Lower Mean Absolute Relative Difference Research Articles

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.

952-P: The Accuracy and Feasibility of Continuous Glucose Monitoring in COVID-19 Non-Critically Ill Hospitalized Patients—A Pilot Study

Background: Optimal glycemic control is associated with favorable outcomes in COVID-19 hospitalized patients. Frequent capillary blood glucose (CBG) monitoring is difficult to perform. Using continuous glucose monitoring (CGM) system to assist glycemic care can reduce the exposure of healthcare personnel and PPE usage. However, data on the accuracy of CGM in this setting are limited. Objective: To evaluate the accuracy and feasibility of real-time CGM (rt-CGM) in non-ICU hospitalized adult COVID-19 patients. Methods: This is a single-center prospective study of non-ICU hospitalized adult patients with COVID-19 infection who had hyperglycemia requiring insulin therapy during admission. Medtronic Guardian Connect rt-CGM sensor and transmitter were placed on the patient’s abdomen. Paired CBG and sensor glucose values were analyzed for accuracy of CGM using mean absolute relative difference (MARD) and Clarke Error Grid Analysis (CEGA). Results: Fifteen patients were enrolled. Mean age was 48.6 ± 17.9 years. Thirteen patients (86.7%) had pre-existing diabetes. Mean HbA1c was 10.6± 3.6%. Mean duration of CGM use was 6 ± 1.2 days and mean calibration was 2.6 ± 0.7 times/day. There were 253 paired CBG and CGM measurements. The overall MARD was 9.9 ± 9.3%. The lowest MARD was observed in the CBG range of 70-180 mg/dl (9.6 ± 9.0%). The percentages of glucose readings within ±15%/15 mg/dL, ±20%/20 mg/dL, and ±30%/30 mg/dL were 80.2%, 89.7%, and 95.3%, respectively. A total of 99.2% of the data points were in zone A and B of CEGA, and none were in zone E. Percent time in range on day 1 was 57.9 ± 22.9 and improved to 64.9 ± 18.1 in the last 72 hours of sensor wear. No adverse events from the CGM were observed. CGM reduced POC testing by 30%. Conclusions: Rt-CGM use in hospitalized patients with COVID-19 infection demonstrates high accuracy and potentially improves glucose control, reduces the frequency of CBG testing, and preserves medical resources. Disclosure C.Sakjirapapong: None. S.Sirinvaravong: None. L.Preechasuk: None. N.Thongtang: None. Funding Specific League Funds (R016421005)

1118-P: Accuracy of Dexcom G6 Continuous Glucose Monitor in Pediatric Diabetic Ketoacidosis Admissions

Continuous glucose monitors (CGMs) are an integral part of care for youth with type 1 diabetes (T1D) though lack FDA labeling for inpatient use. While some adult data on CGM use in inpatient settings is available, pediatric data are minimal. This retrospective chart review evaluated the accuracy of Dexcom G6 CGM versus point of care (POC, Nova Biomedical StatStrip [MARD 6%])) blood glucose values from pediatric inpatient encounters. Blood glucose data, diagnosis codes, and initial labs were collected from the medical record. CGM values were obtained from Dexcom Clarity CSV files. Paired glucose values (N=1191) from 83 patients with T1D (median age 12 yrs, 54% male, 69% non-Hispanic White) were used to calculate mean absolute relative difference (MARD) and Clarke Error Grid. Data from DKA admissions (N=665) had a MARD of 11.1% with 97.8% of values within A&amp;B zones, compared to 11.4% and 98.5% for non-DKA admissions (N=526). Values from severe DKA admissions (N= 307) (pH &amp;lt;7.15 and/or bicarbonate &amp;lt;5 mmol/L) had a lower MARD compared to non-severe admissions (N=358) (8.4% vs 13.4%, p=0.01). In summary, CGM accuracy is comparable between DKA and non-DKA admissions. The accuracy of CGMs, even in severe DKA, suggests potential usability during pediatric hospital encounters. Further analysis will differentiate POC versus lab glucose and the effect of medications, including IV insulin infusions. Disclosure L.A.Waterman: None. L.Pyle: None. L.Towers: None. E.Jost: Other Relationship; Tandem Diabetes Care, Inc. A.J.Karami: None. C.Berget: Consultant; Insulet Corporation, Dexcom, Inc., Other Relationship; Tandem Diabetes Care, Inc. G.P.Forlenza: Advisory Panel; Medtronic, Consultant; Dexcom, Inc., Insulet Corporation, Tandem Diabetes Care, Inc., Lilly Diabetes, Research Support; Medtronic, Abbott, Dexcom, Inc., Insulet Corporation, Tandem Diabetes Care, Inc. R.Wadwa: Consultant; Eli Lilly and Company, Other Relationship; Dexcom, Inc., Eli Lilly and Company, Research Support; Dexcom, Inc., Eli Lilly and Company, Beta Bionics, Inc., Tandem Diabetes Care, Inc. E.C.Cobry: None. Funding National Institutes of Health (5T32DK063687)

