Mean Absolute Relative Difference Values Research Articles

Accuracy of a novel real-time continuous glucose monitoring system: a prospective self-controlled study in thirty hospitalized patients with type 2 diabetes

AimsThe present study aimed to investigate the accuracy of the Glunovo® real-time continuous glucose monitoring system (rtCGMS).MethodsWe conducted a 14-day interstitial glucose level monitoring using Glunovo® rtCGMS on thirty hospitalized patients with type 2 diabetes. The flash glucose monitoring (FGM) was used as a self-control. Consistency tests, error grid analysis, and calculation of the mean absolute relative difference (MARD) were performed using R software to assess the accuracy of Glunovo® rtCGMS.ResultsGlunovo® exhibited an overall MARD value of 8.89% during hospitalization, compared to 10.42% for FGM. The overall percentages of glucose values within ±10%/10, ± 15%/15, ± 20%/20, ± 30%/30, and ±40%/40 of the venous blood glucose reference value were 63.34%, 81.31%, 90.50%, 97.29%, and 99.36% for Glunovo®, respectively, compared with 61.58%, 79.63%, 88.31%, 96.22% and 99.23% for FGM. The Clarke Error Grid Analysis showed that 99.61% of Glunovo® glucose pairs and 100.00% of FGM glucose pairs within zones A and B.ConclusionOur study confirms the superior accuracy of Glunovo® in monitoring blood glucose levels among hospitalized patients with type 2 diabetes.

Performance of subcutaneous continuous glucose monitoring in adult critically ill patients receiving vasopressor therapy.

Background Subcutaneous continuous glucose monitoring (CGM) may facilitate glucose control in the ICU. We aimed to assess the accuracy of CGM (Dexcom G6) against arterial blood glucose (ABG) in adult critically ill patients receiving intravenous insulin infusion and vasopressor therapy. We also aimed to assess feasibility and tolerability of CGM in this setting. Methods We included ICU patients receiving mechanical ventilation, insulin, and vasopressor therapy. Numerical accuracy was assessed by the mean absolute relative difference (MARD), overall, across arterial glucose strata, over different noradrenaline equivalent infusion rates, and over time since CGM start. MARD below 14% was considered acceptable. Clinical accuracy was assessed using Clarke Error Grid (CEG) analysis. Feasibility outcome included number and duration of interrupted sensor readings due to signal loss. Tolerability outcome included skin reactions related to sensor insertion or sensor adhesives. Results We obtained 2946 paired samples from 40 patients (18 with type 2 diabetes) receiving a median (IQR) maximum noradrenaline equivalent infusion rate of 0.18 (0.08-0.33) µg/kg/min during CGM. Overall, MARD was 12.7 (95% CI 10.7-15.3) % and 99.8% of CGM readings were within CEG zones A and B. MARD values ≥14% were observed when ABG was outside target range (6-10 mmol/l [108-180 mg/dl]) and with noradrenaline equivalent infusion rates above 0.10 µg/kg/min. Accuracy improved with time after CGM start, reaching MARD values below 14% after 36 hours. We observed four episodes of interrupted sensor readings due to signal loss ranging from 5 to 20 minutes. We observed no skin reactions related to sensor insertion or sensor adhesives. Conclusions In our ICU cohort of patients receiving vasopressor infusion, subcutaneous CGM demonstrated acceptable overall numerical and clinical accuracy. However, suboptimal accuracy may occur outside glucose ranges of 6-10 mmol/l (108-180 mg/dl), during higher dose vasopressor infusion, and during the first 36 hours after CGM start.

Using Artificial Intelligence to Improve the Accuracy of a Wrist-Worn, Noninvasive Glucose Monitor: A Pilot Study.

