Symptomatic hypoglycaemia during ketosis in children, either induced by fasting or a ketotic diet, has been recognised since the 1920s.^1,2 Ketotic hypoglycaemia or idiopathic hypoglycaemia (KH) is the most common form of hypoglycaemia beyond infancy.³ However, despite numerous efforts to elucidate the underlying pathophysiology, its aetiology remains unclear.^4–7 Typically, children with KH present between 18 months and 7 years of age with recurrent episodes of hypoglycaemia and ketonuria during an intercurrent illness. Seizures may occur, but neurological sequelae are rare. Hormonal counter-regulation remains intact, and diagnostic work-up, including work-up for defects of ketolysis, does not yield any significant abnormalities. Hypoglycaemia responds promptly to oral or intravenous glucose administration. KH improves with age and is rarely observed after puberty.

Ketone bodies are derived from oxidation of fatty acids or certain amino acids (leucine, lysine, and tryptophan) in the liver and transported to other organs such as the brain, heart, skeletal muscle, adrenal cortex, and kidneys.⁸ During ketosis the two main ketone bodies, acetoacetate and 3-hydroxybutyrate, are present in increased concentrations physiologically following fasting, after prolonged exercise, and during consumption of a high fat diet.^9–11 Moderate ketosis may also be present during pregnancy, the neonatal period, and infancy as part of physiological adaptation.¹¹ In addition, ketosis may be found in association with decompensation in diabetes mellitus, corticosteroid or growth hormone deficiencies, intoxication with alcohol or salicylates, rare inborn errors of ketolysis, maple syrup urine disease, and organic acidopathies.^7,8,13–15

Earlier studies of KH suggested that there was a functional defect of hepatic glucose production via gluconeogenesis or a lack of gluconeogenic substrates, including alanine.^6,7 As part of the counter-regulatory mechanism, glucagon causes a modest increase in blood glucose levels during hypoglycaemia and an additional decline in plasma alanine levels. Hypoalaninemia in KH, which may be secondary, is only restored by administration of alanine or prevented by administration of cortisone acetate.⁶ In addition, some patients with KH are known to have increased concentrations of branched chain amino acids during hypoglycaemia, possibly secondary to hypoinsulinism.

KH may be regarded as a disorder of glucose homoeostasis in which glucose utilisation exceeds glucose production. In this context ketosis may represent a physiological response to hypoglycaemia rather than a pathological event.^4,5 However, in addition, ketones do not appear to be utilised normally, particularly by the brain. Data on glucose and leucine kinetics are, to our knowledge, not yet available. Consequently we devised a study in a group of children with KH following a fast, using stable isotopes to measure glucose and leucine kinetics. In addition, basal metabolic rate (BMR) and respiratory quotient (RQ) were measured; all results were compared to another group of children having a diagnostic fast. The hypothesis is that there is a failure of both ketone utilisation and of the increase in glucose production during illness and fasting

METHODS

Subjects

Nine children with KH were studied. All had a past medical history of at least two episodes of symptomatic documented hypoglycaemia following a period of fasting or during intercurrent illness. During hypoglycaemia, ketonuria of 3+ on dipstick was documented; hypoglycaemia resolved rapidly and completely with glucose administration. In addition, all children with KH had had an unremarkable diagnostic work-up in the past that included urinary organic acid, plasma acylcarnitine, intermediate metabolites, and hormone analysis, without identifiable pathology.

Table 1 summarises age, gender, and anthropometric data of the children. There was no statistically significant difference in age, height, weight, or body mass index (BMI) between those with KH and a group of 11 children who were referred to Great Ormond Street Hospital for a diagnostic fast. Indications included recurrent episodes of documented or suspected hypoglycaemia without fulfilling the diagnostic criteria for ketotic hypoglycaemia (n = 9), recurrent ketoacidosis (n = 1), or suspicion of fatty acid oxidation defect (n = 1). There were more males in the KH group than in the non-KH group (7/9 versus 4/11).

