Hyperinsulinaemic hypoglycaemia (HH) is a major cause of recurrent and persistent hypoglycaemia in infancy and childhood.¹ Rapid diagnosis, avoidance of recurrent and repeated episodes of hypoglycaemia and prompt management of the hypoglycaemia are vital in preventing brain damage and mental retardation.² Unfortunately, a large proportion of children with HH still develop brain damage as a consequence of delayed diagnosis and subsequent management. HH can be either congenital or secondary to certain risk factors (such as intrauterine growth retardation). Congenital hyperinsulinism involves either defects in the genes ABCC8 and KCNJ11 (encoding for the two proteins SUR1 and KIR6.2 of the pancreatic β cell K_ATP channel, respectively) or abnormalities in the enzymes glucokinase, glutamate dehydrogenase and short chain acyl-CoA dehydrogenase (SCHAD). Loss of function mutations in the genes ABCC8 and KCNJ11 cause the most severe forms of HH which are usually medically unresponsive.

HH is also observed in newborns with intrauterine growth retardation, in infants with perinatal asphyxia, in infants of diabetic (gestational and insulin dependent) mothers, in some infants with Beckwith-Weidemann syndrome and, more recently, in infants with no predisposing factors.^3–6 In most of these conditions, the HH is a transient phenomenon and resolves spontaneously. However, some small for gestational age and appropriate for gestational age infants can have prolonged hyperinsulinaemic hypoglycaemia which requires treatment with diazoxide, persists for several months and then resolves spontaneously.^7,8 Recognition of this group of patients is important as they are fully responsive to treatment with oral diazoxide. The underlying mechanisms of the HH in these patient groups are unclear at present.

BIOCHEMICAL BASIS OF HYPERINSULINAEMIC HYPOGLYCAEMIA

The biochemical basis of HH (congenital and secondary) involves dysregulated insulin secretion with defects in glucose counter-regulatory hormones.^9–11 It is also possible that there is increased insulin sensitivity in some of these patients (for example, those with intrauterine growth retardation), although this has not been proven directly. The unregulated insulin secretion drives glucose into the insulin sensitive tissues, especially skeletal muscle, adipose tissue and liver, causing profound hypoglycaemia. This is compounded by the fact that insulin simultaneously inhibits glycogenolysis (glycogen breakdown), gluconeogenesis (glucose production from non-carbohydrate sources), lipolysis and ketogenesis (hypoketotic). The normal physiological glucagon and cortisol counter-regulatory hormonal responses to hypoglycaemia are blunted further, exacerbating the hypoglycaemia.^10,11 This biochemical milieu is a recipe for depriving the brain of its most important fuel, namely glucose.

The associated brain glucopaenia is accompanied by a lack of alternative substrates such as ketone bodies and lactate. It is under these conditions that the risk of brain damage is highest. The importance of ketone bodies as an alternative substrate for brain utilisation in patients with HH was illustrated in a study by Plecko et al.¹² Two patients with congenital hyperinsulinism were administered oral beta-hydroxybutyrate and using magnetic resonance spectroscopy Plecko et al were able to show an increased uptake in the brain of beta-hydroxybutyrate with a concomitant reduction in the glucose infusion rate. This uptake of ketone bodies by the brain was proportional to the circulating concentration of beta-hydroxybutyrate.

Recent advances in understanding brain glucose sensing, the role of intrahypothalamic insulin and K_ATP channels in glucose regulation have further highlighted the potential complex central mechanisms that may contribute to the increased risk of hypoglycaemia in patients with HH. Pancreatic β cell K_ATP channels play a pivotal role in regulating insulin secretion by transducing metabolic signals to electrical changes in membrane potential. These K_ATP channels are also present in the hypothalamus and are similar to those in the pancreas.¹³ Animal studies have shown that hypoglycaemia is sensed in several areas of the brain, especially in the ventromedial hypothalamus (VMH)¹⁴ in which K_ATP channels play an important role in mediating responses to hyper- and hypoglycaemia. Closing of K_ATP channels in the VMH (much like the β cell) impairs defence mechanisms against glucose deprivation and therefore could contribute to defects in glucose counter-regulation. In rats intrahypothalamic infusions of insulin can suppress hepatic glucose production by inhibiting gluconeogenesis.¹⁵ This action of insulin is mediated by hypothalamic K_ATP channels via the vagus nerve. As patients with HH have raised serum insulin levels (and presumed raised CSF insulin levels as insulin can cross the blood–brain barrier), it is conceivable that hepatic glucose production is suppressed not just by the peripheral actions of insulin but also by the central actions of insulin. Hence several potentially defective mechanisms (peripheral and central) of glucose homeostasis may be contributing to the hypoglycaemia in patients with HH.