Self-monitoring of glucose is important to the successful management of diabetes; however, existing monitoring methods require a degree of invasive measurement which can be unpleasant for users. This study investigates the accuracy of a noninvasive glucose monitoring system that analyses spectral variations in microwave signals. An open-label, pilot design study was conducted with four cohorts (N = 5/cohort). In each session, a dial-resonating sensor (DRS) attached to the wrist automatically collected data every 60 seconds, with a novel artificial intelligence (AI) model converting signal resonance output to a glucose prediction. Plasma glucose was measured in venous blood samples every 5 minutes for Cohorts 1 to 3 and every 10 minutes for Cohort 4. Accuracy was evaluated by calculating the mean absolute relative difference (MARD) between the DRS and plasma glucose values. Accurate plasma glucose predictions were obtained across all four cohorts using a random sampling procedure applied to the full four-cohort data set, with an average MARD of 10.3%. A statistical analysis demonstrates the quality of these predictions, with a surveillance error grid (SEG) plot indicating no data pairs falling into the high-risk zones. These findings show that MARD values approaching accuracies comparable to current commercial alternatives can be obtained from a multiparticipant pilot study with the application of AI. Microwave biosensors and AI models show promise for improving the accuracy and convenience of glucose monitoring systems for people with diabetes.

Clinical application of real-time continuous glucose monitoring system during postoperative enteral nutrition therapy in esophageal cancer patients.

Enteral nutrition (EN) support therapy increases the risk of abnormal blood glucose (BG). The aim of this study is to evaluate the clinical value of a real-time continuous glucose monitoring (rt-CGM) system in BG monitoring during postoperative EN support therapy in patients with esophageal cancer. Patients without diabetes mellitus (DM) with esophageal cancer who planned to receive postoperative EN were enrolled. With the self-monitoring of BG value as the reference BG, the accuracy of rt-CGM was evaluated by the mean absolute relative difference (MARD) value, correlation efficient, agreement analysis, and Parkes and Clarke error grid plot. Finally, paired t tests were used to compare the differences in glucose fluctuations between EN and non-EN days and slow and fast days. The total MARD value of the rt-CGM system was 13.53%. There was a high correlation between interstitial glucose and fingertip capillary BG (consistency correlation efficient = 0.884 [95% confidence interval, 0.874-0.894]). Results of 15/15%, 20/20%, 30/30% agreement analysis were 58.51%, 84.71%, and 99.65%, respectively. The Parkes and Clarke error grid showed that the proportion of the A and B regions were 100% and 99.94%, respectively. The glucose fluctuations on EN days vs non-EN days and on fast days vs slow days were large, and the difference was statistically significant (P < 0.001). The rt-CGM system achieved clinical accuracy and can be used as a new option for glucose monitoring during postoperative EN therapy. The magnitude of glucose fluctuation during EN therapy remains large, even in the postoperative population without DM.

Comparison of Point Accuracy Between Two Widely Used Continuous Glucose Monitoring Systems.

Safe and effective self-management of glucose levels requires immediate access to accurate data. We assessed the point accuracy of the Dexcom G7 Continuous Glucose Monitoring System (Dexcom, Inc., San Diego, CA, USA) and FreeStyle Libre 3 (Abbott Diabetes Care, Alameda, CA, USA) sensors in a head-to-head comparison. Multicenter, single-arm, prospective, nonsignificant risk evaluation enrolled adults (≥ 18 years) with diagnosed type 1 diabetes (T1D) or type 2 diabetes (T2D). Accuracy was assessed by comparing sensor data to laboratory reference values Yellow Springs Instrument [YSI] and capillary blood glucose values. Outcome measures were differences in mean absolute relative difference (MARD), number and percentage of matched glucose pairs within ±20 mg/dL/±20 of reference values within glucose ranges: < 54, 54 to 69, 70 to 180, 181 to 250, > 250 mg/dL, and combined. Data from 55 adults were included in the analysis. Analysis showed significantly lower MARD with the FreeStyle Libre 3 sensor vs the Dexcom G7 sensor (8.9% vs 13.6%, respectively, P < .0001) with a higher percentage of glucose values within ±20 mg/dL/±20 of reference (91.4% vs 78.6%). The MARD values for both continuous glucose monitoring (CGM) sensors were similar during the first 12 hours; however, the FreeStyle Libre 3 MARD was notably lower than the Dexcom G7 MARD during the next 12 hours (10.0% vs 15.1%, respectively, P < .0001) and throughout the study period. The FreeStyle Libre 3 sensor was more accurate than the Dexcom G7 sensor in all metrics evaluated throughout the study period. This is the first head-to-head study to our knowledge that compares the flagship products currently in widespread use of the two largest CGM manufacturers.

