A case of congenital hyperinsulinism presenting with diabetes after long-term diazoxide therapy

Abstract

Congenital hyperinsulinism (CHI) is the most common form of persistent hypoglycemia in infants, and diazoxide is the most widely used drug for its treatment. Diazoxide suppresses insulin secretion and attenuates hypoglycemia by binding to sulfonylurea receptor 1 and activating K_ATP channels. While the short-term side effects of this drug, such as edema and blood cell abnormalities, are well known, the clinical course after its long-term oral administration remains unclear. Furthermore, there are currently no case reports clearly demonstrating a causal relationship between diazoxide and impaired glucose tolerance. We herein describe the case of a 9-year-old girl with CHI complicated with Kabuki syndrome who presented with impaired glucose tolerance due to decreased initial insulin secretion and insulin resistance caused by obesity resulting from diazoxide medication. This is a rare case of the insufficient effects of insulin due to the oral administration of diazoxide, and provides insights for managing the long-term administration of diazoxide to children.

Introduction

Congenital hyperinsulinism (CHI) is the most common form of persistent hypoglycemia in infants [1, 2] and is estimated to occur in approximately 1 in 30,000 children in Japan [3–5]. The causes of CHI are divided into three categories. The first two categories, non-syndromic CHI, caused by loss-of-function mutations in ATP-gated K channels (K_ATP channels) on the epithelial membrane of pancreatic β cells, and syndromic CHI associated with underlying diseases, such as Kabuki syndrome and Beckwith–Wiedemann syndrome, manifest as persistent CHI. The third category is transient CHI, which is often a complication in short for gestational age (SGA) infants and diabetic mothers.

Diazoxide acts as a K_ATP channel opener and is widely used to treat CHI. Its oral administration may be discontinued in some cases as hypoglycemia improves with age [6]. When diazoxide is ineffective, a somatostatin analog is administered subcutaneously as second-line therapy [6–8]. Octreotide is a long-acting somatostatin analog that exerts inhibitory effects on insulin secretion via somatostatin receptors 2 and 5. Moreover, when medication is not effective, surgical pancreatic resection is performed. As a side effect of pancreatectomy, more of the pancreas may be removed than necessary and, thus, insulin secretion becomes insufficient, resulting in impaired glucose tolerance [9–11]. Focal lesions in the pancreas have recently been identified with ¹⁸F-DOPA PET, thereby allowing segmental resection of the region possible [12, 13]. A genetic diagnosis also facilitates the selection of treatment strategies. The presence of a paternally derived mutation in the K_ATP channel gene, such as ABCC8 or KCNJ11, or a somatic loss-of-function mutation in the paternal allele in fetal pancreatic β-cells is associated with the high probability of a causal focal lesion, for which pancreatic resection is selected as treatment [14, 15]. On the other hand, when CHI is caused by factors other than K_ATP channel genes, diffuse lesions of β cells may be present throughout the whole pancreas and, thus, medication is selected as the first-line treatment. These techniques reportedly decrease the risk of developing impaired glucose tolerance after pancreatic resection [3].

Precautions against the side effects of diazoxide, such as heart failure, fluid retention, and bone marrow suppression in the early stage of treatment, which need to be carefully monitored, have been reported [3, 6, 16–18]. On the other hand, there are currently no definite opinions regarding long-term management by diazoxide. Diazoxide is listed as a drug that may cause hyperglycemia because it suppresses insulin secretion. However, there have been no clinical reports clearly demonstrating that the use of this drug induces impaired glucose tolerance. We encountered a case of Kabuki syndrome with CHI that presented with impaired glucose tolerance after the long-term oral administration of diazoxide, in which insulin resistance had also worsened due to obesity.

Case report

The patient was a 9-year-old girl who had been diagnosed with Kabuki syndrome by a genetic test, which revealed a deletion mutation in the KMT2D gene. She also had a medical history of mild intellectual disability, premature thelarche, post-operative atrial septal defect, and left congenital deafness. There was no relevant family history, including diabetes. She was born at 37 weeks of gestation, without asphyxia, and with an Apgar score of 8 (1 min)/8 (5 min). Her length was 46.0 cm (− 0.85 SD) and weight was 2544 g (− 0.56 SD). Two hours after birth, severe hypoglycemia developed with a blood glucose level < 1 mg/dL and a high insulin level (5.2 μIU/mL). She was diagnosed with hyperinsulinemic hypoglycemia based on normal data derived from other endocrinological examinations and screening for inborn errors of metabolism. She was treated with glucose infusion and then started to be fed with milk 12 times/day. Since pre-prandial hypoglycemia persisted, diazoxide at 15 mg/day (5.0 mg/kg/day) was initiated on the 29th day of age. With a clinical suspicion of Kabuki syndrome, the patient was diagnosed with symptomatic CHI. Kabuki Syndrome was subsequently confirmed by a genetic test. The neonatal course of this case has been reported [19].