Cross-Linkable Polymer-Based Multi-layers for Protecting Electrochemical Glucose Biosensors against Uric Acid, Ascorbic Acid, and Biofouling Interferences.

The lifetime of implantable electrochemical glucose monitoring devices is limited due to the foreign body response and detrimental effects from ascorbic acid (AA) and uric acid (UA) interferents that are components of physiological media. Polymer coatings can be used to shield biosensors from these interferences and prolong their functional lifetime. This work explored several approaches to protect redox polymer-based glucose biosensors against such interferences by designing six targeted multi-layer sensor architectures. Biological interferents, like cells and proteins, and UA and AA interferents were found to have individual effects on the current density and operational stability of glucose biosensors, requiring individual protection and treatment. Protection against biofouling can be achieved using a poly(2-methacryloyloxyethyl phosphorylcholine-co-glycidyl methacrylate) (MPC) zwitterionic polymer coating. An enzyme-scavenging approach was compared to electrostatic repulsion by negatively charged polymers for protection against AA and UA interferences. A multi-layer novel polymer design (PD) system consisting of a cross-linkable negatively charged polyvinylimidazole-polysulfostyrene co-polymer inner layer and a cross-linkable MPC zwitterionic polymer outer layer showed the best protection against AA, UA, and biological interferences. The sensor protected using the novel PD shield displayed the lowest mean absolute relative difference between the glucose reading without the interferent and the reading value with the interferent present and also displayed the lowest variability in sensor readings in complex media. For sensor measurements in artificial plasma, the novel PD extends the linear range (R2 = 0.99) of the sensor from 0-10 mM for the control to 0-20 mM, shows a smaller decrease in sensitivity, and retains high current densities. The application of PD multi-target coating improves sensor performance in complex media and shows promise for use in sensors operating in real conditions.

648-P: Lag Times in a Seventh-Generation Continuous Glucose Monitoring System

Background: Glucose transport from blood to interstitial fluid impacts performance of interstitially-placed continuous glucose monitoring (CGM) sensors. We examined time series offsets (“lag times”) of the G7 CGM system (Dexcom) . Methods: In a pivotal accuracy study conducted in the US, 316 adults with diabetes wearing 6sensors participated in clinic sessions that included glucose manipulations and frequent venous plasma glucose (YSI) determinations. G7 values were obtained every 5 min and interpolated to provide values at 1-min intervals. Each YSI value was associated with 21 G7 values obtained from min before to min after. The absolute relative difference (ARD) was calculated for each YSI:G7 pair and each sensor’s lag was estimated as the time with the lowest mean ARD. Results: A lag time of 2 minutes provided the lowest mean ARD of 8.6% and the highest %20/20 agreement rate of 94.4%. The Figure shows summary statistics for the distribution of sensor-specific lag times. Lag times of ≤0 min (where interstitial glucose leads blood glucose) were observed for 2 (33.0%) of the sensors. Conclusions: The distribution of G7 time series offsets suggests that CGM values may most accurately reflect previous or upcoming blood glucose values. This finding is likely due to technological innovations in G7’s hardware and algorithm. The minimal lag time observed should provide users with confidence in using CGM to manage their diabetes. Disclosure X.Zhang: Employee; Dexcom, Inc. J.L.Reid: Employee; Dexcom, Inc., GenMark Diagnostics. T.C.Walker: Employee; Dexcom, Inc., Stock/Shareholder; Dexcom, Inc. J.Welsh: Employee; Dexcom, Inc. A.Balo: Employee; Dexcom, Inc. Funding Dexcom, Inc.

