Increased referrals for congenital hyperinsulinism genetic testing in children with trisomy 21 reflects the high burden of non‐genetic risk factors in this group

Abstract Background Hyperinsulinism results from inappropriate insulin secretion during hypoglycaemia. Down syndrome is causally linked to a number of endocrine disorders including Type 1 diabetes and neonatal diabetes. We noted a high number of individuals with Down syndrome referred for hyperinsulinism genetic testing, and therefore aimed to investigate whether the prevalence of Down syndrome was increased in our hyperinsulinism cohort compared to the population. Methods We identified individuals with Down syndrome referred for hyperinsulinism genetic testing to the Exeter Genomics Laboratory between 2008 and 2020. We sequenced the known hyperinsulinism genes in all individuals and investigated their clinical features. Results We identified 11 individuals with Down syndrome in a cohort of 2011 patients referred for genetic testing for hyperinsulinism. This represents an increased prevalence compared to the population (2.5/2011 expected vs. 11/2011 observed, p = 6.8 × 10−5). A pathogenic ABCC8 mutation was identified in one of the 11 individuals. Of the remaining 10 individuals, five had non‐genetic risk factors for hyperinsulinism resulting from the Down syndrome phenotype: intrauterine growth restriction, prematurity, gastric/oesophageal surgery, and asparaginase treatment for leukaemia. For five individuals no risk factors for hypoglycaemia were reported although two of these individuals had transient hyperinsulinism and one was lost to follow‐up. Conclusions Down syndrome is more common in patients with hyperinsulinism than in the population. This is likely due to an increased burden of non‐genetic risk factors resulting from the Down syndrome phenotype. Down syndrome should not preclude genetic testing as coincidental monogenic hyperinsulinism and Down syndrome is possible.

Results: We identified 11 individuals with Down syndrome in a cohort of 2011 patients referred for genetic testing for hyperinsulinism. This represents an increased prevalence compared to the population (2.5/2011 expected vs. 11/2011 observed, p = 6.8 Â 10 À5 ). A pathogenic ABCC8 mutation was identified in one of the 11 individuals. Of the remaining 10 individuals, five had non-genetic risk factors for hyperinsulinism resulting from the Down syndrome phenotype: intrauterine growth restriction, prematurity, gastric/oesophageal surgery, and asparaginase treatment for leukaemia. For five individuals no risk factors for hypoglycaemia were reported although two of these individuals had transient hyperinsulinism and one was lost to follow-up.
Conclusions: Down syndrome is more common in patients with hyperinsulinism than in the population. This is likely due to an increased burden of non-genetic risk factors resulting from the Down syndrome phenotype. Down syndrome should not preclude genetic testing as coincidental monogenic hyperinsulinism and Down syndrome is possible.

| INTRODUCTION
Hyperinsulinism (HI) is a disorder of the pancreatic beta-cell where inappropriately high levels of insulin are secreted leading to hypoglycaemia. Prolonged neonatal HI can be transient, often remitting within 6 months, with risk factors including male sex, low birth weight, and perinatal stress. 1 In contrast, persistent HI is likely to be genetic with disease-causing mutations in single genes identified in 50%-70% of cases. 2,3 HI has also been reported as a rare feature in patients with aneuploidies. For example, HI can present in females with Turner syndrome resulting from a complete or partial monosomy of the X chromosome and in children with Patau syndrome resulting from mosaic trisomy 13. 4,5 The most common aneuploidy is trisomy 21, causing Down syndrome, which affects 1 in 794 live births in the USA. 6 Down syndrome is characterized by intellectual disability, microcephaly, congenital heart defects, gastrointestinal disorders, and endocrine disorders which include Type 1 diabetes or neonatal diabetes. 7-9 Whilst HI has not been reported as a feature of Down syndrome, we noted a high number of individuals with the co-existence of these two conditions being referred to our laboratory for genetic testing. Our aim was to assess whether the prevalence of children with HI and Down syndrome was higher than expected in our cohort and if so to determine the reason(s) for this.

