Clarke Error Grid Zones Research Articles

Accuracy of a Continuous Glucose Monitor in the Intensive Care Unit: A Proposed Accuracy Standard and Calibration Protocol for Inpatient Use.

Background and Aims: Guidelines now recommend inpatient continuous glucose monitor (CGM) use with confirmatory blood glucose measurements. However, the Food and Drug Administration has not yet officially approved CGM for inpatient use in large part because its accuracy has not been established in this setting. We tested the accuracy of the Dexcom G6 (G6) in 28 adults on an insulin infusion in a medical-surgical intensive care unit with 1064 matched CGM and arterial point-of-care pairs. Methods: The participants were on average 57.29 (SD 2.39) years, of whom 13 had a prior diagnosis of diabetes and 14 were admitted for a surgical diagnosis. The first 19 participants received the G6 without calibration and had a mean absolute relative difference (MARD) of 13.19% (IQR 5.11, 19.03) across 659 matched pairs, which just meets the critical care expert recommendation of MARD <14%. We then aimed to improve accuracy for the subsequent 9 participants using a calibration protocol. Results: The MARD for calibrated participants was 9.65% (3.03, 13.33), significantly lower than for uncalibrated participants (P < 0.001). Calibration also demonstrated excellent safety with 100% of values within the Clarke Error Grid zones A and B compared with 99.07% without calibration. Our protocol achieved the lowest MARD and safest CEG profile in the critical care setting and well exceeds the critical care expert recommendations. Our large sample of heterogenous critically ill patients also reached comparable accuracy to the MARD of 9% for G6 in outpatients. We believe our calibration protocol will allow G6 to be used with sufficient accuracy in inpatients.

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 <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 <14% after 36 h. We observed four episodes of interrupted sensor readings due to signal loss, ranging from 5 to 20 min. We observed no skin reaction 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 h after CGM start.

Real-world Accuracy of CGM in Inpatient Critical and Noncritical Care Settings at a Safety-Net Hospital.

We sought to determine real-world accuracy of inpatient continuous glucose monitoring (CGM) at multiple levels of acuity in a large safety-net hospital. We analyzed records from hospitalized patients on Dexcom G6 CGM, including clinical, point of care (POC), and laboratory (Lab) glucose, and CGM data. POC/Lab values were matched to the closest timed CGM value. Encounters were divided into not critically ill (NCI) versus critically ill (CI). CGM accuracy was evaluated. Paired readings (2,744 POC-CGM; 3,705 Lab-CGM) were analyzed for 233 patients with 239 encounters (83 NCI, 156 CI). POC-CGM aggregated and average mean absolute relative differences (MARD) were 15.1% and 17.1%. Lab-CGM aggregated and average MARDs were 11.4% and 12.2%. Accuracy for POC-CGM and Lab-CGM was 96.5% and 99.1% in Clarke Error Grid zones A/B. Real-world accuracy of inpatient CGM is acceptable for NCI and CI patients. Further exploration of conditions associated with lower CGM accuracy in real-world settings is warranted.

229-OR: A Calibration Protocol for Continuous Glucose Monitor Accuracy in the ICU

Guidelines now recommend inpatient continuous glucose monitor (CGM) use, assuming that it is sufficiently accurate due to its apparent safety. The FDA issued a non-objection letter for inpatient CGM use, but it is not yet officially approved for hospitals, including in the ICU. To further study CGM in the ICU, we tested the accuracy of the Dexcom G6 (G6) in 32 adults in a medical-surgical ICU in Vancouver, BC using a calibration protocol in 1035 matched CGM and arterial point-of-care (POC) pairs. Our first uncalibrated study had a mean absolute relative differential (MARD) of 13.24%, which barely meets the critical care expert recommendation of MARD &amp;lt;14%. This study aims to show a higher accuracy can be achieved with a simple calibration protocol on Day 1, then none thereafter. The MARD for calibrated patients was 9.27%, significantly lower than 13.19% for uncalibrated patients (p&amp;lt;0.001). Calibration also had excellent safety with 100% of values within Clarke Error Grid Zones A and B compared to 99.07% without calibration. Our protocol achieved the lowest MARD and safest Clarke Error Grid profile of any ICU study comparing CGM to standard of care arterial POC and well exceeds the critical care expert recommendations. Our large sample of heterogenous critically ill patients also reached comparable accuracy to the MARD of 9% for G6 in outpatients. We believe our calibration protocol will allow CGM to be used with sufficient accuracy in the ICU. Disclosure S.Bann: None. J.C.Hercus: None. P.Atkins: None. A.Alkhairy: None. D.M.Thompson: None.