Protocol

All children were admitted to the programmed investigation unit at Great Ormond Street Hospital NHS Trust and fasted according to previously published guidelines.¹⁶ In particular the maximum fasting period was defined prior to the start of the fasting period.¹⁶ Two cannulas were inserted, one in the cubital fossa for tracer infusion (see below), the other in a contralateral superficial hand vein for blood sampling. Blood for intermediate metabolites, hormones, amino acids, acylcarnitine profile, and carnitine concentrations was drawn at specified times.¹⁶ Blood glucose was measured at hourly intervals or more frequently where appropriate. The fast was stopped when symptoms of hypoglycaemia occurred, when the blood glucose dropped below 2.6 mmol/l, or when the fast exceeded the specified duration.¹⁶

Organic acids were analysed in a urine sample at the end of the fast by gas chromatography-mass spectrometry (GC-MS). Plasma NEFA levels were measured using a kit (Wako Chemicals 99475409). Glucose concentrations were measured using a Kodak Ektachem analyser. Lactate and 3-hydroxybutyrate concentrations were measured using kits (Sigma 826B; Randox Laboratories 123530 respectively). Plasma pyruvate and acetoacetate concentrations were analysed by a centrifugal analyser with a fluorimetric attachment as reported previously.¹⁷ Plasma cortisol and insulin concentrations were measured using kits (Abbott 9116-60 and 2410-20). Growth hormone was measured by an immunoradiometric assay (NETRIA). Total and free carnitine levels were measured by a radiochemical enzymatic assay.¹⁷ Acylcarnitine profile was analysed using tandem mass spectrometry. Plasma amino acids were analysed by a LKB Biochrom 4400 automated amino acid analyser (Milton Keynes, UK).

A primed continuous infusion of 1-¹³C leucine (0.5 mg/kg/h) and (6,6)-D₂ glucose (3 mg/kg/h) was started for 4 hours at the beginning of the final 4 hours during the fast; 0.5 mg/kg of 1-¹³C bicarbonate, 0.5 mg/kg of 1-¹³C leucine, and 3 mg/kg of (6,6)-D₂ glucose were given as a prime dose just prior to the start of the infusion. Leucine kinetics were measured in 9/9 children with KH and 6/11 controls. Glucose kinetics were measured in 7/9 children with KH and 9/11 controls as previously reported.^18,19 Blood and breath samples for isotopic enrichment were taken at baseline and every 15 minutes throughout the final 2 hours of the infusion. Blood was transferred to heparinised tubes, spun immediately, and the plasma stored at −70°C until analysis. Analysis for plasma isotopic enrichment of glucose and ketoisocaproic acid (KIC) was done by GC-MS as previously reported.^19,20 Isotopic enrichment of CO₂ was measured by isotope ratio mass spectrometry. Leucine and glucose kinetics were calculated using plasma isotopic enrichment of 1-¹³C leucine and (6,6)-D₂ glucose and carbon dioxide production as described in more detail elsewhere.^18,19

Indirect calorimetry (Deltatrac I, Datex, Finland) was used to measure basal metabolic rate and respiratory quotient under standardised conditions as described previously.²⁰

Statistical significance was assumed at a level of p < 0.05 using Student’s t test. All data were analysed using a standard software package (Microsoft Office 2000) run on a Hewlett Packard N5340 laptop computer.

The study was approved by the Joint Research Ethics Committee, and parents of all participants gave written informed consent. Where appropriate, the participants consented as well.

RESULTS

The length of the fast was 16±2.28 hours (mean±SD) in children with KH and 14.47±5.58 hours in control subjects (not significant). Three children with KH became hypoglycaemic (glucose ⩽2.6 mmol/l) during the fast, while only one child from the control group became hypoglycaemic. Blood glucose levels at the beginning of the fast were 4.01±0.43 in KH versus 4.54±0.34 mmol/l in controls (p < 0.05) compared to 3.08±0.40 in KH versus 3.60±0.69 mmol/l in controls at the end of the fast. Plasma ketone body, non-esterified fatty acid (NEFA), insulin, cortisol, and growth hormone levels at the end of the fast are shown in table 2.

Plasma total and free carnitine levels as well as acylcarnitine species in both groups were within normal limits (data not shown). Plasma alanine concentrations were below our normal range in three subjects with KH, but normal in controls (data not shown). Branched chain amino acid levels were elevated in four subjects with KH but normal in controls (data not shown). Organic acid analysis was unremarkable, with an appropriate ketone response in all but one patient who showed excretion of keto acids.

Plasma insulin levels were significantly lower at the end of the fast (p < 0.05) in KH compared to control subjects. Plasma ketone body levels were significantly higher in KH (acetoacetate, p < 0.006; 3-hydroxybutyrate, p < 0.05) than in controls (table 2).

