PEDIATRICS Vol. 107 No. 3 March 2001, pp. 476-479

Neurologic Outcomes of 90 Neonates and Infants With Persistent Hyperinsulinemic Hypoglycemia

Francesca Menni, MD^*, Pascale de Lonlay, MD^*, Caroline Sevin, MD^*, Guy Touati, MD^*, Catherine Peigné, MPH^*, Valérie Barbier, MPH^*, Claire Nihoul-Fékété, MD, Jean-Marie Saudubray, MD^*, and Jean-Jacques Robert, MD, PhD^*

From the ^* Fédération de Pédiatrie, Diabète-Métabolisme-Neurologie and  Clinique Chirurgicale Infantile, Centre Hospitalier Universitaire Necker-Enfants Malades, Paris, France.

	ABSTRACT

Top Abstract Results Discussion References

Objective.  To evaluate the neurologic outcomes of neonates and infants suffering from persistent hyperinsulinemic hypoglycemia of infancy (PHHI).

Methods.  The neurologic development of 90 PHHI patients was studied retrospectively. Sixty-three patients were treated surgically and 27 were treated medically. Fifty-four patients were neonates, of whom 8 were treated medically and 46 were operated on (19 for a focal adenomatous hyperplasia and 27 for diffuse hyperinsulinism). Thirty-six patients had infancy-onset hyperinsulinism, of whom 19 were treated medically and 17 underwent pancreatectomy (10 patients for a focal adenomatous hyperplasia and 7 for diffuse hyperinsulinism).

Results.  Severe psychomotor retardation was found in 7 patients, 6 with neonatal-onset PHHI. Intermediate psychomotor disability existed in 12 patients; epilepsy existed in 16. Neonatal-onset was the main risk factor for severe retardation or epilepsy. Medically treated patients were less severely affected than those treated by surgery, and there was no difference between the diffuse and focal forms of hyperinsulinism.

Conclusion.  Neonatal hyperinsulinemic hypoglycemia is still a severe disease with an important risk to rapidly develop severe mental retardation and epilepsy.  Key words:  persistent hyperinsulinemic hypoglycemia of infancy, psychomotor retardation, epilepsy, pancreatectomy.

Persistent hyperinsulinism is the most common cause of recurrent hypoglycemia in neonates and infants.^1-3 The inappropriate oversecretion of insulin is responsible for profound hypoglycemia that can cause irreversible brain damage.^4,5 Seizures are the revealing symptom in approximately half of the cases of persistent hyperinsulinemic hypoglycemia of infancy (PHHI), independent of age, whereas hypoglycemia is rarely discovered by a routine assay of blood glucose. Hypoglycemia of neonatal onset seems to be more severe than infancy-onset hypoglycemia. Most cases of PHHI do not respond to diazoxide treatment and need to be treated by pancreatectomy, whereas over half the infancy-onset patients can be managed with medical treatment.⁶ Neurologic sequelae have been reported in patients with PHHI,^7-17 but some studies have clearly shown that abnormal development of the central nervous system is more common and more severe when symptoms occur in neonates rather than in infants.¹⁶

Our understanding of the pathophysiology of PHHI has greatly improved in the last 4 years. Genetic mutations have been described on several genes, particularly SUR1 and Kir6.2.^18,19 In addition, the differentiation between focal and diffuse lesions of the pancreatic -cells have been clearly demonstrated. The therapeutic outcome depends critically on distinguishing between the 2 types of hyperinsulinism. Patients with diffuse hyperinsulinism who are resistant to medical treatment undergo near-total pancreatectomy, with a high risk of later diabetes mellitus. Patients with focal adenomatous hyperplasia can be treated by partial pancreatectomy.³

We have monitored the neurologic development of 90 patients with PHHI, taking into account their age at the onset of symptoms of hypoglycemia, the histologic type of the pancreatic lesion, and the medical or surgical management of the disease.

