Evaluation of Point-of-Care Creatinine Analyzers: eGFR Categorization Accuracy for Safe Contrast Administration

초록

Background Point-of-care testing (POCT) devices facilitate rapid creatinine measurement for timely renal function assessment. Accurate creatinine measurement is crucial for estimating glomerular filtration rate (eGFR), which guides clinical decisions including contrast agent administration (eGFR <60 mL/min/1.73m² threshold) and prophylaxis strategies. This study evaluated analytical performance and clinical concordance of three POCT creatinine analyzers. Methods Three POCT devices were evaluated: NOVA pHOx Ultra (Nova Biomedical), Radiometer ABL90 FLEX (Radiometer Medical), and i-STAT Crea cartridge (Abbott). These were compared against Beckman Coulter AU5800 using an enzymatic method with IDMS-traceable calibration. Venous whole blood samples from 355 adult patients (mean age 62 years, 58% male) were collected over 8 weeks. POCT measurements were performed within 15 minutes of venipuncture. eGFR was calculated using the CKD-EPI 2009 equation and categorized into four clinically relevant stages: =60, 45-59, 30-44, and <30 mL/min/1.73 m². Method comparison employed Passing-Bablok regression and Bland-Altman analysis. Agreement was assessed using weighted kappa statistics. Results A total of 355 samples were included after excluding instrument errors or missing data. The correlation coefficients were highest for ABL90 (r = 0.996), followed by i-STAT (r = 0.995), and pHOx Ultra (r = 0.966). Based on allowable total error (±4.2%) at medical decision levels of creatinine (0.6, 1.6, and 6.0 mg/dL), only ABL90 met acceptability across all concentrations. The accuracy of categorical classification in the four stages of eGFR was highest with ABL90 (98.6%), followed by i-STAT (95.2%), and pHOx Ultra (86.5%). Among patients with a reference eGFR = 60 (n = 287), pHOx Ultra misclassified 30 cases (10.5%) as < 60, compared to 8 (2.8%) with i-STAT and 2 (0.7%) with ABL90. In the reference eGFR <30 group (n = 46), pHOx Ultra misclassified four cases (8.7%) as =30, while no misclassification was observed with i-STAT or ABL90. Conclusion Among the evaluated POCT devices, ABL90 demonstrated the best overall concordance with the reference analyzer in terms of both creatinine measurement and eGFR categorization, followed by i-STAT and pHOx Ultra. In misclassified cases, all POCT devices tended to underestimate the eGFR. The superior performance of ABL90 and i-STAT may be attributable to their manufacturer-stated traceability to the NIST reference standards. Further investigations are warranted to evaluate the impact of POCT devices on patient management and clinical outcomes.