Challenging diagnosis of congenital hyperinsulinism in two infants of diabetic mothers with rare pathogenic KCNJ11 and HNF4A gene variants

Congenital hyperinsulinism (CHI) is the leading cause of persistent hypoglycemia in infants, causing neurodevelopmental and cognitive delays in up to 25–50% of affected children [1]. Early recognition and appropriate treatment are important to prevent such adverse outcomes.

The estimated incidence of CHI is reported to be as low as 1/50,000 live births in Holland to as high as 1/2500 in geographic areas where consanguinity is highly frequent, such as Saudi Arabia [2]. CHI is caused by genetic defects that cause disease through changes in channels expression/activity, transcription factors expression and disturbed enzyme activities. More frequently is increased activity rather than enzymes deficiency All these defects cause the common finding of recurrent episodes of hyperinsulinemic hypoglycemia due to an inappropriate secretion of insulin by the pancreatic beta-cells [3].

In 2016, there were 11 genes known to be associated with monogenic forms of CHI (ABCC8, KCNJ11, GLUD1, GCK, HADH1, UCP2, MCT1, HNF4A, HNF1A, HK1, and PGM1) [4, 5]. Variants in the ABCC8 and KCNJ11 genes together comprise approximately 90% of the identified cases of diazoxide-unresponsive CHI [6]. These genes encode SUR-1 and Kir6.2, respectively, which comprise two subunits of K_ATP-HI, the beta-cell plasma dependent membrane ATP-potassium channel [7, 8]. Most pathogenic variants in these 2 genes are recessive and cause loss of function; however, some cases of dominantly inherited inactivating mutations have been reported. In the past, clinical phenotype of patients with dominant variants was different from those with recessive variants but the biochemical phenotype was not different [9]. However, in 2013 Snider et al. published a large cohort of 417 children with CHI whose genotype-phenotype correlations were most successful if they have GLUD1, GCK, and recessive K_ATP mutations [6]. Many of these correlations were challenging due to the high frequency of novel missense K_ATP mutations because these defects might be either recessive or dominant and if dominant these could be either responsive or unresponsive to diazoxide. In these cases, mutation analysis including parental testing should be considered. However, the existence of novel K_ATPmissense mutations variability (recessive/dominant) in adult carriers is typically unknown because they are frequently asymptomatic unless challenged with fasting or protein tolerance test [6, 9, 10].

Diazoxide, the first-line treatment for CHI, acts by opening the K_ATP-channel and decreasing insulin secretion [11]. The effect of the gene variant on K_ATP-channel expression determines the clinical phenotype and the response to diazoxide [4]. Octreotide, a second-line agent, is a long-acting somatostatin analog that decreases insulin secretion through hyperpolarization of the beta-cells and inhibition of calcium channels but has been associated with treatment failure due to tachyphylaxis and fatal necrotizing enterocolitis [8].

There are two distinct histologic forms of CHI: focal and diffuse. In a published cohort of 52 infants with CHI treated from 1985 to 1998, twenty-two had focal CHI and the rest had diffuse CHI. Only one neonate with focal HI and two with diffuse HI responded to treatment with diazoxide. All the other infants require pancreatic surgery to achieve cure [12].

Variants in a third gene, HNF4A, are associated with transient CHI in early infancy, despite also causing maturity-onset diabetes of the young (MODY 1) [13]. HNF4A variants are typically rare and the incidence varies per population but have been described as the third most common cause of diazoxide-responsive CHI in a large cohort of 300 patients with CHI from United Kingdom [14, 15].

The infants of diabetic mothers (IDM) born with macrosomia typically present frequent metabolic complications including hypoglycemia, jaundice, respiratory distress, asphyxia and are in greater need of intensive care as reported by Looreda-Garcia et al. in a large cohort of 18,005 newborns where 10% were infants of diabetic mothers [16]. The hyperinsulinemic hypoglycemia seen in IDMs occurs secondary to transient hyperinsulinism but it does not last longer than 2 or 3 days [7].

The current 2011 American Academy of Pediatrics guidelines for postnatal glucose homeostasis in healthy late-preterm and term infants mandate that persistent hypoglycemia, lower than 25 mg/dL from birth to 4 h of age, or lower than 35 mg/dL from 4 to 24 h of age after attempts to reefed, requires treatment with intravenous (IV) glucose [17]. These prior AAP guidelines focused on when to screen infants for hypoglycemia and when to provide treatment but did not address the etiology work up procedure for prolonged and recurrent neonatal hypoglycemia. The Pediatric Endocrine Society recommendations published in 2015 considered the interpretation and response to plasma glucose (PG) concentration during the first 2 days of life is controversial however the mean PG concentration in normal newborns increases with time and by 72 h of age is similar to those in older infants and children. Therefore, these guidelines suggest delaying diagnostic evaluations until 2–3 days after birth [17].

Here we report 2 cases of IDMs with rare genetic variants of CHI whose final diagnoses of CHI were complicated and delayed because the patients were IDMs and had concomitant challenging medical conditions.

In these 2 cases, the diagnostic workup for CHI was complicated by their status as infants of diabetic mothers. For both, the standard treatment for neonatal hypoglycemia was provided but the specific work up for CHI was not performed until both infants were older than 3 weeks of life. Case 1 had perinatal asphyxia, prematurity, and sepsis as his underlying diagnoses and therefore his hypoglycemia was initially considered to be secondary to these underlying diagnoses rather than due to CHI, and Case 2 was presumed to have transient hyperinsulinism due to his small for gestational age condition, in addition to being an IDM with concomitant PHI.

