201 ng/mL |
|
|
BUN |
9 mg/dL |
PRL |
21.0 ng/mL |
|
|
Na |
142 mEq/L |
LH |
26.7 mIU/mL |
|
|
K |
4.0 mEq/L |
FSH |
4.8 mIU/mL |
|
|
Cl |
107 mEq/L |
Estradiol |
59.1 pg/mL |
|
|
Ca |
9.4 mg/dL |
Glucagon |
169 pg/mL |
|
|
T-Cho |
112 mg/dL |
Gastrin |
85 pg/mL |
|
|
CK |
25 mg/dL |
Intact PTH |
31 pg/mL |
|
|
CRP |
<0.1 mg/dL |
|
|
|
|
Plasma glucose |
105 mg/dL |
Anti-Insulin Ab |
<125 nU/mL |
|
|
HbA1c |
4.8% |
|
|
WBC, white blood cell; RBC, red blood cell; Hb, hemoglobin; Plt, platelet; AST, aspartate aminotransferase; ALT, alanine aminotransferase; ALP, alkaline phosphatase; LDH, lactate dehydrogenase; T-Bil, total bilirubin; TP, total protein; Alb, albumin; Cr, creatinine; BUN, blood urea nitrogen; T-Chol, total cholesterol; CK, creatine kinase; CRP, C-reactive protein; HbA1c, hemoglobin A1c; IGF-1, insulin like growth factor-1; Anti-Insulin Ab, anti-insulin antibody
After 33-day hospitalization, the patient received prompt enucleation surgery of the pancreatic head tumor because of the insufficient effect of oral diazoxide administration. Pathological examination revealed a tumor 16 mm in diameter (Fig. 3a, b). Immunochemical studies demonstrated positive staining for chromogranin A, synaptophysin, and insulin (Fig. 3c, d, e). The Ki-67 proliferative index of the tumor was less than 1%. Three weeks after surgery, her fasting plasma glucose level was 71 mg/dL.
Comparison between simultaneously measured fCGM and CBG values
As the differences between fCGM and CBG values in the present case (Case 1) seemed relatively large, we added analyses of the mean absolute relative differences (MARDs) and absolute differences (Δ glucose) between fCGM (Freestyle Libre Pro) and CBG values in Case 1 in several different conditions including fasting, post-meal (2 hours after meals), hypoglycemia (defined as CBG values below 70 mg/dL), and others. These values were also retrospectively compared with the results in the previously reported insulinoma cases using fCGM or CBG at Kyoto University Hospital [11]. These procedures were approved by the ethical committee of Kyoto University Hospital (local identifier: R1789) and conducted in accordance with the Declaration of Helsinki. The CBG values were routinely measured using OneTouch® Ultra® (Johnson & Johnson Company, Vacaville, CA, USA). In both cases, self-monitoring of blood glucose using capillary blood was routinely performed in fasting, 2 hours after initiating meals, and at bedtime, in addition to the time just before taking snacks and the time the patients suspected hypoglycemia. The common definitions of all the conditions like fasting, post-meal, and hypoglycemia were used in the analyses of MARDs and Δ glucose. In the present case, fCGM values were largely in accordance with those of CBG (MARDs: 14.0 ± 11.0%, Δ glucose: –10.5 ± 14.0 mg/dL), although MARDs at the time of hypoglycemia were relatively higher in this case (Table 2). Under hypoglycemic condition, the absolute values of CBG were 51–69 [median 59.5, interquartile range (IQR) 6.8] mg/dL in Case 1 and 29–69 [median 58, IQR 12] mg/dL in Case 2, while those of fCGM were 40.0–71.0 [median 60.0, IQR 5.0] mg/dL in Case 1 and 40.0–84.9 [median 55.9, IQR 18.6] mg/dL in Case 2. Moreover, both MARDs and the absolute differences showed no significant dissimilarities between the two insulinoma cases in all examined conditions (Table 2).