Simple Post-Processing of Continuous Glucose Monitoring Measurements Improves Endpoints in Clinical Trials.

Continuous glucose monitoring (CGM) is a powerful tool to be considered both in clinical practice and clinical trials. However, CGM has been criticized for being inaccurate for many reasons including a physiological delay. This study sought to investigate the current delay issue and propose a simple post-processing procedure. More than a million hours of the Dexcom G4 CGM from 472 subjects investigated in a state-of-the-art clinical trial were analyzed by time shifting the CGM measurements and comparing them to plasma glucose (PG) measurements. The resultant CGM measurements were then assessed in relation to real-world clinical research endpoints. A CGM time shift of -9 minutes was optimal and reduced mean absolute relative difference (MARD) statistically significantly with 1.0% point. The MARD reduction resulted in better clinical research endpoints of hypoglycemia and postprandial glucose increments. The delay in CGM is still an issue. The delay in this study was identified to be 9 minutes compared to PG. With a simple post-processing approach of time shifting the CGM measurements with -9 minutes, it was possible to obtain a statistically significantly lower MARD and subsequently obtain clinical research endpoints of improved validity.

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.

Extensive Assessment of Blood Glucose Monitoring During Postprandial Period and Its Impact on Closed-Loop Performance

Closed-loop (CL) systems aims to outperform usual treatments in blood glucose control and continuous glucose monitors (CGM) are a key component in such systems. Meals represents one of the main disturbances in blood glucose control, and postprandial period (PP) is a challenging situation for both CL system and CGM accuracy. We performed an extensive analysis of sensor's performance by numerical accuracy and precision during PP, as well as its influence in blood glucose control under CL therapy. During PP the mean absolute relative difference (MARD) for both sensors presented lower accuracy in the hypoglycemic range (19.4 ± 12.8%) than in other ranges (12.2 ± 8.6% in euglycemic range and 9.3 ± 9.3% in hyperglycemic range). The overall MARD was 12.1 ± 8.2%. We have also observed lower MARD for rates of change between 0 and 2 mg/dl. In CL therapy, the 10 trials with the best sensor spent less time in hypoglycemia (PG < 70 mg/dl) than the 10 trials with the worst sensors (2 ± 7 minutes vs 32 ± 38 minutes, respectively). In terms of accuracy, our results resemble to previously reported. Furthermore, our results showed that sensors with the lowest MARD spent less time in hypoglycemic range, indicating that the performance of CL algorithm to control PP was related to sensor accuracy.

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%.

A Computational Method to Determine Glucose Infusion Rates for Isoglycemic Intravenous Glucose Infusion Study.

