Neonatal Hypo-Ketotic Hypoglycemia Secondary to Transient Hyperinsulinism: Diazoxide Responsiveness and Experience With Fasting Test After Treatment Withdrawal

Abstract

Introduction:: Transient neonatal hyperinsulinism (TNH) is frequently reported in neonates with stress factors (intrauterine growth restriction (IUGR), large for gestational age (LGA), perinatal asphyxia, infants of diabetic mother, etc.). Early recognition and treatment are prioritary to avoid neurological morbidity. Objective: Clinical, molecular characterization and treatment response in neonates with hypoglycemia due to transient hyperinsulinism admitted to a tertiary hospital Neonatal Unit from January 2015 to August 2020. Materials and Methods: Prospective cohort study. Newborns older than 7 days of age, with diagnostic criteria of hyperinsulinism: non ketotic hypoglycemia with detectable insulin, low free fatty acids, glucose infusion rate > 10mg/kg/min, and positive response to glucagon test, were recruited. Results: Out of 5374 patients admitted, 46 (0.85%) presented hypoglycemia secondary to TNH (57% males and 43% females). 78% were delivered by Cesarean section, 59% were European, 17% Latino-Americans, 11% Asians, 9% Africans, and 4% Arabs. 78% were preterm newborns (median 33 weeks gestational age), 70% had birth weights or heights <-1.6 SDS (medians: -1.8 SDS and -2 SDS, respectively). Median age at diagnosis was 22 days (IQE 10–29 days), and feeding was exclusively enteral. Median blood glucose at diagnosis was 37mg/dl (IQE 31-44mg/dl), median insulin: 3mu/ml, median ketonemia: 0.2mmol/L, GH: 15 ng/ml, Cortisol: 16 ug/dl and AAL: 75mg/dl. 90% received diazoxide (dose ranged between 5-10mg/kg/day), presenting as most prevalent side effects hypertrichosis (80%) and edema (13%). Diazoxide median treatment duration was 83 days (IQE 41–110). Response was positive in 100%, with fasting tests response yielding a glycemia > 60mg / dl after 10 hours of fasting post treatment withdrawal. Molecular analysis was carried out with help of a custom NGS panel (MonDIAB.V3; 385 genes) in 80% of the patients. No mutations were identified in known genes implicated in the etiology of congenital hyperinsulinism (ABCC8, KCNJ11, HNF4A, GLUD1, HADH, SLC16A1, GCK, UCP2, HNF1A, AKT2, INSR, CACNA1D), however, predicted deleterious variants were found in other candidate genes such as G6PC2, TH, PMM2, and APPL1, implicated in insulin secretion or glycemic homeostasis. Conclusions: TNH is a prevalent entity to be considered in neonates with risk factors. In our series, TNH is also present in term newborns (22% of patients) and in newborns with weight and/or height appropriate for gestational age (30%). Treatment with diazoxide at low doses is effective in the resolution of these hypoglycemias. The fasting test could be useful for a safe treatment withdrawal when resolution is suspected. No monogenic cause explaining the TNH was identified. Most of the cases molecularly examined presented with 2 or more predicted deleterious variants, suggesting a multifactorial genetic component.