130-LB: GWave—A Novel Noninvasive Glucose Blood Monitoring Radio Wave–Based Technology

Attempts to achieve an accurate noninvasive glucose monitor to date have been unsuccessful. The GWave device (Hagar, Israel) filters out the “white noise” and uses radiofrequency that detects capillary blood glucose across the physiological range. Measuring blood as opposed to interstitial fluid glucose, allows elimination of the “lag time” and poor accuracy seen in the first day of the current invasive continuous glucose monitors. Calibration of the GWave device is not required. Original data with the first prototype device revealed a 96% and 97% zone A (N = 53) in the Clarke Error Grid (CEG) comparing to venous and capillary blood, respectively. Initial results with second generation devices in non-diabetic subjects compared to capillary blood noted 99% in CEG Zone A and a MARD of 7.1% (N = 120 measurements) (Figure). These data were reproducible with two GWave devices (MARD 5.1%). Combining first and second-generation devices (N = 43 measurements) with a range of glucose of 69 to 400 mg/dL), CEG in Zone A was 98%. Subjects with both type 1 and type 2 diabetes are now being studied to examine accuracy at a wider physiological range of glycemia. Early results from the GWave glucose monitor shows promise as the first noninvasive technology. The current on-going studies will provide more insight for its use in diabetes care. Disclosure I. B. Hirsch: Consultant; Abbott Diabetes, Lifecare, Inc., Hagar, Research Support; Beta Bionics, Inc., Insulet Corporation, Dexcom, Inc. A. Navon: Consultant; GWave. Funding Hagar Technologies

Performance of the Dexcom G6 Continuous Glucose Monitoring System During Cardiac Surgery Using Hypothermic Extracorporeal Circulation.

Continuous glucose monitoring (CGM) may be challenged by extreme conditions during cardiac surgery using hypothermic extracorporeal circulation (ECC). We evaluated the Dexcom G6 sensor in 16 subjects undergoing cardiac surgery with hypothermic ECC, of whom 11 received deep hypothermic circulatory arrest (DHCA). Arterial blood glucose, quantified by the Accu-Chek Inform II meter, served as reference. Intrasurgery mean absolute relative difference (MARD) of 256 paired CGM/reference values was 23.8%. MARD was 29.1% during ECC (154 pairs) and 41.6% immediately after DHCA (10 pairs), with a negative bias (signed relative difference: -13.7%, -26.6%, and -41.6%). During surgery, 86.3% pairs were in Clarke error grid zones A or B and 41.0% of sensor readings fulfilled the International Organization for Standardization (ISO) 15197:2013 norm. Postsurgery, MARD was 15.0%. Cardiac surgery using hypothermic ECC challenges the accuracy of the Dexcom G6 CGM although recovery appears to occur thereafter.

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.