The results of the isoglycemic intravenous glucose infusion (IIGI) study need to mimic the dynamic glucose profiles during the oral glucose tolerance test (OGTT) to accurately calculate the incretin effect. The glucose infusion rates during IIGI studies have historically been determined by experienced research personnel using the manual ad-hoc method. In this study, a computational method was developed to automatically determine the infusion rates for IIGI study based on a glucose-dynamics model. To evaluate the computational method, 18 subjects with normal glucose tolerance underwent a 75 g OGTT. One-week later, Group 1 (n = 9) and Group 2 (n = 9) underwent IIGI studies using the ad-hoc method and the computational method, respectively. Both methods were evaluated using correlation coefficient, mean absolute relative difference (MARD), and root mean square error (RMSE) between the glucose profiles from the OGTT and the IIGI study. The computational method exhibited significantly higher correlation (0.95 ± 0.03 versus 0.86 ± 0.10, P = 0.019), lower MARD (8.72 ± 1.83% versus 13.11 ± 3.66%, P = 0.002), and lower RMSE (10.33 ± 1.99 mg/dL versus 16.84 ± 4.43 mg/dL, P = 0.002) than the ad-hoc method. The computational method can facilitate IIGI study, and enhance its accuracy and stability. Using this computational method, a high-quality IIGI study can be accomplished without the need for experienced personnel.

Accuracy of Continuous Glucose Monitoring During Three Closed-Loop Home Studies Under Free-Living Conditions.

Objectives: Closed-loop (CL) systems modulate insulin delivery based on glucose levels measured by a continuous glucose monitor (CGM). Accuracy of the CGM affects CL performance and safety. We evaluated the accuracy of the Freestyle Navigator® II CGM (Abbott Diabetes Care, Alameda, CA) during three unsupervised, randomized, open-label, crossover home CL studies.Materials and Methods: Paired CGM and capillary glucose values (10,597 pairs) were collected from 57 participants with type 1 diabetes (41 adults [mean±SD age, 39±12 years; mean±SD hemoglobin A1c, 7.9±0.8%] recruited at five centers and 16 adolescents [mean±SD age, 15.6±3.6 years; mean±SD hemoglobin A1c, 8.1±0.8%] recruited at two centers). Numerical accuracy was assessed by absolute relative difference (ARD) and International Organization for Standardization (ISO) 15197:2013 15/15% limits, and clinical accuracy was assessed by Clarke error grid analysis.Results: Total duration of sensor use was 2,002 days (48,052 h). Overall sensor accuracy for the capillary glucose range (1.1–27.8 mmol/L) showed mean±SD and median (interquartile range) ARD of 14.2±15.5% and 10.0% (4.5%, 18.4%), respectively. Lowest mean ARD was observed in the hyperglycemic range (9.8±8.8%). Over 95% of pairs were in combined Clarke error grid Zones A and B (A, 80.1%, B, 16.2%). Overall, 70.0% of the sensor readings satisfied ISO criteria. Mean ARD was consistent (12.3%; 95% of the values fall within ±3.7%) and not different between participants (P=0.06) within the euglycemic and hyperglycemic range, when CL is actively modulating insulin delivery.Conclusions: Consistent accuracy of the CGM within the euglycemic–hyperglycemic range using the Freestyle Navigator II was observed and supports its use in home CL studies. Our results may contribute toward establishing normative CGM performance criteria for unsupervised home use of CL.

Accuracy Evaluation of CONTOUR(®)PLUS Compared With Four Blood Glucose Monitoring Systems.

IntroductionThe aim of this study was to compare the accuracy of 5 blood glucose monitoring systems (BGMSs; CONTOUR®PLUS [CP], Accu-Chek® Active [ACA], Accu-Chek® Performa [ACP], FreeStyle Freedom™ [FF], OneTouch® SelectSimple™ [OTSS]).MethodsStudy staff tested fingerstick samples from 106 subjects aged ≥18 years using the 5 BGMSs. Some samples were modified to achieve blood glucose concentrations throughout the measuring range. The primary endpoint was comparison of the mean absolute relative difference (MARD) from the reference value (Yellow Springs Instruments [YSI]) across the overall tested glucose range. Other endpoints were MARD in the low (≤80 mg/dL [≤4.4 mmol/L]), middle (81–180 mg/dL [4.5–10.0 mmol/L]), and high (>180 mg/dL [>10.0 mmol/L]) glucose ranges, and MARD for unmodified samples in the overall glucose range.ResultsCONTOUR®PLUS had a statistically significantly lower MARD than all BGMSs across the overall tested range (27–460 mg/dL [1.5–25.5 mmol/L]) and in the high glucose range. In the low glucose range, CP had a lower MARD than all BGMSs, which was statistically significant except for ACP. For unmodified samples across the overall tested range, CP had a lower MARD than all BGMSs and was statistically significantly lower except for ACA.ConclusionsCONTOUR®PLUS had the lowest mean difference from the reference values (by MARD) when compared with other BGMSs across multiple glucose ranges with modified and unmodified samples.FundingBayer HealthCare LLC, Diabetes Care.Trial registration number: ClinicalTrials.gov Identifier NCT01714232.Electronic supplementary materialThe online version of this article (doi:10.1007/s13300-015-0121-3) contains supplementary material, which is available to authorized users.