Analytical Performance of the Factory-Calibrated Flash Glucose Monitoring System FreeStyle Libre2TM in Healthy Women.

Continuous glucose monitoring (CGM) is used clinically and for research purposes to capture glycaemic profiles. The accuracy of CGM among healthy populations has not been widely assessed. This study assessed agreement between glucose concentrations obtained from venous plasma and from CGM (FreeStyle Libre2TM, Abbott Diabetes Care, Witney, UK) in healthy women. Glucose concentrations were assessed after fasting and every 15 min after a standardized breakfast over a 4-h lab period. Accuracy of CGM was determined by Bland-Altman plot, 15/15% sensor agreement analysis, Clarke error grid analysis (EGA) and mean absolute relative difference (MARD). In all, 429 valid CGM readings with paired venous plasma glucose (VPG) values were obtained from 29 healthy women. Mean CGM readings were 1.14 mmol/L (95% CI: 0.97 to 1.30 mmol/L, p < 0.001) higher than VPG concentrations. Ratio 95% limits of agreement were from 0.68 to 2.20, and a proportional bias (slope: 0.22) was reported. Additionally, 45% of the CGM readings were within ±0.83 mmol/L (±15 mg/dL) or ±15% of VPG, while 85.3% were within EGA Zones A + B (clinically acceptable). MARD was 27.5% (95% CI: 20.8, 34.2%), with higher MARD values in the hypoglycaemia range and when VPG concentrations were falling. The FreeStyle Libre2TM CGM system tends to overestimate glucose concentrations compared to venous plasma samples in healthy women, especially during hypoglycaemia and during glycaemic swings.

Clinical Use of Continuous Glucose Monitoring in Critically Ill Pediatric Patients with Diabetic Ketoacidosis.

Background: The use of continuous glucose monitoring (CGM) in pediatric patients with diabetic ketoacidosis (DKA) remains investigational, and data on its accuracy in pediatric intensive care units (PICU) are limited. This study evaluated the accuracy of three CGM devices in pediatric patients with DKA in the PICU. Methods: We compared 399 matched pairs of CGM and point-of-care capillary glucose (POC) values and grouped patients based on whether they changed their CGM sensor during their PICU stay. Results: Eighteen patients with a mean age of 10.98 ± 4.20 years were included, with three patients in the sensor change group. The overall mean absolute relative difference (MARD) was 13.02%. The Medtronic Guardian Sensor 3 (n = 331), Dexcom G6 (n = 41), and Abbott FreeStyle Libre 1 (n = 27) showed MARD values of 13.40%, 11.12%, and 11.33%, respectively. The surveillance error grid (SEG), Bland-Altman plot, and Pearson's correlation coefficient demonstrated satisfactory clinical accuracy of the CGM devices (SEG zones A and B, 98.5%; mean difference, 15.5 mg/dL; Pearson's correlation coefficient [r2], 0.76, P < 0.0001). MARD was significantly lower in subjects who did not experience a sensor change (11.74% vs. 17.31%, P = 0.048). Also, a statistically significant negative correlation was found between serum bicarbonate levels and POC-CGM values (r = -0.34, P < 0.001). Conclusions: The severity of DKA has a major effect on reducing the accuracy of the CGM, especially during the first several days in the intensive care unit. The reduced accuracy appears to be related to acidosis, as reflected in the serum bicarbonate levels.

A head-to-head comparison between Guardian Connect and FreeStyle Libre systems and an evaluation of user acceptability of sensors in patients with type 1 diabetes.

