Light on the horizon? Will Continuous Glucose Monitoring Allow for Better Management of Congenital Hyperinsulinism?

Glucose is the primary oxidative fuel for the brain. A recent analysis combining PET (positron emission tomography) and MRI (magnet resonance imaging) data to calculate the human brain’s glucose use from birth to adulthood showed that the brain dominates the body’s metabolism during early life (1). Therefore, lack of glucose supply results particularly in symptoms of cerebral dysfunction. The spectrum of neonatal hypoglycemia ranges from silent (asymptomatic) hypoglycemic episodes to severe symptomatic hypoglycemia. Depending on the severity and duration of hypoglycemia, sequalae range from subtle neurodevelopmental consequences up to severe brain damage (2,3). In the latter, MRI studies showed that severe (median glucose level: 1 mmol/L) and symptomatic hypoglycemia resulted in white matter abnormalities in 94% of the babies (2). Besides preterm infants, infants small for gestational age, and offspring of diabetic mothers, patients with the rare condition of genetic congenital hyperinsulinism (CHI) are extremely vulnerable to neonatal hypoglycemia (4).

Recently, several milestones have been achieved in the diagnosis and therapy of genetic CHI. Rapid genetic testing, prompt referral to specialized centers, and 18F-DOPA PET scan are excellently suited to identify patients with focal disease and to provide definitive cure (4). However, adverse long-term neurological outcome is still reported in up to 40% of cases (4, 5). A recent cohort study with a high proportion of patients with severe CHI showed that neurodevelopmental impairment was associated with lowest recorded blood glucose ≤1 mmol/L, and treatment delay from first symptom to expert center (5). Overall neurodevelopmental impairment and abnormal MRI was found in almost half of the patients. Additionally, hyperglycemia due to overtreatment may also contribute to developmental delay. Treatment of dysglycemia is usually guided by frequent blood sampling. However, the use of intermittent blood glucose levels carries the risk of overlooking silent hypo- and hyperglycemic episodes. Furthermore, rapid changes in glucose concentrations resulting in severe hypoglycemia may be detected first by cerebral symptoms rather than by blood glucose control. Therefore, continuous glucose control would be highly desirable.

Within the above context, the recent paper by the Beardsall group (6) is of particular interest. It reports a retrospective study of real-time continuous glucose monitoring (rtCGM) in preterm infants and babies with CHI using 2 different devices. The aim was to provide insights into patterns of dysglycemia by rtCGM and how rtGCM may contribute to the management of persistent neonatal hypoglycemia. Marked and rapid fluctuations were found in glucose levels in babies with CHI. Variation in glucose levels was higher in babies with CHI than in preterm infants. This finding is of practical significance with respect to detection of hypoglycemia because rtCGM measures interstitial fluid glucose levels leading to a physiological lag between interstitial and blood glucose levels. A large number of short hypoglycemic episodes with rtCGM levels <2.6 mmol/L was detected, with 220 out of 233 episodes lasting less than an hour and without clinical symptoms. Some of these hypoglycemic episodes were not noted because nurses only checked CGM every hour as the alarms were switched off. Other rtCGM hypoglycemic events were accompanied by normal blood glucose levels. Only 13 episodes of more prolonged, and therefore considered as clinically significant, rtCGM hypoglycemia (CGM <2.6 mmol/L for >60 minutes) were detected. Of these, 2 episodes were considered to be true hypoglycemia and were treated with glucose. The remaining 11 were associated with normal blood glucose.

The authors conclude that the use of rtCGM to provide reassurance during periods of normoglycemia could potentially limit the need for such frequent and painful blood sampling. In cases of hypoglycemia, the rtCGM may provide continuous trends that would highlight falling glucose levels and alert the clinician to the need for POC blood glucose measurement. However, clinical management during this study was excellent and only a few hypoglycemic episodes were available for analysis. But the Beardsall group reported that rtCGM can reduce exposure to prolonged or severe hyperglycemia and hypoglycemia in 155 preterm infants of less than 1200 g birth weight in a randomized controlled trial (7). Finally, it is important that Beardsall and colleagues showed that rtCGM is safe with respect to skin integrity or infection in both studies (6, 7). However, correct positioning of the sensor is still a problem to be overcome in these tiny patients. A pity that a comparison between the 2 devices for rtCGM has not been made.

Should rtCGM therefore be part of routine care in babies with risk of dysglycemia, particularly in babies with CHI to improve neurodevelopmental outcome? Here, severe hypoglycemia occurs often during manifestation of the disease with the need of measuring blood glucose every 20 minutes and aggressive treatment with intravenous glucose. However, the sensor needs warming up and in those instruments requiring multiple daily calibration, this procedure is quite sophisticated requiring phases of stable glucose levels. Thus application of rtCGM is time consuming and therefore may not be suitable in this early phase of disease before newer rtCGMs devices are tested in this setting (6). Furthermore, medical decision making can currently not be based on a supposedly suspicious rtCGM measured glucose value. A suspicious rtCGM glucose has to be immediately controlled by taking the corresponding blood glucose level, such as by a point of care blood glucose meter. Another hindrance was that rtCGM devices have hypoglycemia alarms for trends and absolute glucose but this was not regarded as useful by the clinical staff (6). Therefore, appropriate choice of thresholds and development of suitable algorithms for alarm settings that meet the needs for neonatal care have to be developed.

And so there is still some way to go before rtCGM can take the role of a routine diagnostic pillar of glucose monitoring in neonates. Further rapid development would be highly desirable since these fragile individuals would benefit 2-fold from the new technique: on the one hand, clinical care would without any doubt be significantly improved; on the other hand, the technique could serve as a unique research tool, such as for correlating the various glucose levels with measures of quality of care, as well as neurodevelopmental or metabolic outcome. In another group of newborns at risk for dysglycemia, studies showed that the risks of obesity and metabolic syndrome in offspring of mothers with gestational diabetes are considerably increased (8).