Continuous glucose monitoring (CGM) in very low birth weight newborns needing parenteral nutrition: validation and glycemic percentiles.

Background Continuous glucose monitoring using subcutaneous sensors is useful in the management of glucose control in neonatal intensive care. We evaluated feasibility and reliability of a continuous glucose monitoring system in a population of very low birth weight neonates needing parenteral nutrition. Moreover, we presented percentiles of glycemia of the studied population. Methods Very low birth weight neonates were enrolled within 24 h from birth. An ENLITE sensor connected to a continuous glucose monitoring system was inserted and maintained for at least 72 h. Data obtained with the continuous glucose monitoring system and with a glucometer were compared. Calibration was performed every 12 h. Results Twenty-three patients (9 males) were included. Median gestational age was 28 weeks (range 23–30) and median birth weight was 860 g (range 500–1092). A total of 299 paired glucose values were obtained. Modified Clarke Error Grid criteria for clinical significance were met. 74 and 33 episodes of hypoglycemia and hyperglycemia were detected, respectively. 31,329 values of glycemia were analyzed and the percentiles calculated. Conclusions This continuous glucose monitoring system is safe and accurate. It allows increasing the detection of hypo- and hyper-glycaemia episodes and it could be routinely used in the management of glucose infusion in very low birth weight neonates under total parenteral nutrition.


Background
Very low birth weight (VLBW) neonates are at high risk of glycemic disorders. Neonates requiring intensive care have impaired glucose control and a wide fluctuation in blood glucose levels. Their developing brain is likely to be more susceptible to these metabolic insults. Hypoglycemia is associated with poor neurodevelopmental outcome [1][2][3] and hyperglycemia is correlated with increased mortality and morbidity in preterm infants [4]. Plasma glucose level detection represents the gold standard to diagnose these metabolic disorders, however, this method produces only punctual values and does not allow a real-time (RT) monitoring of glycaemia and its trends. This could explain the great deal of controversy over the values of normoglycemia in neonatal population and the paucity of proper diagnoses [5,6]. Specifically, there is no consensus on the definition of "significant hypoglycemia" (blood glucose values requiring a medical treatment), and on the threshold of glucose concentration and the time needed to cause neurological damage [7,8]. Continuous glucose monitoring (CGM) can be used to investigate glucose homeostasis during the neonatal period and some studies showed its use is safe and reliable [8][9][10]. Moreover, it is demonstrated that CGM can detect a significantly higher number of hypo/ hyperglycemia episodes as compared to intermittent blood glucose measurement. A better management of glucose administration [11] can be obtained with a RT visualization of glycemic values in the new generation CGM system (CGMS) [12][13][14]. Recently, however, some authors expressed concerns about using CGM in the clinical setting [15]. In fact, they showed that the calibration methods of new generations CGM sensors are designed for higher glucose concentrations of children and adults, and not for neonates. Therefore, the routine usage of a CGMS with its specific sensor, could be proposed only after validation studies conducted in a neonatological setting. This study was conducted using RT Medtronic's CGMS, a system previously tested by our research group in a population of neonates at risk of dysglycemia with a median gestational age (GA) of 32 weeks. Not many dysglicemic episodes were detected, but the CGMS demonstrated reliable in the normoglycemic range [12]. Only VLBW neonates fed by parenteral nutrition were included in this study, with the aim of testing the reliability of the Medtronic's CGMS in this specific new setting. Babies had a lower mean gestational age and birth weight than in the previous study, and both hyper-and hypoglycemic episodes were detected during the study period. We also aimed at describing the distribution of glycemic values in the studied population. Such data is currently lacking in the scientific literature [16] for VLBW infant requiring parenteral nutrition, but it could be very useful to allow a proper modulation of the glucose infusion, thus preventing dysglycemic episodes, to reduce and optimize the need of insulin therapy, and hopefully to lead to the definition of a new protocol for a strict glycemic control in randomized controlled trials.

