Kidney Week

Abstract: SA-PO403

GFR Measurement Method: A Critical Determinant in Estimation Equation Assessment

Session Information

• CKD: Estimating Equations, Incidence, Prevalence, Special Populations
  November 04, 2017 | Location: Hall H, Morial Convention Center
  Abstract Time: 10:00 AM - 10:00 AM

Category: Chronic Kidney Disease (Non-Dialysis)

• 302 CKD: Estimating Equations, Incidence, Prevalence, Special Populations

Authors

• Allen, Céline M, Queen's University, Kingston, Ontario, Canada

Céline M Allen,

• Akbari, Ayub, University of Ottawa, Ottawa, Ontario, Canada

Ayub Akbari,

• Knoll, Greg A., University of Ottawa, Ottawa, Ontario, Canada

Greg A. Knoll,

• Collier, Christine P., Queen's University, Kingston, Ontario, Canada

Christine P. Collier,

• White, Christine A., Queen's University, Kingston, Ontario, Canada

Christine A. White,

Background

Multiple equations exist to translate various serum analyte concentrations into estimates of the glomerular filtration rate (eGFR). Validation studies of these estimates often yield differing results. This is likely the result of differing patient populations and characteristics, analyte assay manufacturer biases, and GFR measurement protocols. This study was designed to examine the impact of GFR measurement methodology on the performance of the creatinine CKD-EPI equation (eGFRCr-EPI).

Methods

Cr was measured and eGFRCr-EPI calculated in 85 research subjects. GFR was measured (mGFR) simultaneously using different methodologies: renal inulin clearance, plasma ^99mTc-DTPA clearance (2-4 hour sampling), plasma iohexol clearance (2-4 hour sampling) and plasma iohexol clearance (2-10 hour sampling). For each method, bias (eGFR - mGFR), precision (standard deviation of mean bias) and accuracy (P30-the percent of all eGFRCr-EPI within 30% of the mGFR) were determined. Bias and accuracies were compared using paired t-tests and McNemar’s test respectively.

Results

Mean age 60 ± 14 yrs, mean BSA 1.98 ± 0.30 m², 95% non-black and 40% female sex. Mean inulin GFR 39.3 ± 28.8 ml/min/1.73m² while mean eGFRCr-EPI was 38.7 ± 28.5 ml/min/1.73m². Performance results are shown in Table 1.

Conclusion

eGFRCr-EPI performance results differ significantly depending on how GFR is measured (tracer, clearance strategy and sample timing). Short plasma-based strategies yield the highest biases and worse accuracies for the eGFRcr-EPI equation. These discrepancies need to be considered when interpreting validation data and designing validation studies.

Table 1: eGFRCr-EPI performance by GFR methodology

* p = 0.0001 compared to inulin, ** p = 0.08 (NS) compared to inulin, ^β p = 0.0012 compared to inulin, ^δ p = 0.006 compared to inulin

Funding

• Government Support - Non-U.S.