Article Text

Download PDFPDF

Glucose monitoring in neonates: need for accurate and non-invasive methods
  1. Hyung Chul Woo1,
  2. Leah Tolosa2,
  3. Dina El-Metwally1,
  4. Rose M Viscardi1
  1. 1Division of Neonatology, Department of Pediatrics, School of Medicine, University of Maryland, Baltimore, Baltimore, Maryland, USA
  2. 2Department of Chemical, Biochemical and Environmental Engineering, Center for Advanced Sensor Technology, University of Maryland, Baltimore County, Baltimore, Maryland, USA
  1. Correspondence to Dr Hyung Chul Woo, Division of Neonatology, Department of Pediatrics, University of Maryland School of Medicine, 110 South Paca Street, 8th Floor, Baltimore, MD 21201, USA; hwoo{at}


Neonatal hypoglycaemia can lead to devastating consequences. Thus, constant, accurate and safe glucose monitoring is imperative in neonatal care. However, point-of-care (POC) devices for glucose testing currently used for neonates were originally designed for adults and do not address issues specific to neonates. This review will address currently available monitoring options and describe new methodologies for non-invasive glucose monitoring in newborns.

  • Neonate
  • Neonatal Hypoglycemia
  • Non-Invasive Monitoring
  • Glucose Binding Protein

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Glucose concentrations are the most frequently measured laboratory values in neonatal medicine.1 Despite considerable research on the adverse impact of neonatal hypoglycaemia, there are extreme variations in the definition of low plasma glucose levels (PGLs) in newborns in the first postnatal days, ranging from <30 to ≤60 mg/dL (<1.67 to ≤3.33 mmol/L).2 At present, there is neither a rational basis, nor sufficient evidence to identify a specific value or a range of PGLs that would define ‘hypoglycaemia’ as a pathologic entity.3–5

There are known risk factors for neonatal hypoglycaemia (table 1) and tighter control of glucose in these vulnerable neonates have proven to be beneficial.5 ,6

Table 1

Risk factors for neonatal hypoglycaemia

In a single institution retrospective review, 45% to 50% of very low birthweight (≤1500 g) and 70% of extremely low birthweight (≤1000 g) neonatal intensive care unit (NICU) admissions had documented episodes of asymptomatic hypoglycaemia during the first week of life.5 ,6 Hypoglycaemia defined as <47 mg/dL (2.61 mmol/L) occurred in 72% in small for gestational age neonates and 16.2% in large for gestational age neonates of non-diabetic mothers.1

Symptomatic hypoglycaemic neonates can have a poor prognosis, with abnormalities ranging from learning disabilities to cerebral palsy and persistent or recurrent seizure disorders, and mental retardation.1 ,7 ,8 Thus, close monitoring of plasma glucose in the transition period after birth in all high-risk neonates is needed to avoid the consequences of hypoglycaemia.9

The ability to measure glucose levels in newborns is critical to good clinical care. Glucose monitoring devices need to give accurate results particularly at the levels of clinical significance that would affect clinical care. However, devices also need to be practical to use and cost efficient. In designing a point-of-care (POC) device for neonatal use, balancing efficiency and accuracy is of utmost importance. In this review, we will describe the current methodologies for neonatal glucose monitoring and their limitations and new promising methodologies including non-invasive approaches with improved sensitivities in the hypoglycaemic range (table 2).

Table 2

Current methodologies for neonatal glucose monitoring

Current glucose monitoring methods

Laboratory analysers

The optimum method for measuring glucose is still the enzymatic-based laboratory analysers10 ,11 that measure plasma/serum glucose utilising glucose oxidase, hexokinase or glucose dehydrogenase reactions.7 These methods are considered the gold standard for clinical measurements of glucose levels and are less affected by interference by metabolites and haematocrit.7 However, even laboratory testing has pitfalls. Results may not be available quickly enough for timely appropriate treatment7 ,10 and frequent blood drawing contributes to blood loss, risk for infection and altered pain sensitivity.12–16

POC devices and their limitations

POC glucose meters, when used in conjunction with appropriate interventional treatments, can effectively improve glycaemic control in children and adults.17 The USA Food and Drug Administration (FDA) has approved >200 POC glucose meters for home or institutional use.17 A major advantage of bedside or POC devices (reflectance glucometers) is their practicality, with rapid turnaround times (≤3 s) requiring small sample volumes (≤0.3 μL)7 ,11 providing efficient clinical care.

