Congenital hyperinsulinism is the most common cause of persistent hypoglycemia in neonates, infants, and children. Since the first case descriptions in the 1950s, the field has advanced significantly. It was the development of the insulin radioimmunoassay by Yalow and Berson a decade later that made it possible to demonstrate that this form of persistent hypoglycemia was caused by insulin, and a few years later, Drash described the successful treatment of children with hyperinsulinism with the antihypertensive diazoxide, which until today remains the only approved treatment for hyperinsulinism. In the mid 1970s, Baker and Stanley described that hyperinsulinism can be recognized by inappropriate responses of metabolic fuels and hormones during the course of a provocative fasting challenge. Later, advances in molecular genetics led to the discovery of the different genetic subtypes of hyperinsulinism. One of the most impactful discoveries in the field was the recognition of the focal form of hyperinsulinism and the development of 18F-DOPA PET for the localization of focal lesions before surgery which has resulted in the possibility of cure for children with focal disease. However, treatment options for children with nonfocal diazoxide-unresponsive hyperinsulinism have continued to be limited. New drug development programs for hyperinsulinism promise to change this in the next few years. Unfortunately, despite all these advances, children with hyperinsulinism around the world continue to experience neurological sequelae at high rates, highlighting the importance of early diagnosis and effective treatment.
Hypoglycemia was first revealed as a human disorder in 1922 when one of the initial patients treated with Banting and Best’s insulin started “climbing the walls” as a consequence of low plasma glucose . Shortly thereafter, endogenous hypoglycemia due to insulin was recognized in patients with insulinoma tumors of the pancreas that could be recognized by Whipple’s triad: symptoms of hypoglycemia, low plasma glucose level, and relief of symptoms following correction of hypoglycemia . This use of Whipple’s triad has been adopted to differentiate patients with true hypoglycemia and to determine who needs to be investigated for the etiology of the hypoglycemia and it is now incorporated into modern guidelines for the diagnosis and management of hypoglycemia .
It is now well known that dysregulated insulin secretion is the most common cause of persistent hypoglycemia in neonates, infants, and children, and our understanding of the pathophysiology of these disorders has advanced significantly, propelled by advances in molecular biology that facilitated the discovery of the genetic underpinnings of the congenital forms of hyperinsulinism (HI). The following historical overview briefly describes the discoveries and advances which have been made in diagnosis, genetics, and medical and surgical treatment of HI in infants and children.
The first clear description of children with hypoglycemia due to HI was provided by Dr. Irvine McQuarrie  in his 1954 Presidential Address to the American Pediatric Society. He felt it was essential to draw the attention of pediatricians to this group of infants because of their high risk of permanent brain injury or seizures. He correctly identified many of the major features of HI in infants including the high risk of “irremediable brain injury” due to “delays in diagnosis and ineffective, especially early, therapy” and that it might have a genetic origin. McQuarrie  named this new disorder, “idiopathic hypoglycemia of infancy,” since he thought insulin was unlikely to be the etiology as these infants did not have insulin-producing tumors in their pancreas. In this, McQuarrie  was wrong, although it took nearly two decades to recognize that these infants had genetic disorders of insulin secretion causing HI. Indeed, one of McQuarrie’s families has recently been shown to have a dominant disease-causing mutation leading to ectopic expression of hexokinase 1 in beta cells . While many of McQuarrie’s concepts about his new disorder have withstood the test of time, some have proven erroneous or have persisted as misconceptions, such as the belief that idiopathic hypoglycemia is a specific disease entity, that HI always spontaneously resolves by early childhood, and that glucocorticoids are effective treatment for neonatal hypoglycemia .
In 1956, based on studies of children seen in England and in Toronto, Cochrane described a familial form of hypoglycemia that is not prevented but instead is aggravated by high protein feeding (unlike so-called reactive hypoglycemia) . The hypoglycemic effect of protein can be reproduced by several of the individual amino acids, especially leucine, and was named “leucine-sensitive hypoglycemia,” the first indication that amino acids, as well as glucose, could be important insulin secretagogues. Leucine-sensitive hypoglycemia was considered a distinct entity, separate from other forms of idiopathic hypoglycemia, but later found to be a feature of some, although not all, types of HI. One of the first applications of Yalow and Berson’s  insulin radioimmunoassay was to demonstrate that hyperinsulinism was responsible for leucine-sensitive hypoglycemia. In 1964, Allan L. Drash  reported that diazoxide, an antihypertensive which inhibits insulin secretion, could control the blood glucose levels in some infants with leucine-sensitive hypoglycemia, providing further evidence that the disorder was due to HI.