A user-calibrated real-time continuous glucose monitoring (rt-CGM) system is compared to a factory-calibrated flash glucose monitoring (FGM) system and assessed in terms of accuracy and acceptability in patients with type 1 diabetes (T1D). Ten participants with T1D were enroled from a specialist diabetes centre in Singapore and provided with the Guardian Connect with Enlite Sensor (Medtronic, Northridge, CA, USA) and first-generation Freestyle Libre System (Abbott Diabetes Care, Witney, UK), worn simultaneously. Participants had to check capillary blood glucose four times per day. At the end of week 1 and week 2, participants returned for data download and were given a user evaluation survey. Accuracy evaluation between Guardian Connect and Freestyle Libre includes the overall mean absolute relative difference value (9.7±11.0% vs. 17.5±10.9%), Clarke Error Grid zones A+B (98.6% vs. 98.1%), sensitivity (78.9% vs. 63.4%), and specificity (93.4% vs. 81.0%). Notably, time below range (<3.9mmol/L) was 10.5% for FGM versus 2% for rt-CGM. From the evaluation survey, 90% of participants perceived rt-CGM to be accurate versus 40% for FGM, although the majority found both devices to be easy to use, educational, and useful in improving glycaemic control. However, due to the cost of sensors, only 30% were keen to use either device for continuous monitoring. Although rt-CGM was superior to FGM in terms of accuracy, the value of glucose trends in both devices is still useful in diabetes self-management. Patients and clinicians may consider either technology depending on their requirements.

An evaluation of remote sensing algorithms for the estimation of diffuse attenuation coefficients in the ultraviolet bands.

In this study, six algorithms (both empirical and semi-analytical) developed for the estimation of Kd in the ultraviolet (UV) domain (specifically 360, 380, and 400 nm) were evaluated from a dataset of 316 stations covering oligotrophic ocean and coastal waters. In particular, the semi-analytical algorithm (Lee et al. 2013) used remote sensing reflectance in these near-blue UV bands estimated from a recently developed deep learning system as the input. For Kd(380) in a range of 0.018 - 2.34 m-1, it is found that the semi-analytical algorithm has the best performance, where the mean absolute relative difference (MARD) is 0.19, and the coefficient of determination (R2) is 0.94. For the empirical algorithms, the MARD values are 0.23-0.90, with R2 as 0.70-0.92, for this evaluation dataset. For a VIIRS and in situ matchup dataset (N = 62), the MARD of Kd(380) is 0.21 (R2 as 0.94) by the semi-analytical algorithm. These results indicate that a combination of deep learning system and semi-analytical algorithms can provide reliable Kd(UV) for past and present satellite ocean color missions that have no spectral bands in the UV, where global Kd(UV) products are required for comprehensive studies of UV radiation on marine primary productivity and biogeochemical processes in the ocean.

Non-invasive determination of blood glucose level using narrowband microwave sensor

The diabetic patients use invasive technique to monitor their Blood Glucose Level (BGL) on a regular basis. This paper mainly presents a system that estimates BGL noninvasively using microwave techniques. A narrowband microstrip antenna resonating at 1.3 GHz is used as a microwave sensor. If human finger containing a specific value of the BGL is placed on a radiating patch of narrowband microstrip patch antenna, microwave sensor structure, then the near field of this radiating patch antenna structure gets interact with human finger and results change in electrical characteristics of the antenna. These electric changes are in relation to blood permittivity changes due to variation in BGL value. The change in electric characteristics of narrow band antenna microwave structure results corresponding frequency shift. The dataset of more than 200 individuals are generated where corresponding frequency shifts for various BGLs are recorded. The prepared dataset provided linear relationship between reference BGL and corresponding frequency shift. Regression analysis for BGL estimation resulted coefficient of determination with value 0.75. The significant improvement in this coefficient of determination is obtained by subband frequency analysis, which resulted in close approximation with expected value. The Surveillance Error Grid (SEG) and Mean Absolute Relative Difference (MARD) analyses are performed on prepared dataset to validate and verify clinical acceptance of microwave based non-invasive BGL estimation system. The developed antenna model has the MARD value of 4.204% and existing split ring resonator non-invasive antenna has 12.5% MARD value for the whole dataset.

Variation of Mean Absolute Relative Differences of Continuous Glucose Monitoring Systems Throughout the Day.