Performance of a factory‐calibrated, real‐time continuous glucose monitoring system during elective abdominal surgery

We assessed the performance of the factory-calibrated, sixth-generation continuous glucose monitoring (CGM) system Dexcom G6® (DexCom Inc., San Diego, California) during elective abdominal surgery. Twenty adults with (pre)diabetes undergoing abdominal surgery (>2 hours; 15 men, age 69 ± 13 years, glycated haemoglobin 53 ± 14 mmol/mol) wore the sensor from 1 week prior to surgery until hospital discharge. From induction of anaesthesia until 2 hours post-surgery, reference capillary glucose values were obtained every 20 minutes using the Accu-Chek® Inform II meter (Roche Diabetes Care, Mannheim, Germany). The primary endpoint was the mean absolute relative difference (ARD) between sensor and reference method during this period. In total, 1207 CGM/reference pairs were obtained. In the peri-operative period (523 pairs), mean ± SD and median (interquartile range [IQR]) ARD were 12.7% ± 8.7% and 9.9 (6.3;15.9)%, respectively, and 67.4% of sensor readings were within International Organization of Standardization 15197:2013 limits. CGM overestimated reference glucose by 1.1 ± 0.8 mmol/L (95% limits of agreement -0.5;2.7 mmol/L). Clarke error grid zones A or B contained 99.2% of pairs (A: 78.8%; B: 20.4%). The median (IQR) peri-operative sensor availability was 98.6 (95.9;100.0)%. No clinically significant adverse events occurred. In conclusion, the Dexcom G6 device showed consistent and acceptable accuracy during elective abdominal surgery, opening new avenues for peri-operative glucose management.

86-LB: Accuracy of a Factory-Calibrated, Real-Time Continuous Glucose Monitoring (CGM) System during Elective Abdominal Surgery

Aim: To assess the accuracy of the Dexcom G6 CGM system during elective abdominal surgery. Methods: Twenty adults with (pre-)diabetes scheduled for abdominal surgery (&amp;gt;2h) (5 females, age 69±13 yrs, BMI 27.3±6.6 kg/m2, HbA1c 7.0±1.3%) wore the Dexcom G6 sensor from 7±2 days preoperatively until hospital discharge. Reference capillary glucose values were obtained every 20min from induction of anaesthesia until 2h post-surgery using the Accu-Chek® Inform II meter. The primary endpoint was the mean absolute relative difference (ARD) between sensor and reference method. Results: Mean±SD operating time was 306±170min and included pancreas/liver (45%), upper GI (10%), lower GI (40%) and kidney (5%) surgeries. The mean±SD and median [interquartile range] ARD of 523 paired CGM/reference values were 12.7±8.7% and 10.4% [4.3%, 19.7%], respectively. A total of 99.4% pairs were in Clarke error grid Zones A or B (A, 76.5%, B 22.9%, Figure 1). Overall, 65.8% of sensor readings were within the limits specified by the ISO 15197:2013 norm. Sensor readings were available for 98.6% [95.9%, 100%] of the perioperative time. Intraoperative sensor replacement was required in one instance. There were no adverse device effects. Conclusion: Perioperative use of the Dexcom G6 system showed satisfying accuracy, opening new avenues for perioperative diabetes care. Disclosure A. Tripyla: None. D. Herzig: None. C. Nakas: None. J. Dehais: None. A. Andreou: None. F. Amiet: None. B. Gloor: None. A. Vogt: None. L. Bally: None. Funding Dexcom, Inc. (HOS-2019-005)

Clinically Accurate Prediction of Glucose Levels in Patients with Type 1 Diabetes.

