Endocrine Abstracts

Background: Point-of-care tests (POCT) for glucose promote timely clinical management. We assessed the precision and accuracy of POCT compared with laboratory measurements in children undergoing dynamic function tests.

Methods: Split samples of venous blood were tested on POCT meters (Precision PCx Plus and Precision Xceed Pro) and in the laboratory (ADVIA 2400). Clinical reliability was assessed against the ISO 15197 standard: In at least 95% of cases, discrepancy between reference and point-of-care glucose should be <0.83 mmol/L for glucose <4.2 mmol/L or <20% for glucose ≥4.2 mmol/L. The data were reanalysed after applying a correction factor for glycolysis in sample tubes of +0.39 mmol/L, derived from a large prospective trial.

Results: A total of 1688 pairs of samples were analysed from 206 patients. The data were highly correlated (R=0.958). Linear regression results, expressed in a Passing-Bablok equation as Y=0.856X+1.048 (P<0.001), and Bland-Altman difference plots (intercept +0.941, slope −0.115) both demonstrated higher POCT results at low glucose values and lower results for high glucose values as compared with laboratory results. Only 84.7% (linear regression) and 87.9% (Bland-Altman) of values fell within the tolerated error margins. This increased to 95.5% and 97.8% after adjusting laboratory values for glycolysis.

Conclusion: The Precision meters overestimate plasma glucose for hypoglycaemic states and underestimate them for hyperglycaemic states. When values are adjusted for glycolysis prior to laboratory measurement, this discrepancy remains significant but the meters meet the ISO 15197 requirements.