On the 94th day of age, the patient was transferred to our hospital for further treatment. Since frequent hypoglycemia still occurred despite the administration of diazoxide, the dose was increased. She was receiving 75 mg/day (10.3 mg/kg/day) at 1 year of age. Although her blood glucose level remained stable with the same dose for several years thereafter, she had frequent episodes of pre-prandial hypoglycemia from 5 years of age. Therefore, the dose of diazoxide was increased from 75 mg/day (3.02 mg/kg/day) to 100 mg/day (4.03 mg/kg/day) at 5 years of age, and from 100 mg/day (3.84 mg/kg/day) to 150 mg/day (5.76 mg/kg/day) at 6 years of age. During the increase in her medication and thereafter, there were no episodes of hypoglycemia.

The patient maintained a normal body weight until 7 years of age; however, her body weight increased and she became obese by 8 years of age. Her parents indicated that she ate more than her siblings. Therefore, weight gain was attributed to an increased appetite. Nutrition education was provided by a dietician; however, weight loss was not achieved. Although insulin levels were high in outpatient care, we did not evaluate blood glucose levels pre-prandially. At 9 years of age, a blood examination showed a casual blood glucose level of 99 mg/dL, an insulin level of 107.3 µIU/mL, and a HbA1c level of 5.6%. The high level of insulin was attributed to the insufficient suppression of insulin secretion due to the inadequate effects of diazoxide or to worsened insulin resistance as a result of obesity. Since difficulties were associated with identifying the underlying cause because her blood glucose levels were normal in outpatient care, she was admitted for a detailed evaluation of obesity and glucose metabolism at 9 years of age.

Her height was 134.7 cm (+ 0.01 SD), weight was 46.2 kg (BMI: 25.5 kg/m², 98.5 percentile for her age), and waist circumference was 86.0 cm. Blood pressure was 110/68 mmHg. She had some features unique to Kabuki syndrome, such as long palpebral fissures, large auricles, a low nasal tip, and short fingers. Puberty developed and Tanner stages were breast 2 and pubic hair 1.

A blood test showed a normal complete blood count. The results of a blood chemistry examination were all within normal ranges, except for a slightly elevated level of uric acid. Regarding glucose metabolism, the levels of HbA1c, blood glucose, and insulin after overnight fasting were 5.6%, 92 mg/dL, and 22.1 μIU/mL, respectively. The level of the anti-glutamic acid decarboxylase antibody was < 0.05 U/mL. An oral glucose tolerance test (OGTT) on diazoxide showed that her blood glucose level increased to 241 mg/dL at 120 min, homeostasis model assessment of insulin resistance (HOMA-R) was 5.18, and the insulinogenic index was 0.38 (reference range for her age, 1.10 ± 0.49) (Table 1A). We performed OGTT again before the administration of diazoxide in the morning and noted no marked differences (Table 1B). In addition, pre- and post-prandial blood glucose levels were frequently higher than 200 mg/dL, leading to a diagnosis of diabetes. Based on OGTT data, indicating a low insulinogenic index for her age, the initial secretion of insulin was considered to be suppressed by diazoxide. Therefore, we decreased the dose of diazoxide and finally discontinued it on the sixth day after drug tapering. A rapid improvement was subsequently observed in post-prandial hyperglycemia (Fig. 1).

Table 1.

Oral glucose tolerance tests

(A) On admission (performed after diazoxide medication)
Time (min)	0	30	60	120
Glucose (mg/dL)	95	192	258	241
IRI (μIU/mL)	22.1	59.2	67.0	178.6
HOMA-R: 5.18, HOMA-β: 248.6%, Insulinogenic index: 0.38
(B) On admission (performed before diazoxide medication)
Time (min)	0	30	60	120
Glucose (mg/dL)	86	178	218	196
IRI (μIU/mL)	18.0	51.9	66.2	111.3
HOMA-R: 3.82, HOMA-β: 281.7%, Insulinogenic index: 0.37
(C) 1 week after the discontinuation of oral diazoxide
Time (min)	0	30	60	120
Glucose (mg/dL)	72	160	203	195
IRI (μIU/mL)	15.8	81.7	85.2	134.4
HOMA-R: 2.80, HOMA-β: 632.0%, Insulinogenic index: 0.74
(D) After 2 months of diet and exercise therapy without diazoxide medication
Time (min)	0	30	60	120
Glucose (mg/dL)	74	162	172	166
IRI (μIU/mL)	11.4	80.0	81.7	116.6
HOMA-R: 2.08, HOMA-β: 373.0%, Insulinogenic index: 0.78

IRI immunoreactive insulin

Fig. 1.