Population
All data were collected in the Neonatal Intensive Care Unit (NICU) of the Catholic University A. Gemelli Hospital between June 1st, 2016 and March 31st, 2017. This non-randomized feasibility study was conducted following the approval of our institutional review board. The study was aimed at testing the reliability of the Medtronic's CGMS in a population of VLBW in parenteral nutrition and to investigate CGM data on the distribution of glycemic values in this specific population. Newborn infants were eligible for inclusion if they were VLBW and, accordingly to our internal protocols, a central line was positioned to administer parenteral nutrition. The criteria used in our unit to start parenteral nutrition, based on ESPGHAN, ESPEN and ESPR guidelines [17][18][19], were applied: Birth Weight (BW) ≤1250 g; 1250 g ≤ BW ≥ 1500 g associated with severe RDS (invasive mechanical ventilation; nCPAP and FI02 ≥ 0,3) or being small for gestational age with a documented prenatal history of severe placental insufficiency (umbilical doppler sonography assessing absent end diastolic (AED) or absent/reverse end diastolic (ARED) flow or brain sparing). Medtronic's CGMS (Northridge, Calif., USA) is routinely used in our NICU to manage glucose infusion in all newborns at high risk of dysglycemia, and eligible neonates were enrolled in the study only if the instrument was available within 24 h from birth. Parenteral nutrition was administered following ESPGHAN, ESPEN, ESPR recommendations [17,18]. The nutritional intake used in our unit is summarized in Table 1.
Newborns with major congenital abnormalities at birth or with skin diseases were excluded. Informed consent was obtained from the parents, and infants enrolled were monitored with CGMS within the first 24 h of life for at least 72 h. In case of detachment or malfunction, the device was replaced (no more than once).

CGMS
The CGMS is composed by the Enlite sensor, the Guardian transmitter and the VEO monitor. The new generation sensor has a cannula length of 8.75 mm. Other studies about CGMS reported the use of this device [11,12,14] but it is the first time, to our knowledge, that this CGMS is validated in this specific population of VLBW neonates fed by parenteral nutrition. The sensor of the CGMS used is equipped with a glucose-oxidase. This enzyme is used to detect the presence of glucose in the interstitial space. It  [20], and hyperglycaemia as a glucose value ≥180 mg/dl. The monitor was programmed in order to give alarms when hyper-or hypoglycemia were detected. As a previous study conducted in our institution [12] demonstrated a good accuracy of this CGMS in detecting normoglycemia in a population of preterm newborns, glucose administration was based on values showed in RT by the monitor and reported by the nurses in the clinical chart, in specific time intervals (hourly or every 3h) according to our protocols.
As there are concerns about the accuracy of CGMS in detecting hypoglycemia and hyperglycemia, a blood glucose test performed with the point of care glucometer was prescribed by the treating physician if episodes of hyper-or hypoglycemia were detected by the CGMS. An additional calibration of the instrument was performed if the CGMS glycemia value (CGMV) was inaccurate (|GTX-CGMV| > 20% CGMV). All the GTX obtained were compared with the correspondent CGMV. We preferred to use GTX (a point of care method) as a comparator, as it is most commonly used in clinical practice in the NICU as compared to laboratory enzymatic methods..

Results
Twenty-three infants were enrolled, 9 males and 14 females, with a median birth weight (BW) of 860 g (range 500-1092 g) and a median gestational age of 28 weeks (range 23-30 weeks). Other population details are summarized in Table 2.
The sensor was well tolerated with an average (DS) duration of 112,7 (31,7) h. We collected 299 pairs of CGMV vs GTX measurements. Figure 1 reports the Bland Altman Analyses for glucose measurements. The mean (95% CI) difference was 1,4 (− 33,5 to 36,3) mg/dl. Figure 2 reports the MCEG: 83,6% of measurements fall in region A, 15,4% of measurements fall in region B, Full enteral feeding during the observed period (n) 0 and 1,0% of measurements in region D, without any value in region C or E. It is to be mentioned that glucose values obtained during the study did not fall under the lower limit detected by CGMS. Seventy-four episodes of hypoglycemia and and 33 of hyperglycemia were detected; mean (DS) durations were 44 (48) and 96 min (130); 15 hypo-and 3 hyperglicemic CGMV turned out to be inaccurate at the blood sample analysis performed with point of care glucometer. 31,329 CGMV were obtained and analyzed.
Tables 3 and 4 present distribution and percentiles of glycemic values of the studied population.