	PATIENTS AND METHODS

We studied the patients who were treated at our hospital for hyperinsulinism between 1982 and 1998. We have not included the patients with transient hyperinsulinism or those with associated syndromes: hyperammonemia, carbohydrate deficiency glycoprotein syndrome, genetic syndromes, and Munchausen syndrome by proxy. The diagnostic criteria for hyperinsulinism were: fasting and postprandial hypoglycemia (mean initial plasma glucose concentration ± standard deviation: 1 ± 0.2 mmol/L) with concomitant hyperinsulinemia (>3 mU/L); in neonates, high rates of intravenous glucose (>10 mg/kg/min) infusion required to maintain the blood glucose >3 mmol/L; a positive response to glucagon administered subcutaneously or intramuscularly (increase in blood glucose by 2-3 mmol/L after 0.5 mg glucagon); for the neonatal forms, persistent hypoglycemia throughout the first month of life.

The patients were initially treated in local pediatric units and were referred to our hospital at a mean age of 1.27 ± 2.1 years (3 days to 12.6 years). Our treatment was designed to maintain the patient's blood glucose at 3 to 6 mmol/L, using continuous drip-feeding and intravenous glucose. Neonates were given glucose through a central venous catheter because of the high rates of infusion. Glucose infusion was often associated with glucagon, given by continuous intravenous infusion at 1 to 2 mg per day. Once the blood glucose had been controlled, patients were given oral diazoxide at 15 mg/kg per day divided into 3 doses. Patients were considered to be diazoxide-resistant when they still had at least 1 plasma glucose value below 3 mmol/L when fed normally. A transhepatic selective pancreatic venous catheterization was performed under general anesthetic to locate the insulin hypersecretion.²⁰ Those patients who were believed to have focal hyperinsulinism underwent surgery. The others also underwent surgery if they resisted or could not tolerate diazoxide treatment. The extent of pancreatic resection was decided on after intraoperative histologic examination.²¹ Focal hyperinsulinism was treated by limited pancreatic resection; diffuse hyperinsulinism required near-total pancreatectomy. The clinical outcome of the operated neonatal-onset patients has been described previously.³ The patients who were sensitive to diazoxide were fed normally, without any intravenous glucose or glucagon.

The neurologic development of the patients with hyperinsulinism was studied retrospectively. The patients' files were judged exploitable for 90 patients, 63 of whom had been treated surgically and 27 treated medically. Of the 90, 54 were neonates, 8 treated medically and 46 operated on (19 for a focal adenomatous hyperplasia and 27 for diffuse hyperinsulinism). The other 36 had infancy-onset hyperinsulinism, 19 were treated medically and 17 were pancreatectomized (10 for a focal adenomatous hyperplasia and 7 for diffuse hyperinsulinism [Table 1]).

The psychomotor outcome was assessed retrospectively, considering the following parameters: 1) intellectual performance (developmental quotient, psychomotor evaluation); 2) academic achievement; 3) psychological, psychomotor, or orthophonic support; 4) behavioral disorders (aggressiveness, attention deficit disorder); 5) neurologic disorders (tremor, myoclonus, pyramidal syndrome, trunk tonus, not walking until after 18 months old). Patients were assigned to 1 of 3 groups: group 1, normal children: development quotient [DQ]/IQ >80 or only 1 minor disability (1 failure at school, psychological, psychomotor or orthophonic support, behavioral disorder, minor neurologic symptom); group 2, children with intermediate disability: DQ <60/IQ <80, 2 failures at school or in a special class, or 2 or more of the above intellectual or motor disorders; group 3, children with major retardation: DQ/IQ <60, major intellectual or motor disability, attending a special school, major neurologic impairment. Head circumference (HC) was followed with the Nellhaus chart. Microcephaly was defined as HC 2 standard deviations. Patients were evaluated during the year of surgery or at the time of investigation (n = 90), and after 3 years (n = 59), 6 years (n = 38), and 10 years (n = 25). The psychomotor outcome was estimated using life-table analysis. Fisher's exact test and the ² test were used when appropriate.

	RESULTS

Top Abstract Results Discussion References

Life-table analysis indicated psychomotor retardation in 26% of the patients, 8% with major retardation (group 3) and 18% with an intermediate disability (group 2). Major retardation was more frequent in neonatal-onset patients than in those with PHHI revealed in infancy (Table 1). It was also more frequent in patients who had undergone surgery than in those who were treated medically. The small number of patients in some subgroups meant that the differences were not statistically significant. There was no difference between the patients who were operated on for focal or diffuse hyperinsulinism. The patients with intermediate disability showed no difference with age at onset of symptoms, medical or surgical treatment and type of histologic lesion (Table 1).