Background and Aims: Accurate prediction of glucose levels in patients with type 1 diabetes mellitus (T1DM) is critical both for their glycemic control and for the development of closed-loop systems. Methods: In this study, we utilized real-life, retrospective, continuous glucose monitoring data from 141 T1DM patients (9,083 connection days, 1,592,506 glucose measurements) and in silico data generated by the UVA/Padova T1DM simulator to evaluate different computational methods for glucose prediction. We evaluated the performance of the models using both measures of numerical accuracy, measured by the root mean square error, and clinical accuracy, measured by the percentage of time in each of the Clarke error grid (CEG) zones, and compared the predictions done by autoregressive (AR) models, tree-based methods, artificial neural networks, and a novel model that we devised and optimized for this task. Results: Our novel model, constructed on real-life data, achieved clinical accuracy of 99.3% and 95.8% in predicting the glucose level 30 and 60 min ahead, respectively, and reduced the percentage of glucose predictions in zones C-E of the CEG by 60.6% and 38.4% in these prediction horizons, compared with a standard AR model. The model was superior to all other models across all age groups and achieved higher clinical accuracy in subgroups of patients with high glucose variability and greater time spent in hypoglycemia. Compared with real-life data, when evaluated on in silico data, the model had a higher clinical and numerical accuracy. Conclusions: A model that optimizes for CEG zones may significantly improve clinical accuracy and clinical outcomes of treatment decisions in T1DM patients. Results obtained from simulated data may overestimate the performance of models on real-life data.

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 of Subcutaneous Continuous Glucose Monitoring in Critically Ill Adults: Improved Sensor Performance with Enhanced Calibrations

Accurate real-time continuous glucose measurements may improve glucose control in the critical care unit. We evaluated the accuracy of the FreeStyle(®) Navigator(®) (Abbott Diabetes Care, Alameda, CA) subcutaneous continuous glucose monitoring (CGM) device in critically ill adults using two methods of calibration. In a randomized trial, paired CGM and reference glucose (hourly arterial blood glucose [ABG]) were collected over a 48-h period from 24 adults with critical illness (mean±SD age, 60±14 years; mean±SD body mass index, 29.6±9.3 kg/m(2); mean±SD Acute Physiology and Chronic Health Evaluation score, 12±4 [range, 6-19]) and hyperglycemia. In 12 subjects, the CGM device was calibrated at variable intervals of 1-6 h using ABG. In the other 12 subjects, the sensor was calibrated according to the manufacturer's instructions (1, 2, 10, and 24 h) using arterial blood and the built-in point-of-care glucometer. In total, 1,060 CGM-ABG pairs were analyzed over the glucose range from 4.3 to 18.8 mmol/L. Using enhanced calibration median (interquartile range) every 169 (122-213) min, the absolute relative deviation was lower (7.0% [3.5, 13.0] vs. 12.8% [6.3, 21.8], P<0.001), and the percentage of points in the Clarke error grid Zone A was higher (87.8% vs. 70.2%). Accuracy of the Navigator CGM device during critical illness was comparable to that observed in non-critical care settings. Further significant improvements in accuracy may be obtained by frequent calibrations with ABG measurements.

Real-Time Continuous Glucose Monitoring Shows High Accuracy within 6 Hours after Sensor Calibration: A Prospective Study

Accurate and timely glucose monitoring is essential in intensive care units. Real-time continuous glucose monitoring system (CGMS) has been advocated for many years to improve glycemic management in critically ill patients. In order to determine the effect of calibration time on the accuracy of CGMS, real-time subcutaneous CGMS was used in 18 critically ill patients. CGMS sensor was calibrated with blood glucose measurements by blood gas/glucose analyzer every 12 hours. Venous blood was sampled every 2 to 4 hours, and glucose concentration was measured by standard central laboratory device (CLD) and by blood gas/glucose analyzer. With CLD measurement as reference, relative absolute difference (mean±SD) in CGMS and blood gas/glucose analyzer were 14.4%±12.2% and 6.5%±6.2%, respectively. The percentage of matched points in Clarke error grid zone A was 74.8% in CGMS, and 98.4% in blood gas/glucose analyzer. The relative absolute difference of CGMS obtained within 6 hours after sensor calibration (8.8%±7.2%) was significantly less than that between 6 to 12 hours after calibration (20.1%±13.5%, p<0.0001). The percentage of matched points in Clarke error grid zone A was also significantly higher in data sets within 6 hours after calibration (92.4% versus 57.1%, p<0.0001). In conclusion, real-time subcutaneous CGMS is accurate in glucose monitoring in critically ill patients. CGMS sensor should be calibrated less than 6 hours, no matter what time interval recommended by manufacturer.