Table 2
Comparisons between Factory-calibrated Continuous Glucose Monitoring (fCGM) and Capillary Blood Glucose (CBG) Values
|
Case 1 (Present Case) |
Case 2 (Reference #11) |
p value (Case 1 vs. 2: <0.05) |
| N |
MARD
(%) |
Δ glucose
(mg/dL) |
N |
MARD
(%) |
Δ glucose
(mg/dL) |
MARD
(%) |
Δ glucose
(mg/dL) |
| Fasting |
27 |
15.4 ± 14.2 |
–9.4 ± 16.2 |
9 |
3.5 ± 15.2 |
–2.7 ± 8.3 |
0.51 |
0.25 |
| Post-meal |
76 |
12.6 ± 10.1 |
–10.2 ± 13.3 |
28 |
11.6 ± 16.1 |
–10.2 ± 12.3 |
0.09 |
0.99 |
| Hypoglycemia |
4 |
21.5 ± 17.8 |
–2.3 ± 16.9 |
35 |
0.8 ± 18.8 |
–0.4 ± 10.0 |
0.32 |
0.75 |
| Total |
229 |
14.0 ± 11.0 |
–10.5 ± 14.0 |
69 |
8.1 ± 17.7 |
–7.6 ± 12.0 |
0.15 |
0.42 |
The fCGM values obtained by Freestyle Libre Pro. We calculated the mean absolute relative difference (MARD) and the absolute difference of fCGM vs. CBG (Δ glucose). We checked CBG at fasting, post meal (two hours after meals), and hypoglycemic range (≦70 mg/dL). For the comparison of MARD and Δ glucose at each timing between the two cases, p values were calculated for Welch’s t-test. All data were expressed as mean ± standard deviation (SD). MARD is defined as 100 × |fCGM value – CBG value|/CBG value; Δ Glucose is defined as “fCGM value-CBG value”.
Discussion
Insulinoma is a rare neuroendocrine tumor that produces excess endogenous insulin, resulting in hypoglycemia [4]. As previously shown, hypoglycemia unawareness is not rare in patients with insulinoma, and it can obscure a patient’s notice of hypoglycemia [3].
CGM systems have been used in insulinoma cases with hypoglycemia unawareness [5-7]. However, CGM systems require frequent CBG-based calibrations, which can be a major obstacle to the use of CGM systems as a modality for identifying hypoglycemia in insulinoma cases with hypoglycemia unawareness. fCGM systems have been recently introduced as a tool to continuously estimate blood glucose levels based on interstitial fluid glucose levels. fCGM systems avoid frequent CBG-based calibrations because of their factory calibration [9, 10]. Although we previously identified the possible utility of fCGM for detecting hypoglycemia unawareness and preventing severe hypoglycemia [11], the method not only facilitates the diagnosis of insulinoma but also is useful in monitoring therapeutic responses against hypoglycemia such as glucose load via a central venous catheter and oral diazoxide administration (Fig. 1a, b), which contributed to establishing the need for prompt surgery in the present case. In addition, indirect and limited access to CGM values with a time lag was also an obstacle to the use of CGM systems as a modality for identifying hypoglycemia in insulinoma cases with hypoglycemia unawareness. Recently, real-time CGM, which can directly alert patients to hypoglycemia, has become clinically available. Comparison of the clinical usefulness between fCGM and real-time CGM in insulinoma cases with hypoglycemia unawareness remains to be examined.
The accuracy of fCGM has been examined in healthy people and subjects with type 1 and 2 diabetes [8, 9, 12]. However, according to some reports [9, 10, 13, 14], there remains debate concerning the accuracy of fCGM values under hypoglycemic conditions. It might be difficult to observe frequent hypoglycemic episodes in studies of patients with diabetes [14], and they have been rarely examined in the case of insulinomas, which are instances of inappropriate hyperinsulinemic hypoglycemia. To this point, we analyze the comparison between fCGM and CBG values in two pre-operative insulinoma cases (Table 2). As shown in Table 2, MARD values and Δ glucose between fCGM and CBG in both cases were acceptable, indicating that fCGM values were largely in accordance with those of CBG. The tendency is consistent under hypoglycemic condition, although the frequency of CBG-recorded hypoglycemia in Case 1 was limited due to the continuous glucose infusion via central venous catheter. Thus, fCGM should be useful under hypoglycemia equivalent to our cases. We did not directly compare the fCGM values with plasma glucose levels. The fCGM values were a little lower than the CBG values, a tendency that is compatible with the previous report of fCGM values within relatively lower glycemic ranges [12]. Based on our results of fCGM and today’s limited access to CGM values with a time lag, Freestyle® Libre also can be an option in patients with insulinoma.
In the present case, the values of MARDs and Δ glucose were relatively larger than in the previously reported case [11], although the values were acceptable and there was no significant difference between the two cases (Table 2). It is of note that these case-to-case comparisons might be affected by lot-to-lot variations of the fCGM sensors. Thus, in research settings, case-to-case comparisons should be performed with simultaneous multiple fCGM and/or other CGM sensors. In another point of view, the relatively larger values of MARDs and Δ glucose might be related to the delay in fCGM changes, as the previous study mentioned [10, 13]. In the present case, the glycemic fluctuations were relatively obvious because of the frequent bolus 50% glucose intravenous administrations and changes of continuous infusion rates of 50% glucose via central venous catheter (Fig. 1a, b), both of which were not performed in Case 2. Moreover, the differences of MARDs and Δ glucose were not obvious in post-meal conditions, which might also support the possible influence of frequent intravenous glucose infusions on MARDs and Δ glucose values since intravenous glucose infusions were performed mainly in fasting and hypoglycemic conditions. In addition, the influence of meal intake on glycemic fluctuations might be limited under hyperinsulinemic status with insulinomas. Therefore, we should note possible increased differences between fCGM and CBG values in cases using frequent intravenous glucose infusions as well as case-dependent differing levels of consistency.