Table 2 shows selected clinical features of patients with severe and intermediate mental retardation. Six of the 7 patients with severe retardation had symptoms of hypoglycemia in the first few hours of life. Three neonates had repeated seizures (patients 2, 3, and 4) and 3 had generalized hypotonia (patients 1, 5, and 6). The plasma glucose concentrations were low in all cases, before the onset of symptoms in 3 patients (patients 3, 5, and 6). Four of these 6 patients were rapidly treated in neonatal or intensive care units. Hyperinsulinism was not recognized immediately in the other 2 patients (patients 2 and 6) and they suffered hypoglycemic seizures at 5 weeks (patient 6) and status epilepticus at 9 months (patient 2) of age. The seventh patient (patient 7) had repeated brief episodes of loss of consciousness at 5 months of age, after an apparently symptom-free early life. All 7 patients were judged neurologically abnormal with hypotonia and poor contact, on admission in our unit, before investigations and/or surgery. These patients have now been followed-up for 15 months to 15 years. Three started walking at between 2 and 4 years of age. The other 4 patients cannot sit without support, at 15 to 24 months of age. Four suffer from epilepsy and 4 have microcephaly (Table 2). One patient has blindness and 1 has strabismus. Three of these 7 patients had focal adenomatous hyperplasia, 3 had diffuse hyperinsulinism, and 1 was treated medically.

Among the 12 patients with intermediate disability, 7 were found to be hyperinsulinemic in the first 2 days of life; 4 because of seizures (1 status epilepticus), 2 because of neurologic symptoms and/or cyanosis, and 1 by a routine assay of blood glucose. The other patients developed the first symptoms at 1.5 to 9 months of age; 4 had seizures and 1 lost consciousness. Six patients had neurologic abnormalities when first admitted to our unit. The patients are now 4.6 to 14 years old. None of them began to walk late. Seven have epilepsy; 4 microcephaly, 2 deafness, and 1 strabismus (Table 2). Four of these patients were treated medically and 8 underwent surgery; 4 for focal adenomatous hyperplasia and 4 for diffuse hyperinsulinism.

The patients in groups 2 and 3 did not differ from each other, or from the patients in group 1, in terms of year of birth (Table 2), gestational age, weight and HC at birth, or in the frequency of seizures as first symptoms. The mean age (± standard deviation) of the patients is now 8.5 ± 4.6 years, 11.6 ± 4.2 years, and 8.9 ± 6.0 years, in groups 1, 2 and 3, respectively (not significant). The frequency of acute fetal distress tended to increase from group 1 (4%) to group 2 (9%) and group 3 (16%), but this was not significantly different. Seizures were the first symptom in 55% of patients of group 1 and 66% of those of group 2. As shown in Table 2, seizures were not the first symptom in most patients of group 3. However, 4 patients (patients 1, 2, 3, and 4) presented seizures shortly after diagnosis. The severity and duration of these episodes could not be precisely evaluated in this retrospective study. The first neurologic examination in our unit was abnormal in 22% of group 1 patients, 50% of those in group 2, and 100% of those in group 3. The frequency of hypoglycemic relapses after surgery, either clinical or subclinical (blood glucose monitoring), did not differ significantly between the patients operated on in the 3 groups. It even tended to be less frequent in group 3 (17%) than in group 1 (45%) and group 2 (37%).

Epilepsy developed in 16 patients (Table 3). Thirteen had neonatal onset and 3 had infancy-onset hyperinsulinism. This represents 24% of the neonatal-onset patients and 8% of the infancy-onset patients, but the difference was not statistically significant (P = .09, Fisher's Exact test). Three of these patients were treated medically (11% of medical patients) and 13 underwent pancreatectomy (21% of surgical patients), but the difference was not significant (P = .24, Fisher's Exact test). The frequencies of epilepsy in patients with focal (24%) and diffuse (17%) lesions were the same. Epilepsy occurred more frequently in mentally retarded patients, with the same frequency in groups 2 and 3, than in nonretarded patients (Table 4). The percentage of patients with microcephaly also tended to increase from group 1 to 3 (Table 4).