Continuous glucose monitoring: quality of hypoglycaemia detection

To evaluate the accuracy of a (widely used) continuous glucose monitoring (CGM)-system and its ability to detect hypoglycaemic events. A total of 18 patients with type 1 diabetes mellitus used continuous glucose monitoring (Guardian REAL-Time CGMS) during two 9-day in-house periods. A hypoglycaemic threshold alarm alerted patients to sensor readings <70 mg/dl. Continuous glucose monitoring sensor readings were compared to laboratory reference measurements taken every 4 h and in case of a hypoglycaemic alarm. A total of 2317 paired data points were evaluated. Overall, the mean absolute relative difference (MARD) was 16.7%. The percentage of data points in the clinically accurate or acceptable Clarke Error Grid zones A + B was 94.6%. In the hypoglycaemic range, accuracy worsened (MARD 38.8%) leading to a failure to detect more than half of the true hypoglycaemic events (sensitivity 37.5%). Furthermore, more than half of the alarms that warn patients for hypoglycaemia were false (false alert rate 53.3%). Above the low alert threshold, the sensor confirmed 2077 of 2182 reference values (specificity 95.2%). Patients using continuous glucose monitoring should be aware of its limitation to accurately detect hypoglycaemia.

Benefits of Blinded Continuous Glucose Monitoring during a Randomized Clinical Trial

Real-time, personal continuous glucose monitoring (CGM) is a validated technology that can help patients improve glycemic control. Blinded CGM is a promising technology for obtaining retrospective data in clinical research where the quantity and quality of blood glucose information is important. This study was designed to investigate the use of novel procedures to enhance data capture from blinded CGM. Following a 4-week run-in, 46 patients with type 1 diabetes were randomized to one of two prandial insulins for a 12-week treatment period, after which they were crossed over to the alternate treatment for 12 weeks. Continuous glucose monitoring was implemented at the end of run-in (practice only) and during the last 2 weeks of each treatment period. Eighty percent of 288 possible daily glucose values were required for at least three days. Continuous glucose monitoring was extended for an additional week if these criteria were not met, and patients were allowed to insert sensors at home when necessary. Continuous glucose monitoring results were compared to reference eight-point self-monitoring of blood glucose (SMBG). Higher than expected sensor failure rate was approximately 25%. During run-in, 12 of 45 attempted profiles failed adequacy criteria. However, treatment periods had only 1 of 82 attempted profiles considered inadequate (6 cases required an additional week of CGM). Using SMBG as reference, 93.7% of 777 CGM values were in Clarke error grid zones A+B. With appropriate training, adequate practice, and opportunity to repeat blinded CGM as needed, nearly 100% of attempted profiles can be obtained successfully.

Continuous Glucose Monitoring in Subcutaneous Tissue Using Factory-Calibrated Sensors: A Pilot Study

Commercial continuous subcutaneous glucose monitors require in vivo calibration using capillary blood glucose tests. Feasibility of factory calibration, i.e., sensor batch characterization in vitro with no further need for in vivo calibration, requires a predictable and stable in vivo sensor sensitivity and limited inter- and intra-subject variation of the ratio of interstitial to blood glucose concentration. Twelve volunteers wore two FreeStyle Navigator (Abbott Diabetes Care, Alameda, CA) continuous glucose monitoring systems for 5 days in parallel for two consecutive sensor wears (four sensors per subject, 48 sensors total). Sensors from a prototype sensor lot with a low variability in glucose sensitivity were used for the study. Median sensor sensitivity values based on capillary blood glucose were calculated per sensor and compared for inter- and intra-subject variation. Mean absolute relative difference (MARD) calculation and error grid analysis were performed using a single calibration factor for all sensors to simulate factory calibration and compared to standard fingerstick calibration. Sensor sensitivity variation in vitro was 4.6%, which increased to 8.3% in vivo (P < 0.0001). Analysis of variance revealed no significant inter-subject differences in sensor sensitivity (P = 0.134). Applying a single universal calibration factor retrospectively to all sensors resulted in a MARD of 10.4% and 88.1% of values in Clarke Error Grid Zone A, compared to a MARD of 10.9% and 86% of values in Error Grid Zone A for fingerstick calibration. Factory calibration of sensors for continuous subcutaneous glucose monitoring is feasible with similar accuracy to standard fingerstick calibration. Additional data are required to confirm this result in subjects with diabetes.