Discussion
CGM has a great potential for optimizing glycaemic control [21] and there is growing interest in using CGM devices in neonatal intensive care. However, as remarked in a recent review [15], CGM has a clear role in adult and child care but there are some concerns about its use in neonatology. This study evaluated feasibility and safety of the Medtronic's CGMS in VLBW fed by parenteral nutrition. This CGMS proved to be safe to use in this population and it was well tolerated after the sensor's insertion. Moreover, it didn't interfere with nursing care and it showed an unexpected satisfaction by the neonatal staff. Furthermore, MCEG demonstrated the accuracy of the tool tested in the studied population. However, the data presented here show that this CGMS was inaccurate in some episodes of both hyper-and hypoglycemia, although all such inaccuracies by CGMV evidenced by GTX, fell in B region of MCEG, thus not leading to an inappropriate treatment. Some authors have published previous studies focused on CGM in neonates, suggesting the CGM as a useful instrument to be used in neonates requiring intensive care and in VLBW [9,10,22]. However, in all these studies both an old sensor and an old (not RT) continuous glucose monitoring system were used. More recently, the usefulness and reliability of RT-GCM was demonstrated in babies born at 32 weeks gestation who were at risk of hypoglycemia [23]. Moreover, a randomized controlled trial focused on VLBW babies, describes the CGM as a more effective tool in detecting and treating hypoglycemic episodes when compared to   [11]. New glucose infusion protocols [13,24], based on real time visualization of glycemic values, are incoming and will be proposed for newborns at high risk of hyper-or hypoglycemia. The main strengths of our study are the focus on a selected population at high risk of disglycemia and the use of a new generation sensor and a RT-CGM device. We also described the distribution of glycemic values of the studied population. At our knowledge it is the first time such an analysis has been performed. This distribution is very interesting, as in the future new glucose infusion rate protocols, based on the CGM usage, may be designed to tailor patient-based nutritional intake. Glycemic percentiles' generation could allow to overcome the current limits of the "significant hypoglycemia" definition, correlating it to the glycemic trend rather than to punctual values. The main limitation of our study is the small size of the population. However, the observational design allowed to assess the efficacy and safety of the CGMS in this specific population. In conclusion, our results suggest that the presented CGM system with its sensor can be used, being both safe and accurate, in the management of glucose infusion and insulin therapy in VLBW neonates in total parenteral nutrition during the first week of life.

Conclusions
Medtronic's CGMS is safe and reliable in VLBW neonates fed by parenteral nutrition and is well tolerated after the sensor's insertion. It doesn't interfere with nursing care, as shown by the good satisfaction by the neonatal staff. Collecting glycemic values allowed to generate glycemic percentiles that could be useful to tailor patient-based glucose intakes. The data presented here are promising but a larger population of VLBW is needed in a randomized controlled clinical trial setting to properly define the place in therapy of CGMS. Availability of data and materials All data and materials are available in our computerized data base.
Authors' contributions A.P. conceptualized this study. A.P., L.G. and E.T. designed the data base used in this study, performed measurements, collected data. A.P. and L.G. analyzed the data. F.P. and M.C. wrote the paper. G.V. and E.Z. reviewed and approved the final manuscript as submitted. All authors read and approved the final manuscript and agree to be accountable for the article and to ensure that all questions regarding the accuracy or integrity of the article are investigated and resolved.

Ethics approval and consent to participate
This non-randomized feasibility study was conducted following the approval of our institutional review board. Consent was obtained before data collection.

Consent for publication
Not applicable.