	DISCUSSION

Top Abstract Results Discussion References

Hyperinsulinism in infancy is responsible for recurrent very profound hypoglycemia that is often <1 mmol/L. These episodes are especially dangerous, as there is no alternative fuel available for the brain (no lactate and no ketone bodies).²² As glucose is an important energy source and the precursor of the macromolecular constituents required for the rapid growth of the brain during the neonatal period, it is essential for cerebral functioning.^4,5 Psychomotor retardation, learning disability, seizures, and diverse neurologic sequelae have been reported in children with hyperinsulinism.^7-17 There is little information available on the factors associated with these complications of hypoglycemia. However, they seem to be more frequent when the hypoglycemia is of neonatal onset,¹⁶ when diagnosis is delayed,¹⁴ or when the child requires surgical treatment.⁹ Our results confirm that hyperinsulinemic hypoglycemia is a serious pediatric disease, as it is frequently associated with cerebral sequelae. Severe mental retardation was found in 8% of our patients, less severe disability in 18% of them, 6% of the patients with normal mental development also had epilepsy, and 8% had microcephaly. Our study included a fairly large number of patients for a rare disease, but statistical analyses did not show significant correlations with the occurrence of developmental problems, probably because of the very small number of patients within subgroups. However, our results strongly suggest that severe psychomotor retardation and epilepsy is more common in patients with neonatal hypoglycemia than in those with infancy-onset hypoglycemia. They also suggest that patients treated medically were less severely affected than those treated surgically. A delayed diagnosis was probably deleterious in specific cases, such as patients 2 and 6, whose hypoglycemia was recognized in the first day of life. They became asymptomatic and went back home, but they suffered dramatic seizures some weeks later. Conversely, there seemed to be no difference between the diffuse and focal forms of hyperinsulinism.

These observations confirm that focal and diffuse forms of hyperinsulinism are clinically indistinguishable, although this difference is of great importance for the surgical indications. They reemphasize the greater severity of the neonatal forms of hyperinsulinism, which are also resistant to diazoxide treatment and extremely difficult to manage medically. It is particularly striking that hypoglycemia was diagnosed in 4 of our severely retarded patients very soon after birth, and they were treated in specialized neonatal or intensive care units. This shows how very difficult it is to manage this condition adequately. It is also striking that the first neurologic examination in our unit was abnormal in 100% of the patients who are now severely retarded, that no specific event at follow-up could be related to the mental status, and that the evaluation of these patients did not vary over the years. This suggests that early events are responsible for the development of mental retardation and indicates the need for better understanding of the mechanisms of mental retardation in hyperinsulinemic hypoglycemia to establish clear recommendations for the management of this severe condition in the neonates.²³

	FOOTNOTES

Received for publication Jan 20, 2000; accepted Jul 24, 2000.

Reprint requests to (J.J.R.) Diabète de l'Enfant et de l'Adolescent, Hôpital Necker-Enfants Malades, 149 rue de Sèvres, 75743 Paris Cedex 15, France. E-mail: jean-jacques.robert{at}nck.ap-hop-paris.fr

	ABBREVIATIONS

PHHI, persistent hyperinsulinemic hypoglycemia of infancy; DQ, development quotient; HC, head circumference.

