Assessing the Accuracy of Continuous Glucose Monitoring (CGM) Calibrated With Capillary Values Using Capillary or Venous Glucose Levels as a Reference.
2016 (English)In: Journal of Diabetes Science and Technology, ISSN 1932-2968, E-ISSN 1932-2968, Vol. 10, no 4, 876-884 p.Article in journal (Refereed) Published
Background: Using the standard venous reference for the evaluation of continuous glucose monitoring (CGM) systems could possibly negatively affect measured CGM accuracy since CGM are generally calibrated with capillary glucose and venous and capillary glucose concentrations differ. We therefore aimed to quantify the effect of using capillary versus venous glucose reference samples on estimated accuracy in capillary calibrated CGM.less thanbr /greater thanMethods: We evaluated 41 individuals with type 1 diabetes mellitus (T1DM) using the Dexcom G4 CGM system over 6 days. Patients calibrated their CGM devices with capillary glucose by means of the HemoCue system. During 2 visits, capillary and venous samples were simultaneously measured by HemoCue and compared to concomitantly obtained CGM readings. The mean absolute relative difference (MARD) was calculated using capillary and venous reference samples.less thanbr /greater thanResults: Venous glucose values were 0.83 mmol/L (15.0 mg/dl) lower than capillary values over all glycemic ranges, P less than .0001. Below 4 mmol/l (72 mg/dl), the difference was 1.25 mmol/l (22.5 mg/dl), P = .0001, at 4-10 mmol/l (72-180 mg/dl), 0.67 mmol/l (12.0 mg/dl), P less than .0001 and above 10 mmol/l (180 mg/dl), 0.95 mmol/l (17.1 mg/dl), P less than .0001. MARD was 11.7% using capillary values as reference compared to 13.7% using venous samples, P = .037. Below 4 mmol/l (72 mg/dl) MARD was 16.6% and 31.8%, P = .048, at 4-10 mmol/l (72-180 mg/dl) 12.1% and 12.6%, P = .32, above 10 mmol/l (180 mg/dl) 8.7% and 9.2%, P = .82.less thanbr /greater thanConclusion: Using capillary glucose concentrations as reference to evaluate the accuracy of CGM calibrated with capillary samples is associated with a lower MARD than using venous glucose as the reference. Capillary glucose concentrations were significantly higher than venous in all glycemic ranges.less thanbr /greater than (© 2016 Diabetes Technology Society.)
Place, publisher, year, edition, pages
Diabetes Technology Society , 2016. Vol. 10, no 4, 876-884 p.
Endocrinology and Diabetes Clinical Laboratory Medicine
IdentifiersURN: urn:nbn:se:liu:diva-134874DOI: 10.1177/1932296815626724PubMedID: 26810924OAI: oai:DiVA.org:liu-134874DiVA: diva2:1077564