Comparative Accuracy of 3 Blood Glucose Monitoring Systems that Communicate with An Insulin Pump

This study compared the accuracy of the Contour® Next Link blood glucose monitoring system (BGMS) with 2 other BGMSs (OneTouch® UltraLink® and Nova Max Link™), all 3 of which can communicate wirelessly with an insulin pump. Study staff tested fingerstick samples from 112 subjects aged ≥18 years with each BGMS. All BGMS results were compared to YSI 2300 STAT Plus™ Glucose Analyzer results from the same sample source. To safely obtain a wide range of glucose values, 3 blood samples were obtained from each subject (1 sample was tested fresh from the fingertip [unmodified], and the other 2 samples were modified to raise or lower the glucose concentration). The primary endpoint evaluated was difference in accuracy among the BGMSs across the entire tested glucose range (34 to 561 mg/dL). Additional comparisons were made of accuracy in the low (≤80 mg/dL) and high (>180 mg/dL) glucose ranges. BGMS accuracy was assessed by mean absolute relative difference (MARD). The Contour® Next Link BGMS had a statistically significantly lower MARD from the reference value than the other BGMSs across the entire tested glucose range, calculated using all samples as well as only unmodified (i.e., natural) samples. In the low and high glucose ranges, the Contour® Next Link also had a statistically significantly lower MARD from the reference value than the other BGMSs. As assessed by MARD, the Contour® Next Link BGMS was significantly more accurate than the other BGMSs evaluated, all of which can communicate wirelessly with an insulin pump.

Evaluation of the Performance of a Novel System for Continuous Glucose Monitoring

The performance of a continuous glucose monitoring (CGM) system in the early stage of development was assessed in an inpatient setting that simulates daily life conditions of people with diabetes. Performance was evaluated at low glycemic, euglycemic, and high glycemic ranges as well as during phases with rapid glucose excursions. Each of the 30 participants with type 1 diabetes (15 female, age 47 ± 12 years, hemoglobin A1c 7.7% ± 1.3%) wore two sensors of the prototype system in parallel for 7 days. Capillary blood samples were measured at least 16 times per day (at least 15 times per daytime and at least once per night). On two subsequent study days, glucose excursions were induced. For performance evaluation, the mean absolute relative difference (MARD) between CGM readings and paired capillary blood glucose readings and precision absolute relative difference (PARD), i.e., differences between paired CGM readings were calculated. Overall aggregated MARD was 9.2% and overall aggregated PARD was 7.5%. During induced glucose excursions, MARD was 10.9% and PARD was 7.8%. Lowest MARD (8.5%) and lowest PARD (6.4%) were observed in the high glycemic range (euglycemic range, MARD 9.1% and PARD 7.4%; low glycemic range, MARD 12.3% and PARD 12.4%). The performance of this prototype CGM system was, particularly in the hypoglycemic range and during phases with rapid glucose fluctuations, better than performance data reported for other commercially available systems. In addition, performance of this prototype sensor was noticeably constant over the whole study period. This prototype system is not yet approved, and performance of this CGM system needs to be further assessed in clinical studies.