There is an increasing use of continuous glucose monitoring (CGM) by people with diabetes. Measurement performance is often characterized by the mean absolute relative difference (MARD). However, MARD is influenced by a number of factors and little is known about whether MARD is stable throughout the day. A total of 24 participants with type 1 diabetes were enrolled in the study. The study was performed for seven in-patient days. Participants wore two CGM systems in parallel and performed additional frequent blood glucose (BG) measurements. On two days, glucose excursions were induced.MARD was calculated between pairs of CGM and BG values, with BG values serving as reference values. ARD values calculated from CGM-BG pairs were grouped by hour of the day. Results were analyzed separately for glucose excursion days and for regular days. Total MARDs for the complete study duration were 12.5% ± 3.6% and 13.2% ± 2.4% (n = 24). Throughout the day marked variability of MARD was observed (8.0% ± 1.3%-16.3% ± 2.9% (G5); 9.1% ± 1.4%-16.3% ± 5.3% (FL), up to n = 157 each). Low(est) MARD values were observed before breakfast and dinner, when subjects were in or near a fasting state. Especially after breakfast and lunch, MARD values were higher than average. Analytical performance of the two CGM systems, assessed by MARD, was found to vary markedly throughout the day. Activities of daily life likely triggered these variations. An increasing number of CGM users base therapeutic decisions on CGM values, and they should be aware of these variations of performance throughout the day.

The individualized statistical analysis of the continuous glucose monitoring data

Background: Continuous glucose monitoring (CGM) has shown its benefits in pregnant women with diabetes. Flash glucose monitoring (FGM), as one of the CGM types, has not been well assessed in this patient group. The interpretation of a big volume of information on glycaemia obtained with various CGM devices is possible with statistical analysis according to the algorithms proposed by manufacturers. While these algorithms cannot be comprehensive, evaluation of alternative approaches to the CGM data statistical analysis and comparison of the results obtained with different devices seem reasonable. No unified algorithm for modification of antidiabetic treatment according to the CGM results has been yet developed. This study was performed in a pregnant patient with type 1 diabetes mellitus (T1DM) to demonstrate the methods to individualized analysis of the data from various devices (CGM, FGM, glucometer) that could be used in routine clinical practice.Aim: To evaluate the individual advantages and disadvantages of the simultaneous use of FGM, CGM and SMBG in a pregnant woman with type 1 diabetes.Materials and methods: This was an observational case study with a retrospective assessment of the patient's data obtained with FGM, CGM and a glucometer in a 31-year female patient with T1DM of 6-year duration and 9 weeks of gestation, who had been on pump insulin therapy for one year and had an HbA1c level of 5.4%. During the study the patient continued her pump therapy and performed blood glucose self-monitoring (BGSM) and simultaneously used FGM and CGM. The following FGM data were compared with CGM and glucometer results: measurement numbers, time in range, mean daily glucose, mean absolute difference (MAD), and mean absolute relative difference (MARD).Results: The FGM-derived mean daily glucose was lower than that measured with the glucometer: 5.1±1.9 mmol/L vs 6.4±2.2 mmol/L (p&lt;0.001). The number of measurements with FGM was 32.0±12.9 times daily and with a glucometer 15.1±5.5 times daily (p&lt;0.001). MAD values were minimal in the hypoglycemic range (0.5±0.3 mmol/L) and maximal in the hyperglycemic range (1.6±1.2 mmol/L, р&lt;0.001). The MARD values were significantly smaller in the hyperglycemic than in the normoglycemic (16.6±12.6% vs 21.3±14.0%, р=0.035). The highest MAD and MARD were observed on the Day 1 of the sensor installation. The comparison of FGM and the glucometer readings with the Clarke consensus error grid showed that 82% of the FGM readings were in zone A or B. The FGM accuracy was higher from Day 2 to Day 9 (72.5% of the FGM readings in zone A). MAD between FGM and CGM readings was not different from that between FGM and the glucometer: 1.3±1.0 mmol/L and 1.2±0.9 mmol/L, respectively (p=0.09). MARD for the FGM and CGM comparison was higher than that for FGM and glucometer comparison: 24.4±23.0% and 18.8±13.5%, respectively (р&lt;0.001). The Pearson's correlation coefficient FGM and CGM seemed lower than that between FGM and the glucometer (0.837 and 0.889, respectively). FGM has identified more hypoglycemic events compared to CGM: time below range was 29.4% and 8.8%, respectively, p&lt;0.001).Conclusion: The FGM readings highly correlate with the glucometer. The FGM difference with the glucometer was lower in the hypo- and hyperglycemic ranges. FGM shows higher values for time below range than CGM. It is necessary to continue the study of the clinical acceptability of FGM in pregnant women and determination of its optimal regimen for the treatment of this patient category, as well as to develop an algorithm for treatment modification based on the results of FGM.

