Department of Clinical Chemistry, University of Liège, CHU Sart-Tilman, 4000 Liège, Belgium.
Nephrology, Dialysis, Transplantation, University of Liège, CHU Sart-Tilman, 4000 Liège, Belgium.
Clin Chim Acta. 2019 Dec;499:123-127. doi: 10.1016/j.cca.2019.08.032. Epub 2019 Aug 31.
Determination of creatinine and estimation of Glomerular Filtration Rate (GFR) rapidly before injection of contrast media provides early detection of high-risk patients for acute kidney failure. Hence, a rapid point-of-care (POC) device (result in 30 s) allowing creatinine measurement and eGFR could be of interest. To validate this method, we considered a population referred for measuring GFR.
Iohexol plasma clearance was used to measure GFR. For each subject, enzymatic creatinine was quantified with two different devices: in plasma with the Roche Cobas analyzer and in capillary blood with the Nova Biomedical POC device. Both values of creatinine were used in the CKD-EPI equation for estimated glomerular filtration rate (eGFR). eGFR using POC was compared to eGFR using Cobas and to mGFR by Passing Bablok regression, calculation of bias, precision and accuracy (or concordance) within 30%. Also, we calculated the rate of discrepant staging (eGFR >60 or ≤ 60 when mGFR is actually ≤60 and > 60) with both creatinine methods.
120 subjects (52 ± 13 years, 49% of women) were included. Mean mGFR was 77 ± 27 mL/min/1.73m with 29 patients presenting mGFR <60 mL/min/1.73m. Passing- Bablok regression comparing eGFR obtained with the POC and the Cobas was: eGFR = -0.1 (95% CI: -7.4; 3.0) + 1.06 (95% CI: 1.00; 1.15) x eGFR. Mean bias was 3.7 ± 14.1 mL/min/1.73m. Concordance within 30% was 82%. Compared to mGFR, Passing-Bablok with POC was: eGFR = -11.5 (95% CI: -22.9; -0.7) + 1.15 (95% CI: 1.02; 1.29) x mGFR. Mean bias was 0.1 ± 17.6 mL/min/1.73m. Accuracy within 30% was 81%. Between eGFR and mGFR, mean bias was -3.7 ± 12.5 mL/min/1.73m. Accuracy within 30% was 95%. With POC (and Cobas), 3.3% (0.8%) of patients would have been considered with GFR > 60 mL/min/1.73m whereas mGFR it was ≤60 and 10% (9.2%) of them would have been considered with GFR ≤60 mL/min/1.73m when mGFR was >60.
Creatinine measured with the POC has an acceptable performance when used with the CKD-EPI equation to estimate GFR. Its ability to detect GFR <60 mL/min/1.73m is not significantly different from the classical Roche assay. StatSensor Creatinine (Nova Biomedical) can be used for GFR screening before contrast media injection.
在注射造影剂前快速测定肌酐并估算肾小球滤过率（GFR）可早期发现急性肾衰高危患者。因此，一种可快速（30s 内）检测肌酐并估算 eGFR 的即时检验（POC）设备可能具有应用价值。为验证该方法，我们对一个接受 GFR 检测的患者群体进行了研究。
采用碘海醇清除率来检测 GFR。对于每个患者，我们分别用两种不同的设备检测血浆中的酶法肌酐：罗氏 Cobas 分析仪检测血浆肌酐，Nova Biomedical POC 设备检测毛细血管血肌酐。两种肌酐值均用于 CKD-EPI 方程估算肾小球滤过率（eGFR）。使用 POC 检测的 eGFR 与使用 Cobas 检测的 eGFR 及 mGFR 进行 Passing-Bablok 回归分析，计算偏差、30%内的精度和准确度（或一致性）。此外，我们还计算了两种肌酐检测方法估算的 GFR 分期差异（eGFR >60 或 mGFR ≤60 但实际上 mGFR >60）发生率。
共纳入 120 例患者（52±13 岁，49%为女性）。平均 mGFR 为 77±27mL/min/1.73m，29 例患者的 mGFR <60mL/min/1.73m。POC 和 Cobas 检测的 eGFR 间 Passing-Bablok 回归为：eGFR=-0.1（95%CI：-7.4；3.0）+1.06（95%CI：1.00；1.15）×eGFR。平均偏差为 3.7±14.1mL/min/1.73m。30%内的一致性为 82%。与 mGFR 相比，POC 与 Passing-Bablok 的回归方程为：eGFR=-11.5（95%CI：-22.9；-0.7）+1.15（95%CI：1.02；1.29）×mGFR。平均偏差为 0.1±17.6mL/min/1.73m。30%内的准确性为 81%。eGFR 和 mGFR 之间的平均偏差为-3.7±12.5mL/min/1.73m。30%内的准确性为 95%。使用 POC（和 Cobas）时，3.3%（0.8%）的患者会被认为 GFR>60mL/min/1.73m，而实际上 mGFR ≤60mL/min/1.73m；10%（9.2%）的患者会被认为 GFR≤60mL/min/1.73m，而实际上 mGFR >60mL/min/1.73m。
使用 POC 检测的肌酐值与 CKD-EPI 方程联合估算 GFR 时具有良好的性能。其检测 GFR <60mL/min/1.73m 的能力与经典的罗氏检测法无显著差异。StatSensor 肌酐检测试剂盒（Nova Biomedical）可用于造影剂注射前的 GFR 筛查。
