Are point-of-care measurements of glycated haemoglobin accurate in the critically ill?

Abstract

IntroductionCritically ill patients with type 2 diabetes mellitus (T2DM) and chronic hyperglycaemia may benefit from a more liberal approach to glucose control than patients with previously normal glucose tolerance. It may therefore be useful to rapidly determine HbA1c concentrations. Point-of-care (POC) analysers offer rapid results but may be less accurate than laboratory analysis. Aim(s)The aim of this study was to determine agreement between POC and laboratory HbA1c testing in critically ill patients with T2DM. MethodsCritically ill patients with T2DM had concurrent laboratory, capillary-, and arterial-POC HbA1c measurements performed. Data are presented as mean (standard deviation) or median [interquartile range]. Measurement agreement was assessed by Lin's concordance correlation coefficient, Bland–Altman 95% limits of agreement, and classification by Cohen's kappa statistic. ResultsHbA1c analysis was performed for 26 patients. The time to obtain a result from POC analysis took a median of 9 [7, 10] minutes. Laboratory analysis took a median of 328 [257, 522] minutes from the time of test request to the time of report. Lin's correlation coefficient showed almost perfect agreement (0.99%) for arterial- vs capillary-POC and both POC methods vs arterial laboratory analysis. Bland–Altman plots showed a mean difference of 2.0 (3.7) with 95% limits of agreement of −5.4 to 9.3 for capillary vs laboratory, 1.6 (3.4) and −5.1 to 8.4 for arterial vs laboratory, and −0.137 (2.6) and −5.2 to 4.9 for capillary vs arterial. Patient classification as having inadequately controlled diabetes (>53 mmol/mol) showed 100% agreement across all tests. ConclusionsHbA1c values can be accurately and rapidly obtained using POC testing in the critically ill.