However, commercially available meters were originally developed for use in adult diabetic patients to periodically monitor and/or detect normoglycaemia or hyperglycaemia.7 ,10 ,11 ,18 ,19 These meters measure a current generated by an enzyme reaction (glucose oxidase or glucose dehydrogenase) that is proportional to the amount of glucose in the blood sample.7 However, the accuracy of these devices is of concern in the neonatal population.3 ,9–11 ,17 POC devices correlate well with laboratory measurements in the normoglycaemic and hyperglycaemic ranges, but the accuracy is limited at the low glucose concentrations,7 ,10 ,19–22 with variations from actual levels as much as 10 to 20 mg/dL (0.55 to 1.11 mmol/L).4 ,9 ,19 The limits of accuracy of these devices are at the threshold of glucose levels for changes in clinical management in newborns.7 ,10 Therefore, although useful for screening, such devices cannot be relied upon for the accurate diagnosis of neonatal hypoglycaemia.7

Since POC meter readings are used to inform therapeutically important decisions, it is essential that accuracy of these measurements be comparable to laboratory analysers.11 However, in the hypoglycaemic range, differences greater than 20% and 10% occurred 57% and 75% of the time, respectively, between some POC analysers and laboratory values.11 The current FDA (and the International Organisation for Standardisation or ISO) minimum acceptable system accuracy for POC meters states that 95% of individual glucose results shall fall within ±15 mg/dL (±0.83 mmol/L) of results of reference measurements at glucose concentrations <75 mg/dL (<4.17 mmol/L).23 ,24 So, when the ‘true’ glucose level of a hypoglycaemic neonate is 35 mg/dL (1.94 mmol/L), it is acceptable for POC meters to give values between 20 and 50 mg/dL (1.11 to 2.78 mmol/L). At the higher limit, there is the potential for mistakenly assuming the blood glucose level to be normal. Furthermore, factors that cause inaccuracies of POC testing in neonates are haemoglobin concentrations and packed cell volumes,10 ,19 metabolic acidosis,25 high blood oxygen tension levels,26 high bilirubin levels27 and/or oedema.28 Thus, there is an adjunct FDA guidance for glucose monitoring devices for use in neonates. Specifically, data demonstrating the precision and accuracy of the system in the range of 10 to 50 mg/dL (0.55 to 2.78 mmol/L) should be provided when seeking FDA approval. At present, there is no POC method that is sufficiently reliable and accurate in this low range of blood glucose levels to allow it to be used as the sole method for screening for neonatal hypoglycaemia.3 ,4 ,7 ,10 ,19 ,29

New invasive and non-invasive neonatal glucose monitoring methods

Invasive—role for continuous glucose monitoring

New methods such as continuous subcutaneous glucose monitoring devices are increasingly being used to help improve glycaemic control in patients with diabetes and have the potential to help provide improved glucose monitoring and management in high-risk neonates.7 In adults and children with diabetes, continuous glucose monitoring using a subcutaneous sensor has recently been validated and shown to be useful in the management of glucose control, revealing episodes of asymptomatic hypoglycaemia and hyperglycaemia.5 The current means of intermittent glucose sampling can result in many hours between measurements and data from the continuous glucose monitoring system (CGMS) revealed that there were often prolonged periods, sometimes up to 12 h, between patient blood glucose measurements. In one instance, CGMS detected hypoglycaemia almost 3 h before the routine blood glucose check was scheduled to be done.5 Continuous glucose monitoring has demonstrated that glucose levels can fluctuate widely, particularly in infants requiring intensive care, and may result in undetected periods of hypoglycaemia.5 ,7

However, there are limitations to the use of current CGMS devices in the NICU. Current CGMS devices use a probe that needs to be inserted into subcutaneous tissue, leading to puncturing/breakage of skin and the potential for infection.5 Another limitation of CGMS is that the sensor device measures interstitial rather than whole blood or plasma/serum glucose. In two infants with oedema, the sensor was noted to have periods when no measurements were recorded.5 Reliability and accuracy of CGMS has not been tested in hypoglycaemic infants.5 Although the sensors can be left in situ for 7 days without loss of accuracy, they need to be calibrated with whole blood or plasma/serum glucose measurements at least four times per day. Thus, the need for blood glucose sampling is not completely avoided.7 ,30

Continuous glucose monitoring with subcutaneous perfusion devices has been used to a limited extent in preterm infants.3 But larger studies are required to determine the safety and effectiveness of such monitoring.3 Alternative methods of continuous glucose monitoring include microdialysis.7 However, these devices are invasive, need calibration, require a significant lag time in collection and measurement and are expensive.7 Therefore, these devices are limited to use in research within NICUs31 ,32