Surgical management of patients who were unresponsive to medical therapy made it possible to examine the histological features of HI; in 1971, Lester Baker and Yakovac  described “nesidioblastosis” in infants with idiopathic hypoglycemia of infancy. This term was first coined by Laidlaw  in 1938 when he said that “in contrast with the concentration of excess islet tissue in a tumor (nesidoblastoma) there is some evidence pointing to a diffuse or disseminated proliferation of islet cells as a possible cause of hypoglycemia. Such a diffuse proliferation of nesidioblasts would be a nesidioblastosis” . It was proposed to be a distinct pathologic entity involving budding of beta-cell precursors from pancreatic ductal epithelium that was interpreted to represent persistence of the embryologic formation of islets . Later work by Jaffe et al.  and Rahier et al.  demonstrated that the histology of nesidioblastosis is not a specific entity but is rather the normal appearance of the pancreas during fetal development and in early infancy when islet development is still ongoing. Unfortunately, the term, nesidioblastosis, was picked up by adult endocrinologists who often continue to use it to refer to infants with congenital genetic forms of HI.
In 1975, several groups of investigators began to suggest that the term “idiopathic hypoglycemia” be discarded in favor of “congenital hyperinsulinism” as a more accurate and specific term [13-15]. Finally, in the 1990s, application of new genetic molecular biology technologies  opened the way to recognize more than a dozen genetic loci now known to be involved in HI.
The history of diagnosis in hyperinsulinism is closely linked to the story of the understanding of diabetes and the introduction of insulin therapy. Measurement of glucose in the blood first occurred in the 1880s, but it was not until 1915 that Lewis and Benedict  developed a method to measure blood glucose in small volumes of blood.
In 1910, Cobliner from Germany first reported hypoglycemia in children. In 1937, Hartmann and Jaudon  described hypoglycemia in infants. It was not until 1954 that McQuarrie’s  emphasis on the recognition of infants with persistent hypoglycemia led to a greater awareness of hypoglycemia. Also in 1954, Komrower  noted the changes in blood sugar that occurred after birth in normal babies and infants of diabetic mothers [IDM]. Marvin Cornblath  first described hypoglycemia in infants born to mothers with toxemia in 1959 and noted that many went on to have brain damage. This is likely the first description of what is now known as perinatal stress-induced HI.
Shortly following the discovery of insulin in 1921, James Collip injected insulin into a rabbit and noted it lowered blood glucose levels and caused the rabbit to go into a coma and die. In 1927, Wilder reported pancreatic cancer in a patient with symptoms of hypoglycemia and, following surgery by William Mayo, a large tumor was removed from the pancreas in addition to an orange-sized lesion from the liver. Mayo injected extracts of these tumors into rabbits and caused the blood glucose of the rabbit to fall precipitously. Subsequently, the hypoglycemia in adults with a single insulinoma was cured by surgery. Later, Alan Whipple  noted that there were many patients with symptoms of hypoglycemia and that most of them had no need for surgery. Thus, he developed Whipple’s triad and suggested that no surgery of the pancreas be performed unless the full triad was present.
Ernest Starling  coined the term “hormone” in 1905. Subsequently, De Meyer  and Sharpey-Schafer  suggested the name “insuline/insulin” for an internal secretion of the pancreas that controlled glucose metabolism.
In 1968, Isles and Farguaha  noted that 6 infants of diabetic mothers (IDM) who were noninsulin-dependent had lower glucose levels (mean 22 ± 8 mg/dL vs. 44 ± 20 mg/dL) compared to infants of nondiabetic mothers. In addition, the IDM had mean insulin levels of 34 μU/mL at the time of hypoglycemia indicating that the hypoglycemia of babies born to mothers with diabetes during the pregnancy was due to high insulin levels.
By the 1970s, it became clear that idiopathic hypoglycemia of infancy was caused by excessive secretion of insulin following a series of papers by Baker et al. , Morey Haymond , and Anthony Pagliara . The criteria for the clinical diagnosis of hyperinsulinism were described by Baker and Charles A. Stanley  based on descriptions by Cahill  and others of the metabolic changes that occur during normal fasting adaptation. Stanley and Baker  showed that patients with HI can be recognized by inappropriate responses of metabolic fuels and hormones during the course of a provocative fasting challenge in 1976. The response to glucagon at a time of hypoglycemia as demonstrated by David Feingold and Stanley  in 1980 was the final part of the diagnostic triad for hyperinsulinism: (1) inappropriately elevated plasma insulin concentration at a time of hypoglycemia, (2) evidence of insulin effect with suppressed free fatty acid and ketone concentrations in the critical sample, and (3) a glycemic response to glucagon. This triad became the standard of care for diagnosis of HI .