Locally Smoothed Median Absolute Difference Curves and the First Global Performance Cooperative

Abstract We present US multicenter results, multinational performance, and a global performance cooperative. We use locally smoothed median absolute difference (LS MAD) curves to reveal bedside glucose testing performance through rapid visual discrimination and hypothesize that (a) inaccuracy and open-loop feedback systems cause glycemic variability adversely affecting outcomes in tight glycemic control (TGC) and (b) poor performance increases a risk for underserved populations where hospital and primary care staff may use glucose meter systems (GMSs) in critical therapeutic decisions. The US methods are the following: modified glucose oxidase enzyme with a multilayer-gold, multielectrode, 4-well test strip (StatStrip, NOVA Biomedical, Waltham, Mass); pragmatic comparison of GMS results versus paired plasma glucose level measurements from chemistry analyzers in clinical laboratories; venous samples (2767) at 59 US hospitals with 21 chemistry analyzer types; erroneous results identified using a Bland-Altman plot and the International Organization for Standardization 15197 criteria; and analysis of values discrepant for TGC. The global method is synopsis of LS MAD curves. The US GMS met the International Organization for Standardization 15197 tolerances: 98.86% (607/614) of observations were within tolerance for a glucose level lower than 75 mg/dL and 99.81% (2149/2153) for 75 mg/dL or higher. The mean (SD) paired difference was −2.92 (9.6) mg/dL (median, −2; range, −96 to +45). Locally smoothed MAD curve analysis revealed a satisfactory performance below 179 mg/dL; above 179 mg/dL, the recommended error tolerance limit (5 mg/dL) was not met. No discrepant values appeared. All paired observations fell within Clarke error grid zone A. Linear regression showed y = 1.019 × −0.233 mg/dL; r2 = 0.995. In contrast, LS MAD curves for other types of GMSs used outside the United States showed inconsistencies. The US LS MAD curve results presented here benchmark a high standard. We invite readers to submit data sets for analysis and to participate in a new point-of-care performance cooperative. This global forum can advance bedside glucose testing expectations by motivating the more accurate and precise devices better standardized and harmonized for use with patients who have diabetes, TGC, and low-resource settings.

Pre- and Postoperative Accuracy and Safety of a Real-Time Continuous Glucose Monitoring System in Cardiac Surgical Patients: A Randomized Pilot Study

Our objective was to evaluate the accuracy and safety of a real-time (RT) continuous glucose monitoring system (CGMS) in patients before and after cardiothoracic surgery and to investigate whether activation of the alarm function of the RT-CGMS had an effect on glucose control. Patients scheduled for elective cardiothoracic procedures, without a history of insulin-requiring diabetes, were perioperatively monitored with RT-CGMS for 72 h and were randomized into two groups: with or without the alarm function (set at 4 and 10 mmol/L) of the device activated. Sensor values were compared with capillary, arterial, and venous blood glucose values. Percentages of time spent in various glucose ranges were compared between groups. There were no adverse effects of the RT-CGMS. Of the 1,001 sensor value comparisons with capillary or arterial measurements, 96.6% fell within Clarke Error Grid zones A and B, with relative absolute differences ranging from 15% (preoperative period) to 12% (intensive care unit period) to 14% (postoperative period on the ward). Seventeen (7.9%) arterial and 16 (2.0%) capillary comparisons fell within zone D or E. Whether or not the alarm function, as used in this pilot study, was activated did not affect time spent in different glucose ranges. Although the RT-CGMS is safe and accurate according to accepted standards, there are still small aberrations, which in our opinion preclude unlimited use in its present form in a clinical setting. The effect of the alarm function at different glucose levels remains to be investigated.