Basal metabolic rate was significantly higher in subjects with KH compared with control subjects (45.48±7.41 v 31.81±6.72 kcal/kg/day, p < 0.02), although the respiratory quotients were similar in both groups (0.84±0.05 v 0.8±0.04). Hepatic glucose production rates were significantly lower in KH than controls (3.84±0.46 v 6.6±0.59 mg/kg/min, p < 0.0005). Leucine oxidation rates were significantly lower in children with KH compared to controls (12.25±6.25 v 31.96±8.59 μmol/kg/h, p < 0.05).

DISCUSSION

Ketosis is part of the physiological response to fasting or pregnancy, or a result of a high fat (ketogenic) diet.^9–12 It may also occur during decompensation of diabetes mellitus, panhypopituitarism with secondary growth hormone and corticosteroid deficiencies, intoxication with alcohol or salicylates, or in children with KH.^3,9,14–16 Rare causes include maple syrup urine disease and inborn errors of ketolysis, such as succinyl CoA:3oxoacid CoA transferase (SCOT) and mitochondrial acetoacetyl CoA thiolase (MAT) deficiencies.⁹

KH is the most common form of hypoglycaemia during infancy and childhood.³ The typical patient is male and between the ages of 2 and 7 years. This age distribution is also reflected in the patient population which was studied here. These children usually present with hypoglycaemia associated with ketonuria during an intercurrent illness associated with fasting.⁴ Hypoglycaemia responds promptly to glucose administration. Seizures secondary to hypoglycaemia may occur, but neurological sequelae are very rare. In addition, hypoglycaemia may reoccur and cataracts may develop in some patients.⁶ KH typically improves with age and is usually not observed after children reach puberty. Diagnostic work-up excludes other causes of ketosis as outlined above.

The pathophysiology of KH is largely unknown. Some evidence suggested a disorder of hepatic glucose production, specifically a functional defect of gluconeogenesis secondary to a lack of gluconeogenic substrates, particularly alanine.^7,8 However, only three of our subjects with KH had low plasma alanine levels. Other studies indicated a failure of the adrenergic response during episodes of KH.²¹ Based on a simple model, KH may be regarded as an imbalance of energy supply and demand during episodes of increased metabolic stress. Surprisingly, few data are available for glucose and leucine kinetics or basal metabolic rate. We have studied a group of children with KH and compared the data to a group of children of similar age, height, weight, and BMI using stable isotopes.^22–24

Hepatic glucose production rates were significantly lower in KH, reflecting a failure of glucose supply as the main cause of ensuing hypoglycaemia. Obviously, this may be secondary to depletion of glycogen stores and/or failure of gluconeogenesis. As all tracer studies were done in steady state, glucose production equalled glucose utilisation rates excluding increased glucose oxidation rates as the cause of hypoglycaemia. However, there was little evidence to suggest the failure of gluconeogenic substrate supply as plasma alanine levels were only lower in three subjects with KH.

While the BMR was significantly elevated in subjects with KH by about 25% compared to controls, the respiratory quotient was comparable in both groups. The increase of BMR is not due to differences in BMI between the two groups but may be secondary to differences in intermediary metabolism. One explanation might be enhanced lipolysis for ketogenesis in KH which in part is reflected by increased ketone body concentrations in KH. However, this hypothesis is not supported by the respiratory gas exchange measurements as the RQs were similar in both groups. Both were oxidising similar substrates and glucose oxidation still exceeds fatty acid oxidation in both groups. This is also consistent with our hypothesis that children with KH fail to utilise ketone bodies.

Surprisingly leucine oxidation rates were significantly lower in KH compared to controls. The rates were comparable to those previously reported in children with intermediate maple syrup urine disease (MSUD).²⁴ This might explain the observation of raised concentrations of branched chain amino acids in KH during hypoglycaemia, but the corresponding ketoacids are not found as would be expected in maple syrup urine disease. The explanation for these findings is yet unknown.

In summary, we have shown that KH is caused by a failure to sustain sufficient hepatic glucose production rather than by increased glucose oxidation. Ketone utilisation does not explain the difference and hence the neurological symptoms. Surprisingly, leucine oxidation is reduced while energy expenditure is significantly increased. This may be of pathophysiological relevance and warrants further investigation.