The incidence of CHI in IDMs is not well known considering the absence of published reports in the literature. However, considering IDMs have higher rates of prolonged hospitalization at the NICU services is remarkably important to bring these findings to the attention of pediatricians and neonatologists to increase awareness about the need of considering CHI as the main etiology of prolonged and recurrent hypoglycemia in all neonates including the infants of diabetic mothers.

The most common risks for neonates developing abnormal postnatal glucose homeostasis include being SGA, LGA, IDMs, history of perinatal asphyxia and prematurity [17]. Case 1 was born LGA, IDM, premature and had asphyxia at birth, leading perinatal stress-induced hyperinsulinism (HI) to be included on his differential diagnosis. Stress-induced HI can be seen in asphyxia, as well as maternal toxemia and intrauterine growth retardation [22]. This former medical condition can persist for several weeks after birth, but it usually remits spontaneously. Perinatal asphyxia in Case 1 was probably one of the confounding factors contributing to the delay in considering CHI as the most likely etiology of his prolonged and recurrent hypoglycemia. Case 2 was also an IDM, premature but born SGA and these three risk factors played an important role in the delay of considering CHI as the final etiology of his persistent neonatal hypoglycemia.

Pertinent workup for the etiology of persistent neonatal hypoglycemia includes obtaining critical samples in a timely manner and performing genetic testing. Critical samples for evaluation of hypoglycemia include plasma glucose, beta-hydroxybutyrate, carbon dioxide, lactate, free fatty acids, cortisol, growth hormone, total/free carnitine, acyl carnitine profile, C-peptide, and pro-insulin levels. In 2015, the Pediatric Endocrine Society (PES) of the United States published new recommendations for evaluation and management of persistent hypoglycemia in neonates, infants, and children and the specific timeframes to start performing work up for CHI were also described [23]. Adherence to the current PES recommendations in all NICUs and nurseries would potentially avoid delays on neonatal hypoglycemia etiology work up and this potentially would lead to a prompt diagnosis of CHI. Prior AAP guidelines published in 2011 failed to address the diagnosis and management of disorders causing recurrent and prolonged hypoglycemia beyond the neonatal period [17].

DNA sequencing is remarkably important when CHI is suspected because it contributes to accurate medical treatment, genetic counseling, potential distinction between focal and diffuse disease, determination of the likelihood of good response to medications, and consideration of early pancreatic surgery. In a cohort study from the United Kingdom of infants who underwent rapid sequencing, 27 patients (26%) were positive for ABCC8 variants and 5 of them (5%) were positive for KCNJ11 variants [24]. Of the 5 with KCNJ11-variants, 2 underwent subtotal pancreatectomy. A significant proportion (18%) who tested positive for ABCC8 or KCNJ11 variants achieved euglycemia with medical treatment. All infants with maternal heterozygous mutations achieved remission, leading to the conclusion that maternally inherited variants may confer a higher rate of remission.

Genetic testing is also very helpful in CHI because even though patients with inactivating mutations of GCK (MODY 2), inactivating mutations of hepatocyte nuclear factorHNF1A (MODY 3) and HNF4A (MODY 1) can present as congenital HI in neonates they can later evolve in early adulthood into insulinopenic diabetes; and activating mutations of the KATP channel subunits can cause either severe permanent neonatal diabetes or a mild increased susceptibility to adult-onset diabetes [5].Case 1 was heterozygous for a maternally inherited likely pathogenic variant in KCNJ11. This variant has been reported in one family where it segregated with disease in four symptomatic hypoglycemic individuals [19]. It is rare (present in only 2 of ~ 12,000 alleles) in the general population and predicted by in silico tools to be damaging. However, this is the first time this variant has been reported in an IDM.

The phenotype of Case 2 (SGA) did not follow the classical phenotype of IDMs who are typically born with macrosomia. Most infants reported with HNF4A variants had increased birth weight and macrosomia [25], and they also had transient CHI requiring diazoxide therapy between 3 months and 8 years [26]. Case 2 was born SGA in contrast to the prior reports of HNF4A-associated CHI infants (E.g. p.R63W mutation) who are typically born with macrosomia [26]. However, one recent report of an infant with HNF4A-CHI and carrier of the p.R63W mutation described that infant was born with normal weight but this infant was not an IDM [27]. It is not clear if the SGA phenotype of Case 2 was secondary to his underlying HNF4A variant, his prematurity (considering that insulin is an important growth factor in the third trimester of pregnancy) or if this was a consequence of the intrauterine growth retardation associated with his maternal PIH.

With regard to his response to medical treatment, diazoxide therapy in Case 2 was rapidly tapered and successfully discontinued similar to the prior HNF4A case reported by Flanagan in 2010 [14].

It is important to describe this transient hyperinsulinism clinical course on Case 2 because his successful response to diazoxide could be associated to his R267C gene variant, and could be also linked to future progression to MODY. Even though we recommended genetic testing for both parents, and mother was referred to genetics for testing and counseling (father was not available and his family medical history was unknown) she was not able to pursue genetic testing due to unavailable medical insurance coverage. We discussed with mother the potential implications of genetic testing for her diagnosis (she has high risk to be a HNF4A carrier) but she reported to be healthy and refused further evaluation with OBGYN.

This case series is one of the few reports describing two IDMs with persistent recurrent hypoglycemia secondary to genetic CHI, and an atypical clinical phenotype for HNF4A-associated CHI. Prompt consideration of CHI as the most likely etiology of recurrent and prolonged neonatal hypoglycemia even in IDMs is important to facilitate prompt consultation with endocrinology, and appropriate timely workup including genetic testing as recommended by the current PES clinical guidelines [23]. Transient neonatal hypoglycemia in an IDM is not expected to last more than 2–3 days. In all the cases of neonatal hypoglycemia persisting beyond this timeframe, including infants of diabetic mothers, CHI should be considered as the presumed etiology.