Changes in blood glucose levels. The vertical axis shows blood glucose levels (mg/dL) and the horizontal axis indicates the number of days after hospitalization

One week after the discontinuation of diazoxide, a third OGTT showed a borderline blood glucose level (195 mg/dL) at 120 min; however, the insulinogenic index had increased to 0.74, indicating an improvement in early insulin secretion (Table 1C). During hospitalization, continuous glucose monitoring (CGM) was also conducted and showed that hypoglycemia was not observed during drug tapering or after its discontinuation under diet and exercise therapy for obesity. After a 2-month hospitalization, a fourth OGTT was performed, showing a blood glucose level of 166 mg/dL at 120 min, an insulinogenic index of 0.78, and HOMA-R of 2.08, which also indicated an improvement in insulin resistance (Table 1D). Her body weight was 42.3 kg (BMI: 23.3 kg/m², 96.7 percentile for her age) when she was discharged from hospital.

Discussion

Diazoxide binds to sulfonylurea receptor 1, which consists of the K_ATP channel, and keeps the channel open to inhibit insulin secretion. Although the mechanisms underlying CHI in Kabuki syndrome caused by the KMY2D gene mutation remains unknown, diazoxide was clearly effective for hypoglycemia in this patient from the neonatal period.

Adverse events of diazoxide in the early stage of oral administration, including heart failure and edema [6, 16–18, 20], have been noted. Therefore, monitoring after the initiation of the drug is recommended. While the short-term side effects of diazoxide are well known, there is no consensus on how to manage patients on long-term administration. Furthermore, there have been no case reports clearly demonstrating a causal relationship between diazoxide and impaired glucose tolerance. Only one case report described a patient with hyperglycemia in whom was both diazoxide and steroids were ineffective [21]. Moreover, in a review of drugs that may cause diabetes, diazoxide was described as a drug causing diabetes together with steroids, l-asparagine, and calcineurin inhibitors [22]. The present case suggests that diazoxide needs to be included as a drug that impairs glucose tolerance.

Regarding the mechanisms underlying diabetes in the present case, initial insulin secretion was impaired by diazoxide based on OGTT. Since hyperglycemia improved soon after the discontinuation of diazoxide, diazoxide appeared to have worsened impaired glucose tolerance. The exacerbation of obesity simultaneously increased insulin resistance. Children receiving long-term CHI therapy are more likely to overeat to prevent hypoglycemia, and this excessive calorie intake leads to obesity and a fatty liver [21]. We suggest stricter weight control for children receiving diazoxide medication for CHI. In contrast, after the discontinuation of diazoxide, continuous follow-ups are needed to detect the potential recurrence of hypoglycemia after the attenuation of obesity.

There are some issues in the management of patients with diazoxide that need to be improved. Although diazoxide may exacerbate glucose intolerance due to the suppression of initial insulin secretion, it is impossible to fully comprehend the conditions of glucose metabolism based on regular examinations at an outpatient clinic. Therefore, regular assessments of glucose metabolism by OGTT and CGM are recommended for children receiving long-term diazoxide medication for CHI.

There are a number of limitations that need to be addressed. The mechanisms by which CHI is caused by the KMT2D gene mutation, which encodes a protein that prevents proper histone modifications and causes genomic instability, remain unknown. However, this case report provides important information to health care providers who prescribe diazoxide, regardless of the pathogenesis and epidemiology of hyperinsulinism in cases of Kabuki syndrome caused by the KMT2D gene mutation. Furthermore, at the time of OGTT 1 week after the discontinuation of diazoxide, the patient had already been hospitalized and was receiving appropriate diet and exercise therapy. However, the amelioration of glucose intolerance was mainly attributed to the discontinuation of diazoxide because of the immediate attenuation of hyperglycemia after drug discontinuation and the improvement in the insulinogenic index based on OGTT performed 1 week after discontinuation, which indicated an increase in initial insulin secretion. A previous study suggested that Kabuki syndrome itself was a cause of diabetes or insulin resistance [23]. Although we cannot exclude this possibility, we attributed insulin resistance in the present case to overeating and being overweight.

Conclusion

The long-term prognosis of children treated with diazoxide remains unclear, and there is no consensus on long-term management. Based on the present case, diazoxide may aggravate glucose intolerance. In long-term management with diazoxide for hypoglycemia, OGTT and CGM need to be regularly performed to fully comprehend the conditions of glucose metabolism in patients. When CGM confirms no episodes of hypoglycemia, the discontinuation of diazoxide needs to be considered under careful monitoring.

Declarations

Conflict of interest

We declare no conflicts of interest.

Ethical approval

This article does not contain any studies with human or animal subjects performed by any of the authors.