Practical implications for managing patients with HH

Confusion and controversy surrounds the definition and management of hypoglycaemia in infancy and especially in the newborn period. However, regardless of the definition used, it is recognised that hypoglycaemia can cause neonatal encephalopathy resulting in permanent brain injury.¹⁶ Even transient HH secondary to intrauterine growth retardation and maternal diabetes mellitus can lead to neonatal seizures and occipital brain injury.¹⁷ A blood glucose concentration of 2.6 mmol/l was suggested by Koh et al¹⁸ as the definition of hypoglycaemia based on the neurophysiological changes associated with hypoglycaemia. It is important to remember that these studies were performed in infants who were not hyperinsulinaemic and thus were able to generate alternative fuels such as ketone bodies in response to hypoglycaemia. Cornblath et al¹⁹ have recently suggested using “operational” thresholds (blood glucose levels at which clinical interventions should be considered) in different groups of individuals. For most patients it difficult to define a blood glucose level that will require intervention (especially in neonates) since there is uncertainty over the level and duration of hypoglycaemia that can cause neurological damage. Many centres accept a blood glucose level of 2.6 mmol/l as the operational threshold for neonates and infants with HH.

Patients with HH are clearly very different and the definition of an operational threshold of hypoglycaemia should be re-evaluated in this group of patients. Given the biochemical basis (hypoketotic and hypofattyacidaemic) of the hypoglycaemia, maintaining blood glucose concentrations below the normal range (3.5–6 mmol/l) will lead to brain glucopaenia. As discussed above, this brain glucopaenia will be compounded by the fact that there are no alternative substrates (ketone and lactate) for the brain to use. In these patients the brain is constantly dependent on a normal circulating blood glucose concentration as the oxidative fuel for neuronal function. There are no studies in this group of patients to suggest that the brain is able to compensate for the glucopaenia either by reducing brain glucose utilisation, increasing cerebral blood flow, increasing the fractional extraction of glucose from the circulation or using glycogen stored in the astrocytes. On the contrary, as insulin is able to cross the blood–brain barrier, it is conceivable that the breakdown of stored glycogen in astrocytes will be inhibited. Insulin stimulates glucose uptake and enhances glycogen accumulation in neonatal rat brain astrocytes preventing glucose release.²⁰

Table 1 shows the total serum ketone bodies (acetoacetate and beta-hydroxybutyrate) and lactate concentrations in seven patients with different forms of HH. These serum ketone bodies were measured at three different blood glucose concentrations (<3 mmol/l, 3–5 mmol/l and >5 mmol/l) in all seven patients. The different blood glucose concentrations were maintained by increasing or decreasing the intravenous glucose infusion rate. As is shown, serum insulin concentrations are detectable at all three blood glucose concentrations in each patient and serum ketone bodies are concurrently suppressed at all time points. The serum lactate level remained within the normal range in all patients. These simple observations reinforce the fact that glucose is the only substrate available for brain utilisation in patients with HH.

Hence these principles have important implications for managing patients with HH. It is therefore recommended that in patients with suspected HH blood glucose concentrations should be maintained within the normal range (3.5–6 mmol/l) during routine management. This is to avoid the neurological consequences of neuroglycopaenia and its long-term complications. Neonatologists and paediatricians should have a high index of suspicion of diagnosing HH in any patient who has persistent hypoglycaemia.

Given the complex management issues with these patients, in the UK and in some other European countries the development of designated regional referral centres is being considered. Referring paediatricians and neonatologists will be able to seek advice about diagnosis, treatment and transfer of patients with HH to regional centres. The increased awareness, early diagnosis and transfer of patients with severe forms of hyperinsulinaemic hypoglycaemia to dedicated centres will not only improve the quality of care delivered to these patients but more importantly the long-term neurological outcome.