Evaluation of continuous flash glucose monitoring in a pediatric ICU setting.

Glucose monitoring is of great importance among patients in intensive care units (ICU). The purpose of this study is to assess the performance of a new flash glucose monitoring (FGM) system in a pediatric ICU setting. Sixteen consecutive patients admitted in pediatric ICU aged > 4years, expected length stay > 2days and with no medication or existing diagnosis affecting glucose metabolism were enrolled. FreeStyle Libre sensor was applied to the upper arm of the patients (8 boys). FGM measurements were compared to 3 "references": arterial blood gas analysis, capillary blood analysis and biochemical serum analysis. Mean age of patients was 8.03 ± 2.91years. Sensors remained in situ for a median of 9.71 ± 5.35days. Removal of the sensor was mainly attributed to the completion of the predefine life-span of the sensor or discharge of the patient from the ICU. We compared 711 pairs of measurements between the sensor and other glucose measurement methods. Glucose values from the sensor were consistently lower with mean absolute relative difference (MARD) being 28.34%, 25.11% and 18.99% compared to the blood gas analyzer, capillary blood glucose meter, and biochemical serum analysis, respectively, but a wide interindividual variability. Significant linear correlations between age and MARD values were observed. Surveillance error grid (SEG) analysis showed 92.04%, 94.67% and 95.52% of the readings in the none or slight risk zone respectively. FreeStyle Libre is well tolerated although not adequately accurate with a tendency to underestimate glucose levels in critically ill pediatric patients.

A Prospective Multicenter Evaluation of the Accuracy and Safety of an Implanted Continuous Glucose Sensor: The PRECISION Study.

Background: A prior study (PRECISE II) demonstrated that an implantable continuous glucose monitoring (CGM) system (Eversense® CGM System) provided accurate glucose readings through the 90-day sensor life with a favorable safety profile in participants with type 1 or type 2 diabetes (T1D, T2D). This study was performed to further characterize the accuracy of the system.Methods: PRECISION was a prospective multicenter study that evaluated the accuracy and safety of Eversense among adults with T1D or T2D through 90 days (NCT02647905). Accuracy measures included percentage system agreement and mean absolute relative difference (MARD) between Eversense and Yellow Springs Instrument reference measurements from 40 to 400 mg/dL. The primary safety endpoint was incidence of device-related or sensor insertion/removal procedure-related serious adverse events (SAEs) through 90 days. An updated glucose calculation algorithm was also applied to the sensor data from the PRECISE II study to evaluate consistency of accuracy results.Results: Thirty-five participants received the CGM system. Eighty-five percent of CGM values were within 15/15% of reference and the MARD value against reference was 9.6% (95% confidence interval [CI]: 8.9–10.4). All sensors were functional through day 90. No device- or procedure-related SAEs occurred. Application of the updated algorithm to PRECISE II sensor data resulted in 87% of readings within 15/15% of reference and an MARD value against reference of 8.5% (95% CI: 8.0%–9.1%).Conclusions: PRECISION corroborated prior accuracy and safety findings of the Eversense CGM System through the 90-day sensor life. The updated algorithm improved accuracy of measurements in PRECISE II.

Seven-Year Clinical Surveillance Program Demonstrates Consistent MARD Accuracy Performance of a Blood Glucose Test Strip.