| METHODS
We studied 2011 individuals referred for HI genetic testing to the Exeter Genomics Laboratory between 2008 and 2020. Clinical information was provided at referral using a standardized request form.
Follow-up data by case note review were requested for all individuals with HI and Down syndrome.
We performed targeted next-generation sequencing of 13 known HI genes including ABCC8, CACNA1D, CDKN1C, GCK, GLUD1, HADH, HNF1A, HNF4A, INSR, KCNJ11, PMM2, SLC16A1, and TRMT10A in all individuals with HI and Down syndrome using previously described methods. 10 We used Stata/SE v16.0 to perform a one-sample binomial test to assess if the prevalence of Down syndrome in our cohort was significantly higher than the population prevalence (Stata Corp, College Station, TX, USA).
Informed consent was obtained from the parents or guardians of all probands. This study was approved by the North Wales Research Ethics Committee (517/WA/0327).

| RESULTS
Within our international cohort of 2011 individuals, we identified 11 cases with Down syndrome (n = 11/2011 [0.55%]). This represents a minimal prevalence as we do not routinely screen for aneuploidies, and some clinicians may not have provided this information on the genetic request form. The number of children with Down syndrome was significantly higher than expected by chance given the population prevalence of Down syndrome of 12.6/10,000 6 (2.5/2011 expected vs. 11/2011 observed, p = 6.8 Â 10 À5 ).
We identified a mutation in a known HI gene in 1/11 (9%) patients. This individual had a pathogenic paternally inherited ABCC8 mutation. 11 Of the 10 individuals without a mutation in a known gene, two were born with intrauterine growth retardation (IUGR) (birth weight Z-score < À2). The median age at diagnosis at HI of the 10  Patient 5 previously reported in. 13 ** indicates that mutation was previously reported in. 11 information was not available (n = 1). One individual with persistent HI demonstrated side-effects to diazoxide and did not respond to octreotide, necessitating a near-total pancreatectomy. 12 Consanguinity was reported in this individual.
Seven individuals, including the child with an ABCC8 mutation, had undergone gastric or oesophageal surgery for duodenal atresia, duodenal stenosis, tracheomalacia, or gastro-oesophageal reflux disease (GORD). In two cases surgery had been performed prior to the onset of HI. One of these cases had also undergone surgery to repair a portosystemic shunt. 13 Table 1.

| DISCUSSION
We identified 11 individuals with HI and Down syndrome. Given that Down syndrome has an approximate incidence of one in 794 live births, we would have expected two or three individuals with Down syndrome in our cohort of 2011 individuals. 6 The statistically significant enrichment and higher prevalence therefore suggest that the two conditions are related.
The prevalence of mutations in the known genes was low in the Down syndrome and HI cohort (n = 1/11, 9%) although this increased to 20% in those with confirmed persistent HI (n = 1/5). This pick-up rate is lower than anticipated given previous studies have reported mutations in the known genes in 50%-70% of HI cases. 2,3 While this may reflect the small sample size, it is also possible that the Down syndrome is increasing the risk of the child developing HI.
We identified risk factors for developing HI in five of the 10 individuals without a mutation in a known gene. Two children had surgery to correct a gastrointestinal (GI) disorder prior to the onset of HI (Table 1). GI disorders are common in individuals with Down syndrome and surgical management of this can lead to iatrogenic hypoglycaemia as a result of dumping syndrome. 14,15 Furthermore, one of these individuals had confirmed post-prandial hypoglycaemia following surgery lending further support to this diagnosis. 16 This patient also had a portosystemic shunt, with surgical closure resulting in a resolution of the hypoglycaemia. 13 In four further cases, gastric surgery was performed but this occurred after the onset of HI in three cases suggesting that the HI was unlikely to be due to gastric surgery induced post-prandial hypoglycaemia. The age at gastric surgery in the remaining patient was unknown.
IUGR or biochemical evidence of perinatal and postnatal stress associated with prematurity, was reported in two individuals. These are well-recognized risk factors for prolonged neonatal hypoglycaemia. 1 IUGR was reported in a second individual however the HI was ongoing at the age of 13 years suggesting it was not causative of the hypoglycaemia. 17 One individual had been diagnosed with acute lymphoblastic leukaemia that had been treated with an L-asparaginase based chemo-

CONFLICT OF INTEREST
The authors have nothing to disclose.