Non-invasive neonatal glucose monitoring

Attempts have been made to develop non-invasive techniques to measure glucose levels to prevent the need for frequent blood sampling.7 Such methodologies include optical sensors or transdermal devices, but attempts have generally been unsuccessful.7 Several spectroscopic methods have been in development for years including near-infrared, mid-infrared, Raman, photoacoustic and terahertz spectroscopy.33–37 These methods generally involve exposing the skin to light or other forms of radiative energy and isolating the minute signal of the blood and tissue glucose from the large background interferences. Multiple corrections on the hardware and software are required to accomplish this, and to date, none have resulted in a commercial device. Thus far, there have been no satisfactory methods for non-invasive monitoring of glucose or alternate substrates.3

Novel approach to glucose monitoring

Investigators at the Center for Advanced Sensor Technology (CAST) at the University of Maryland Baltimore County (UMBC) are developing a neonatal glucose sensor that will address the issues of current practice. The sensor is non-invasive and painless, sensitive at very low glucose concentrations, easy to use by medical personnel, and produces rapid results at the bedside. The sensor takes advantage of the underdeveloped neonatal skin, which is highly permeable to small neutral molecules such as glucose. This permeability is a function of gestational and postnatal age38 ,39 and can be measured as transepidermal water loss (TEWL). Other investigators have evaluated this increased permeability as a potential non-invasive route for delivery of drugs to sick neonates.40–42

To non-invasively monitor drugs given to neonates, a process called reverse iontophoresis has been attempted.43 Iontophoresis is a method of applying a mild electric current to induce the flow of molecules through the skin. It has been generally used to deliver drugs through the skin while the reverse process has been used to non-invasively monitor drugs administered to patients. It was found that for neutral molecules, reverse iontophoresis on neonates was no better than passive diffusion through the skin.43 While this is not a desirable outcome for drug monitoring, it is of particular significance for monitoring small neutral molecules like glucose that could easily diffuse through the skin together with TEWL. As the amount of glucose on the skin is much less than in blood, current glucose oxidase–based devices are not sensitive enough to measure the glucose on the skin. The glucose sensor developed by the CAST group is based on the glucose binding protein (GBP) and addresses this issue as it is up to four orders of magnitude more sensitive than any available glucose sensor in the market today.44–50 The GBP is found in the periplasmic space of gram-negative bacteria. The GBP is not an enzyme and signal transduction is effected through the change in conformation of the protein upon binding of glucose (figure 1—left). The advantage of this change in protein conformation has been taken up by strategically introducing a cysteine mutation at position 255 and labelling that mutation with the polarity-sensitive dye, acrylodan. Changes in the fluorescent properties of the dye are then correlated with the concentration of the analyte (glucose) in the sample (figure 1—right). The CAST group 44–47 ,49 ,51–55 and others56–61 have shown this method of signal transduction for several binding proteins for biosensing applications as a broad class of reagentless biosensors. The CAST group, however, have tailored the photophysical properties of the labelled proteins to complement the engineering of an optical glucose monitoring biosensor. This biosensor is highly specific for glucose even in a complex mixture. This novel glucose biosensor methodology is currently being incorporated into a low-cost and user-friendly POC device51–55 while maintaining sensitivity, accuracy and miniaturisability.51 ,55

Figure 1

(Left) The glucose binding protein (GBP) undergoes conformational change in the presence of glucose. (Right) The fluorescence spectra of the GBP labelled with acrylodan and a ruthenium metal ligand complex at increasing concentrations of glucose. Note that the glucose concentrations are in the μM levels.

To collect the transdermal glucose (TG), which is present in micromolar (µmolar) concentrations, a small amount of buffer is applied on clean skin, recirculated on the surface for a brief period of time, collected and analysed for its glucose concentration using the GBP assay. The method is painless and gentle and ideal for infants.

Differentially tape-stripped porcine skin was developed and validated as an in vitro model for transdermal drug delivery studies.62 ,63 Using this in vitro model to test the GBP-based glucose sensor, a linear correlation was observed between the concentration of the glucose under the skin and glucose collected on the skin (R^2=0.9433, p value for slope=0.001 and p value for intercept=0.351). In repeated experiments on adults, the TG collected appears to track blood glucose levels.64 Preliminary results on preterm neonates show about 40 times more glucose collected on skin compared to adults with the same blood glucose levels. This is consistent with the higher permeability of the neonatal skin relative to that of adults.


Measurements obtained using POC glucose meters should be interpreted with caution and when possible, verified with a laboratory analyser, especially at critical hypoglycaemic levels, in order to minimise misdiagnosis.11 In addition, these glucose meters require breaking of skin, causing pain and the potential for infection. However, no methods currently available painlessly and non-invasively evaluate glucose levels accurately in neonates. For good reasons, there is a call for a development of such devices. The GBP biosensor that can detect the very low dermal glucose concentrations is proof of principle that non-invasive glucose monitoring will be possible in the future.



  • Funding The CAST-UMBC work was supported by the NIH grant R41HD069207 and partly by GE Healthcare.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.