Persistent hyperinsulinemic hypoglycemia of infancy (PHHI) (Online Mendelian Inheritance in Man [OMIM] 256450), formerly known as nesidioblastosis, is a glucose metabolism disorder characterized by profound hypoglycemia and inappropriate secretion of insulin (1). Affected children run the risk of severe neurological damage unless immediate and adequate steps are taken (2). Treatment with diazoxide and/or somatostatin analogue is the first line of therapy. However, it not always effective, especially in familial cases, which may necessitate an alternative intervention such as pancreatectomy (3).
Several studies have suggested that partial pancreatectomy endangers future islet cell function (4,5). The incidence of diabetes increases with age and correlates with the extent of surgical resection (6,7). However, there was no report of occurrence of overt diabetes in medically treated patients (8). In this report, we described an adolescent female with neonatal nesidioblastosis who developed diabetes after medical treatment with diazoxide/octreotide. To our knowledge, this is the first nesidioblasosis case subject who developed diabetes following medical therapy.
A 14-year-old Saudi female presented with severe persistent hypoglycemia during the first few days of life. She was diagnosed with hyperinsulinemic hypoglycemia of infancy based on her intravenous glucose requirement of >14 mg · kg−1 · min−1, an insulin-to-glucose ratio of >0.3 (her insulin level was 98 μU/ml at a serum glucose of 32 mg/dl), negative urinary ketones, a 30-min glucose increment of ≥30 mg/dl in response to intramuscular 0.5 mg glucagon, and normal blood spot acylcarnitine profile determined by tandem mass spectrometry. She also had a normal growth hormone level of >20 mU/l and a normal cortisol level of >500 nmol/l during hypoglycemia. She was treated initially with frequent feeding supplemented with complex carbohydrates (polycose/corn starch) and then started on diazoxide 15 mg · kg−1 · day−1 divided three times a day, which kept her euglycemic with occasional hypoglycemic episodes. In 1992, octreotide was first introduced in our hospital as an adjunctive therapy to diazoxide. She was started on 25 μg · kg−1 · day−1 of subcutaneous octreotide divided four times a day. She responded to medical treatment with no hypoglycemic episodes. She was continued on diazoxide and octreatide until the age of 10 years, when she became euglycemic and these two medications were stopped. At the age of 14, she developed hyperglycemia associated with weight gain. Her blood glucose ranged from 200 to 300 mg/dl, and her weight was 75 kg (>95%). She had an insulin level of 10 μU/ml and C-peptide level of 0.16 nmol/l at a serum glucose level of 350 mg/dl. Antiglutamic acid decarboxylase, insulin, and islet cell antibodies were negative. She responded to metformin 250 mg twice a day with a serum mean glucose level of 109 mg/dl and HbA1c of 7.5%.
The long-term outcome of PHHI is not well documented. Previous reports suggested that subtotal or near total pancreatectomy in infants will endanger the future islet function (4–8). Long-term follow-up studies in medically treated patients with diazoxide or octreotide showed that some of these patients responded to medical therapy and became euglycemic (9–11). Some patients were weaned off medical therapy and continued to be euglycemic; however, none of them became hyperglycemic or diabetic. Leibowitz et al. (8) followed six conservatively treated patients with PHHI. Intravenous glucose tolerance was performed in all patients and showed a blunted insulin response in two with no overt hyperglycemia. Histologically, Kassem et al. (12) showed that β-cell proliferation and apoptosis, which normally occurrs in the normal developing human pancreas, also occurs in the PHHI pancreas with a higher frequency of apoptosis. They suggested that this phenomenon will result in a slow, progressive, and complete loss of β-cell mass. This histological report and the development of diabetes in our nonpancreatectomized PHHI patient may suggest that patients with PHHI will naturally develop diabetes whether they were treated medically or surgically or even if they are left untreated. This hypothesis was further raised when the natural history of this disease was discussed in knockout mouse models. Transgenic mice engineered to express a dominant- negative form of Kir6.2 or mice with ATP-sensitive K+ channel deficiency developed hyperinsulinemic hypoglycemia followed by hypoinsulinemic hyperglycemia. Diabetes in these transgenic mice was thought to be due to sustained unregulated Ca influx and premature β-cell apoptosis (burn-out phenomenon) (13,14). Seino et al. (15,16) reported another possible predisposing factor to hyperglycemia in PHHI patients. They showed that hyperglycemia in Kir6.2 knockout mice was more evident with age and increasing weight. They suggested that the Kir6.2 knockout mouse provides a model of type 2 diabetes, and that both the genetic defect in glucose-induced insulin secretion and the acquired insulin resistance due to environmental factors are necessary to develop diabetes in the Kir6.2 knockout mouse.
We hypothesized that diabetes was induced by weight gain and obesity in our patient. She responded to metformin, which may suggest that her diabetes is due to insulin resistance induced by both weight gain and insulin insufficiency. Simple type 2 diabetes is still a possibility, although there was no history of diabetes in the family. This patient could be the human example of the Kir6.2 knockout mouse model. We recommend, based on this human clinical evidence, weight control in aged PHHI patients to decease the incidence of diabetes.