	REFERENCES

Top Abstract Results Discussion References

1. Stanley CA Hyperinsulinism in infants and children. Pediatr Clin North Am 1997; 44:363-374 [CrossRef][Medline]
2. Bruining GJ Recent advances in hyperinsulinism and pathogenesis of diabetes mellitus. Curr Opin Pediatr 1990; 2:758-765
3. De Lonlay P, Poggi-Travert F, Fournet JC, Clinical features of 52 neonates with hyperinsulinism. N Engl J Med 1999; 340:1169-1175 [Abstract/Free Full Text]
4. Koh THHG, Aynsley-Green A, Tarbit M, Eyre JA Neural dysfunction during hypoglycaemia. Arch Dis Child 1998; 63:1353-1358 [Abstract]
5. Lucas A, Morley R, Cole TJ Adverse neurodevelopmental outcome of moderate neonatal hypoglycaemia. Br Med J 1988; 297:1304-1308
6. Touati G, Poggi-Travert F, Ogier de Baulny H, Long-term treatment of persistent hyperinsulinaemic hypoglycaemia of infancy with diazoxide: a retrospective review of 77 cases and analysis of efficacy-predicting criteria. Eur J Pediatr 1998; 157:628-633 [CrossRef][Medline]
7. Thomas CG, Underwood LE, Carney CN, Dolcourt JL, Whitt JJ Neonatal and infantile hypoglycemia due to insulin excess: new aspects of diagnosis and surgical management. Ann Surg 1977; 185:505-517 [Medline]
8. Landau H, Perlman M, Meyer S, Persistent neonatal hypoglycemia due to hyperinsulinism: medical aspects. Pediatrics 1982; 70:440-446 [Abstract/Free Full Text]
9. Thomas CG, Cuenca RE, Azizkhan RG, Underwood LE, Carney CN Changing concepts of islet cell dysplasia in neonatal and infantile hyperinsulinism. World J Surg 1988; 12:598-609 [CrossRef][Medline]
10. Spitz L, Bhargava RK, Grant DB, Leonard JV Surgical treatment of hyperinsulinaemic hypoglycaemia in infancy and childhood. Arch Dis Child 1992; 67:201-205 [Abstract]
11. Glaser B, Hirsch HJ, Landau H Persistent hyperinsulinemic hypoglycemia of infancy: long-term octreotide treatment without pancreatectomy. J Pediatr 1993; 123:644-650 [CrossRef][Medline]
12. Soliman AT, Alsalmi I, Darwish A, Asfour MG Growth and endocrine function after near total pancreatectomy for hyperinsulinaemic hypoglycaemia. Arch Dis Child 1996; 74:379-385 [Abstract]
13. Shilyansky J, Fisher S, Cutz E, Perlman K, Filler RM Is 95% pancreatectomy the procedure of choice for treatment of persistent hyperinsulinemic hypoglycemia of the neonate? J Pediatr Surg 1997; 32:342-346 [CrossRef][Medline]
14. Meissner T, Brune W, Mayatepek E Persistent hyperinsulinaemic hypoglycaemia of infancy: therapy, clinical outcome and mutational analysis. Eur J Pediatr 1997; 156:754-757 [CrossRef][Medline]
15. Cade A, Walters M, Puntis JWL, Arthur RJ, Stringer MD Pancreatic exocrine and endocrine function after pancreatectomy for persistent hyperinsulinaemic hypoglycaemia of infancy. Arch Dis Child 1998; 79:435-439 [Abstract/Free Full Text]
16. Beinbreech B, Meissner T, Schilke K, Persistent hyperinsulinemic hypoglycemia of infancy (PHHI)-Clinical findings, therapeutic interventions and outcome in 80 patients [abstract]. J Inher Metab Dis 1998; 21:130
17. Cresto JC, Abdenur JP, Bergada I, Martino R Long term follow up of persistent hyperinsulinaemic hypoglycaemia of infancy. Arch Dis Child 1998; 79:440-444 [Abstract/Free Full Text]
18. Thomas PM, Cote GJ, Wohllk N, Mutations in the sulfonylurea receptor gene in familial persistent hyperinsulinemic hypoglycemia of infancy. Science 1995; 268:426-429 [Abstract/Free Full Text]
19. Thomas P, Ye Y, Lightner E Mutation of the pancreatic islet inward rectifier Kir6.2 also leads to familial persistent hyperinsulinemic hypoglycemia of infancy. Hum Mol Genet 1996; 5:1809-1812 [Abstract/Free Full Text]
20. Brunelle F, Negree V, Barth MO, Pancreatic venous samplings in infants and children with primary hyperinsulinism. Pediatr Radiol 1989; 19:100-103 [CrossRef][Medline]
21. Rahier J, Sempoux C, Fournet JC, Partial or near-total pancreatectomy for persistent neonatal hyperinsulinaemic hypoglycaemia: the pathologist's role. Histopathology 1998; 32:15-19 [CrossRef][Medline]
22. Evans M, Amiel SA Carbohydrates as a cerebral metabolic fuel. J Pediatr Endocrinol Metab 1998; 11:99-102
23. Baker L, Thornton PS, Stanley CA Management of hyperinsulinism in infants. J Pediatr 1991; 119:755-757 [CrossRef][Medline]