Assessing the Performance of Handheld Glucose Testing for Critical Care

We assessed the performance of a point-of-care (POC) glucose meter system (GMS) with multitasking test strip by using the locally-smoothed (LS) median absolute difference (MAD) curve method in conjunction with a modified Bland-Altman difference plot and superimposed International Organization for Standardization (ISO) 15197 tolerance bands. We analyzed performance for tight glycemic control (TGC). A modified glucose oxidase enzyme with a multilayer-gold, multielectrode, four-well test strip (StatStriptrade mark, NOVA Biomedical, Waltham, MA) was used. There was no test strip calibration code. Pragmatic comparison was done of GMS results versus paired plasma glucose measurements from chemistry analyzers in clinical laboratories. Venous samples (n = 1,703) were analyzed at 35 hospitals that used 20 types of chemistry analyzers. Erroneous results were identified using the Bland-Altman plot and ISO 15197 criteria. Discrepant values were analyzed for the TGC interval of 80-110 mg/dL. The GMS met ISO 15197 guidelines; 98.6% (410 of 416) of observations were within tolerance for glucose <75 mg/dL, and for > or =75 mg/dL, 100% were within tolerance. Paired differences (handheld minus reference) averaged -2.2 (SD 9.8) mg/dL; the median was -1 (range, -96 to 45) mg/dL. LS MAD curve analysis revealed satisfactory performance below 186 mg/dL; above 186 mg/dL, the recommended error tolerance limit (5 mg/dL) was not met. No discrepant values appeared. All points fell in Clarke Error Grid zone A. Linear regression showed y = 1.018x - 0.716 mg/dL, and r2 = 0.995. LS MAD curves draw on human ability to discriminate performance visually. LS MAD curve and ISO 15197 performance were acceptable for TGC. POC and reference glucose calibration should be harmonized and standardized.

FreeStyle Mini™ Blood Glucose Results Are Accurate and Suitable for Use in Glycemic Clamp Protocols

We assessed the accuracy of the FreeStyle Mini (FSM) meter for use in glycemic clamp and meal protocols in comparison with the HemoCue Glucose 201 DM Analyzer (HemoCue) and the YSI 2300 STAT Glucose Oxidase Analyzer (YSI). Seven volunteers with type 2 diabetes mellitus, 35-69 years old, underwent a frequently sampled meal test and a graded hyperglycemic test, on two separate days, with one of the volunteers undergoing each test twice. Samples for glucose measurements were obtained from arterialized venous blood. A total of 420 samples (with glucose levels ranging from 63 to 388 mg/dl) were available for comparison. On average, 10 measurements were available for every 5 mg/dl increment in glucose level in the range of 130-310 mg/dl. Blood glucose measurements were done on each sample with the FSM, HemoCue, and YSI. FreeStyle Mini blood glucose values correlated closely with the YSI readings. Of the FSM measurements, 99.0% were within the Clarke error grid zone A; 51.3%, 84.7%, and 96.2% of the FSM readings were within 5%, 10% and 15% of the YSI values, respectively. The FSM was significantly more accurate than the HemoCue (84.7% vs 76.6% of results within 10% of the YSI results; p = .0038). The mean average relative difference of the FSM (5.8%) was also significantly lower than that of the HemoCue (6.8%; p = .0013) The FSM provides accurate results and constitutes a suitable alternative for bedside blood glucose measurements in experimental procedures, helping to reduce sample size, turnaround time, and cost.