In summary, we describe here a case of insulinoma with hypoglycemia unawareness in which diagnosis of insulinoma was facilitated by fCGM-mediated monitoring of hypoglycemia. The fCGM values were largely in accordance with the CBG values in insulinoma cases, although we should pay attention to the glycemic fluctuations induced by frequent intravenous glucose infusions.
Disclosure
None of the authors have any potential conflicts of interest associated with this research. This study did not receive any specific grants from funding agencies in the public, commercial or not-for-profit sectors.
References
- 1 Ito T, Igarashi H, Nakamura K, Sasano H, Okusaka T, et al. (2015) Epidemiological trends of pancreatic and gastrointestinal neuroendocrine tumors in Japan: a nationwide survey analysis. J Gastroenterol 50: 58–64.
- 2 Service FJ, McMahon MM, O’Brien PC, Ballard DJ (1991) Functioning insulinoma-incidence, recurrence, and long-term survival of patients: a 60-year study. Mayo Clin Proc 66: 711–719.
- 3 Wexler DJ, Macias-Konstantopoulos W, Forcione DG, Xiong L, Cauley CE, et al. (2018) Case 23-2018: a 36-year-old man with episodes of confusion and hypoglycemia. N Engl J Med 379: 376–385.
- 4 Imamura M, Nakamoto Y, Uose S, Komoto I, Awane M, et al. (2015) Diagnosis of functioning pancreaticoduodenal neuroendocrine tumors. J Hepatobiliary Pancreat Sci 22: 602–609.
- 5 Murakami T, Yamashita T, Yabe D, Masui T, Teramoto Y, et al. (2017) Insulinoma with a history of epilepsy: still a possible misleading factor in the early diagnosis of insulinoma. Intern Med 56: 3199–3204.
- 6 Munir A, Choudhary P, Harrison B, Heller S, Newell-Price J (2008) Continuous glucose monitoring in patients with insulinoma. Clin Endocrinol 68: 912–918.
- 7 Kikuchi T, Chujo D, Takahashi K, Takahashi N, Tanno Y, et al. (2017) Insulinoma presenting with reactive hypoglycemia: evaluating the effect of tumor resection via continuous glucose monitoring. Intern Med 56: 3067–3071.
- 8 Aberer F, Hajnsek M, Rumpler M, Zenz S, Baumann PM, et al. (2017) Evaluation of subcutaneous glucose monitoring systems under routine environmental conditions in patients with type 1 diabetes. Diabetes Obes Metab 19: 1051–1055.
- 9 Bailey T, Bode BW, Christiansen MP, Klaff LJ, Alva S (2015) The performance and usability of a factory-calibrated flash glucose monitoring system. Diabetes Technol Ther 17: 787–794.
- 10 Heinemann L, Freckmann G (2015) CGM Versus FGM; or, continuous glucose monitoring is not flash glucose monitoring. J Diabetes Sci Technol 9: 947–950.
- 11 Sugawa T, Murakami T, Yabe D, Kashima R, Tatsumi M, et al. (2018) Hypoglycemia unawareness in insulinoma revealed with flash glucose monitoring systems. Intern Med 57: 3407–3412.
- 12 Fokkert MJ, van Dijk PR, Edens MA, Abbes S, de Jong D, et al. (2017) Performance of the FreeStyle Libre Flash glucose monitoring system in patients with type 1 and 2 diabetes mellitus. BMJ Open Diabetes Res Care 5: e000320.
- 13 Sekido K, Sekido T, Kaneko A, Hosokawa M, Sato A, et al. (2017) Careful readings for a flash glucose monitoring system in nondiabetic Japanese subjects: individual differences and discrepancy in glucose concentration after glucose loading. Endocr J 64: 827–832.
- 14 Kumagai R, Muramatsu A, Fujii M, Katakura Y, Ito K, et al. (2019) Comparison of glucose monitoring between Freestyle Libre Pro and iPro2 in patients with diabetes mellitus. J Diabetes Investig 10: 851–856.