In the 1990s, the race to discover the genetic etiology of HI was on. In 1991, Paul Thornton et al.  suggested that HI was transmitted as an autosomal recessive genetic condition and in 1994 suggested it could also be inherited in an autosomal dominant manner [30, 31]. Shortly thereafter, Glaser et al.  found linkage to chromosome 11p14-15 and 1 year later the gene encoding the SUR1 protein was discovered by Aguilar-Bryan et al.  in this location. Subsequently, Thomas [33, 34] described the first patients with HI to have a genetic mutation in 1995 in what later became known as the ABCC8 gene and 1 year later in Kir6.2 (later known as KCNJ11). During 1997–1998, Stanley played a pivotal role in determining the genetic etiology for Cochrane’s leucine-sensitive hypoglycemia, also known as the hyperinsulinism hyperammonemia syndrome [35, 36], and glucokinase hyperinsulinism . In addition to the discovery of novel genes, a novel genetic mechanism encompassing paternal inheritance of a recessive mutation in ABCC8 or KCNJ11 and somatic loss of the maternal 11p15 region was identified as the genetic explanation for focal disease of the pancreas [38, 39]. The list of genetic loci and syndromes associated with HI continues to grow and currently, rapid turnaround genetic testing can discover the genetic diagnosis of diazoxide-unresponsive HI within 4–7 days .
The first breakthrough in the medical treatment of HI was in 1964 when Drash  described the use of diazoxide for treatment of children with “idiopathic hypoglycemia” and “leucine-sensitive hypoglycemia.” Diazoxide is an antihypertensive used for acute treatment of hypertensive crisis but not for long-term hypertension treatment because of its propensity to induce diabetes. Diazoxide suppresses insulin secretion by opening the beta-cell KATP channels . In 1976, diazoxide was approved in the USA by the Food and Drug Administration (FDA) for HI in children. To this date, diazoxide continues to be the only drug with regulatory approval for the treatment of HI and the first line of therapy for this condition ; however, up to 60% of children with HI do not respond to therapy with diazoxide.
In 1977, initial reports described the use of infusions of native somatostatin to raise plasma glucose levels in infants with HI . Subsequent reports from Israel by Glaser and from Philadelphia by Thornton [44, 45] described the short- and long-term use of octreotide, an intermediate-acting somatostatin analog, in the treatment of HI to avoid the need for pancreatectomy. Approximately 20 years later, Le Quan Sang  and Dalit Modan-Moses  reported the successful use of long-acting somatostatin analogs, octreotide LAR and lanreotide, respectively, for the treatment of HI, which allowed for the simplification of the treatment regimens with monthly dosing rather than multiple daily doses or continuous subcutaneous infusion. In the largest series to date, Diva D. De Leon and colleagues  reported significant improvements on glycemic control and fasting tolerance in children treated with lanreotide. While somatostatin analogs continue to be commonly used as second-line treatment of HI, this indication has not been approved by the FDA (USA).
Multiple new therapies for HI are under development and promise to make possible a personalized approach to treatment of children with HI and to improve their long-term outcomes. Programs currently in clinical trials include a peptide antagonist of the GLP-1 receptor, a short-acting soluble glucagon analog, a long-acting glucagon analog, a selective nonpeptide somatostatin receptor 5 agonist, and an allosteric inhibitor of the insulin receptor.
A surgical approach to managing disorders of endocrine hormone oversecretion was considered a great triumph of modern surgery in the first half of the twentieth century . Following the successful treatment of hyperthyroidism by surgical removal of the thyroid gland, pancreatectomy was established as a treatment approach for protracted hypoglycemia suspected to be due to pancreatic hypersecretion. However, it was clear even from those early days that a cure was only achieved when a tumor was present and completely removed (as summarized by McCaughan and colleagues ).
The recognition that persistent hypoglycemia due to hypersecretion of insulin could occur at any age led to the use of pancreatectomy as a way of managing persistent hypoglycemia in children, even before the diagnosis of HI could be established based on the metabolic response to hypoglycemia [50-52]. As in adults, some children were cured after resection of an adenoma, but for some, only amelioration but not complete resolution of the hypoglycemia was achieved after resection of 80–90% of pancreas [51-53]. By the mid 1960s, pancreatectomy was considered for children with persistent hypoglycemia despite dietary measurements and corticosteroid therapy . The introduction of diazoxide in 1964  held then the promise of further reducing the number of children with HI that would require pancreatectomy.