MARD (mean absolute relative difference) is increasingly used to describe performance of glucose monitoring systems, providing a single-value quantitative measure of accuracy and allowing comparisons between different monitoring systems. This study reports MARDs for the OneTouch Verio® glucose meter clinical data set of 80 258 data points (671 individual batches) gathered as part of a 7.5-year self-surveillance program Methods: Test strips were routinely sampled from randomly selected manufacturer's production batches and sent to one of 3 clinic sites for clinical accuracy assessment using fresh capillary blood from patients with diabetes, using both the meter system and standard laboratory reference instrument. Evaluation of the distribution of strip batch MARD yielded a mean value of 5.05% (range: 3.68-6.43% at ±1.96 standard deviations from mean). The overall MARD for all clinic data points (N = 80 258) was also 5.05%, while a mean bias of 1.28 was recorded. MARD by glucose level was found to be consistent, yielding a maximum value of 4.81% at higher glucose (≥100 mg/dL) and a mean absolute difference (MAD) of 5.60 mg/dL at low glucose (<100 mg/dL). MARD by year of manufacture varied from 4.67-5.42% indicating consistent accuracy performance over the surveillance period. This 7.5-year surveillance program showed that this meter system exhibits consistently low MARD by batch, glucose level and year, indicating close agreement with established reference methods whilste exhibiting lower MARD values than continuous glucose monitoring (CGM) systems and providing users with confidence in the performance when transitioning to each new strip batch.

Accuracy and Longevity of an Implantable Continuous Glucose Sensor in the PRECISE Study: A 180-Day, Prospective, Multicenter, Pivotal Trial

It is known that continuous glucose monitoring (CGM) systems can lower mean glucose compared with episodic self-monitoring of blood glucose. Implantable CGM systems may provide additional benefits. We studied the Eversense (Senseonics Inc.) implantable CGM sensor in 71 participants aged 18 years and older with type 1 and type 2 diabetes in a 180-day multinational, multicenter pivotal trial. Participants used the CGM system at home and in the clinic. CGM accuracy was assessed during eight in-clinic visits with the mean absolute relative difference (MARD) for venous reference glucose values >4.2 mmol/L as the primary end point. Secondary end points included Clarke Error Grid Analysis and alarm performance. The primary safety outcome was device-related serious adverse events. This trial is registered with ClinicalTrials.gov, number NCT02154126. The MARD value against reference glucose values >4.2 mmol/L was 11.1% (95% CI 10.5, 11.7). Clarke Error Grid Analysis showed 99.2% of samples in the clinically acceptable error zones A and B. Eighty-one percent of hypoglycemic events were detected by the CGM system within 30 min. No device-related serious adverse events occurred during the study. Our results indicate the safety and accuracy of this new type of implantable CGM system and support it as an alternative for transcutaneous CGM.

Significance and Reliability of MARD for the Accuracy of CGM Systems.

There is a need to assess the accuracy of continuous glucose monitoring (CGM) systems for several uses. Mean absolute relative difference (MARD) is the measure of choice for this. Unfortunately, it is frequently overlooked that MARD values computed with data acquired during clinical studies do not reflect the accuracy of the CGM system only, but are strongly influenced by the design of the study. Thus, published MARD values must be understood not as precise values but as indications with some uncertainty. Data from a recent clinical trial, Monte Carlo simulations, and assumptions about the error distribution of the reference measurements have been used to determine the confidence region of MARD as a function of the number and the accuracy of the reference measurements. The uncertainty of the computed MARD values can be quantified by a newly introduced MARD reliability index (MRI), which independently mirrors the reliability of the evaluation. Thus MARD conveys information on the accuracy of the CGM system, while MRI conveys information on the uncertainty of the computed MARD values. MARD values from clinical studies should not be used blindly but the reliability of the evaluation should be considered as well. Furthermore, it should not be ignored that MARD does not take into account the key feature of CGM sensors, the frequency of the measurements. Additional metrics, such as precision absolute relative difference (PARD) should be used as well to obtain a better evaluation of the CGM performance for specific uses, for example, for artificial pancreas.

The Quantitative Relationship Between ISO 15197 Accuracy Criteria and Mean Absolute Relative Difference (MARD) in the Evaluation of Analytical Performance of Self-Monitoring of Blood Glucose (SMBG) Systems.