In 1978, Gauderer and colleagues  from the Children’s Hospital of Philadelphia outlined the indications for surgery for children with HI emphasizing that surgery should be the treatment of choice for children with onset after 1 year of age in whom a resectable islet cell adenoma would be more likely. For children under 1 year of age, the experience at the time was that diffuse disease was more common, and therefore, a trial of diazoxide should be undertaken before considering surgery. The approach recommended by the Philadelphia team was to do a thorough intraoperative exploration of the pancreas with histological examination of frozen biopsies. If a lesion was not found, a subtotal pancreatectomy was performed. The high frequency of significant residual hypoglycemia in neonates with severe HI led to the recommendation of a near-total pancreatectomy as the procedure of choice for these infants [56-59]. This overall approach continues to be the preferred approach for diazoxide-unresponsive HI; however, the development of molecular and imaging diagnostic tools has made the assessment for the possibility of a curable form of HI more precise.
In 1984, Jacque Rahier  described the basic structural lesion of the pancreas in patients with diffuse hyperinsulinism and reported the finding of focal lesions in 3 patients. He and his co-workers later described recognizable differences between focal and diffuse forms of HI. Their methods lead to the development of a new surgical strategy pioneered by Claire Fékété in Paris by which focal HI could be cured by partial resection of the pancreas, based on transhepatic portal venous sampling (THPVS) . A description of pre-operative localization of focal HI lesions using selective pancreatic arterial calcium stimulation with hepatic vein insulin sampling was proposed by Stanley and colleagues  as an alternative to THPVS described earlier by Brunelle et al. . The interventional radiology techniques of arterial calcium stimulation with hepatic vein insulin sampling and THPVS offered modest help in targeted resection of focal HI lesions. Later, a landmark paper published by Timo Otonkoski  described the use of 6-fluoro-(18F)-L-3,4-dihydroxyphenylalanine (18F-FDOPA) positron emission tomography for lesion localization in focal HI. This imaging technique was subsequently demonstrated to lead to cure of focal hyperinsulinism with the first large series published by Olga Hardy and colleagues .
Enucleation of focal lesions sparing the normal pancreas continues to be the treatment of choice for focal hyperinsulinism with a high cure rate in the most experienced centers , while near-total pancreatectomy is reserved for the most severe diffuse cases that fail to respond to intensive medical therapy. In the largest single-center series published today, Adzick and colleagues  reported surgical outcomes of 500 patients evaluated at the Children’s Hospital of Philadelphia. The cure rate for focal HI, accounting for 49% of patients, was 97% and most of these children required less than 50% pancreatectomy. Near-total pancreatectomy for diffuse HI, which accounted for 40% of all patients, resulted in adequate glycemic control in 31% of cases, persistent hypoglycemia in 49%, and persistent hyperglycemia in 20% at the time of hospital discharge.
While much has changed regarding the diagnosis and management of HI, unfortunately, neurodevelopmental sequelae continue to be very common. In 1954, McQuarrie  described irreparable brain damage from severe hypoglycemia in children with what was then known as idiopathic hypoglycemia and since then, published series from across the world have reported rates of neurodevelopmental differences in 23–48% of children affected by HI [59, 66-70]. Importantly, children with transient and curable forms of HI are equally affected, even though the exposure to hypoglycemia resolves within the first year of life [71, 72].
The field of hyperinsulinism has seen significant progress in the last 25 years thanks to advances in molecular biology contributing to an improved understanding of the genetic and phenotypic heterogeneity of this condition. More recent progress in therapeutics has started to pave the way toward a personalized approach to managing HI and improving long-term outcomes.
Conflict of Interest Statement
Disclosures: Paul S. Thornton performed industry-sponsored research for Zealand, Ascendis, Pfizer, Novo Nordisk, OPKO Spruce, and Rezolute. Paul S. Thornton acted as an unpaid consultant for Zealand, Rezolute, Crinetics, Eiger, and Spruce and received Honorarium from the Human Growth foundation. Charles A. Stanley serves in Data Safety Monitoring Boards for Jansen Pharmaceuticals and Zealand Pharma. Diva D. De Leon has received research funding from Zealand Pharma, Tiburio Therapeutics, Twist Bioscience, and Crinetics Pharmaceuticals. Diva D. De Leon receives funding support by National Institutes of Health grants R01-DK056268 and R01-DK098517. These funders played no role in the completion of this manuscript. Diva D. De Leon has received consulting fees from Zealand Pharma, Crinetics Pharmaceuticals, Hanmi Pharmaceutical, Eiger Biopharmaceuticals, Poxel Pharma, and Heptares Therapeutics. Charles A. Stanley and Diva D. De Leon are named inventors in patents # USA Patent Number 9,616,108, 2017, USA Patent Number 9,821,031, 2017, Europe Patent Number EP 2120994, 2018, and Europe Patent Number EP2818181, 2019.
There was no external funding for this project.
Paul S. Thornton, Charles A. Stanley, and Diva D. De Leon reviewed the literature and co-wrote and edited the manuscript.
Data Availability Statement
There were no data generated for this report.