The relationship between International Organization for Standardization (ISO) accuracy criteria and mean absolute relative difference (MARD), 2 methods for assessing the accuracy of blood glucose meters, is complex. While lower MARD values are generally better than higher MARD values, it is not possible to define a particular MARD value that ensures a blood glucose meter will satisfy the ISO accuracy criteria. The MARD value that ensures passing the ISO accuracy test can be described only as a probabilistic range. In this work, a Bayesian model is presented to represent the relationship between ISO accuracy criteria and MARD. Under the assumptions made in this work, there is nearly a 100% chance of satisfying ISO 15197:2013 accuracy requirements if the MARD value is between 3.25% and 5.25%.

Performance Evaluation of Three Blood Glucose Monitoring Systems Using ISO 15197: 2013 Accuracy Criteria, Consensus and Surveillance Error Grid Analyses, and Insulin Dosing Error Modeling in a Hospital Setting.

Blood glucose monitoring is an essential component of diabetes management. Inaccurate blood glucose measurements can severely impact patients' health. This study evaluated the performance of 3 blood glucose monitoring systems (BGMS), Contour® Next USB, FreeStyle InsuLinx®, and OneTouch® Verio™ IQ, under routine hospital conditions. Venous blood samples (N = 236) obtained for routine laboratory procedures were collected at a Spanish hospital, and blood glucose (BG) concentrations were measured with each BGMS and with the available reference (hexokinase) method. Accuracy of the 3 BGMS was compared according to ISO 15197:2013 accuracy limit criteria, by mean absolute relative difference (MARD), consensus error grid (CEG) and surveillance error grid (SEG) analyses, and an insulin dosing error model. All BGMS met the accuracy limit criteria defined by ISO 15197:2013. While all measurements of the 3 BGMS were within low-risk zones in both error grid analyses, the Contour Next USB showed significantly smaller MARDs between reference values compared to the other 2 BGMS. Insulin dosing errors were lowest for the Contour Next USB than compared to the other systems. All BGMS fulfilled ISO 15197:2013 accuracy limit criteria and CEG criterion. However, taking together all analyses, differences in performance of potential clinical relevance may be observed. Results showed that Contour Next USB had lowest MARD values across the tested glucose range, as compared with the 2 other BGMS. CEG and SEG analyses as well as calculation of the hypothetical bolus insulin dosing error suggest a high accuracy of the Contour Next USB.

Role of Interleukin-1/Interleukin-1 Receptor Antagonist Family of Cytokines in Long-Term Continuous Glucose Monitoring In Vivo

Glucose-sensor-induced tissue reactions (e.g., inflammation and wound healing) are known to negatively impact sensor function in vivo. The roles of cytokine networks in controlling these tissue reactions (i.e., sensor biofouling) is not understood. In the present study, we investigated the role of interleukin-1 receptor antagonist (IL-1Ra), a key anti-inflammatory antagonist of the proinflammatory interleukin-1 cytokines [i.e. interleukin-1 (IL-1) alpha and IL-1 beta] in controlling continuous glucose monitoring (CGM). To investigate the role of IL-1Ra in long-term CGM in vivo, we compared CGM in transgenic mice that overexpress IL-1Ra [interleukin-1 receptor antagonist overexpresser (IL-1Ra~OE), B6.Cg-Tg(IL1rn)1Dih/J] or are deficient in IL-1Ra [interleukin-1 receptor antagonist knockout (IL-1Ra~KO), B6.129S-IL1rn(tm1Dih)/J] with mice that have normal levels of IL-1Ra (C57BL/6) over a 28-day time period. Mean absolute relative difference (MARD) analysis of CGM results among the mice of varying IL-1Ra levels demonstrated that during the first 21 days, IL-1~KO mice had the greatest tissue inflammation and the poorest sensor performance (i.e., higher MARD values) when compared with normal or IL-1Ra~OE mice. By 28 days post-sensor implantation, the inflammatory reactions had subsided and were replaced by varying degrees of fibrosis. These data support our hypothesis on the importance of the IL-1 family of agonists and antagonists in controlling tissue reactions and sensor function in vivo. These data also suggest that local delivery of IL-1Ra genes or recombinant proteins (anakinra) or other IL-1 antagonists such as antibodies or soluble IL-1 receptors would suppress sensor-induced tissue reactions and likely enhance glucose sensor function by inhibiting inflammation and wound healing at sensor implantation sites.