Quantifying Individual-Level Inaccuracy in Glomerular Filtration Rate Estimation A Cross-Sectional Study

Background: Although the population-level differences between estimated glomerular filtration rate (eGFR) and measured glomerular filtration rate (mGFR) are well recognized, the magnitude and potential clinical implications of individual-level differences are unknown. Objective: To quantify the magnitude and consequences of the individual-level differences between mGFRs and eGFRs. Design: Cross-sectional study. Setting: Four U.S. community-based epidemiologic cohort studies with mGFR. Patients: 3223 participants in 4 studies. Measurements: The GFRs were measured using urinary iothalamate and plasma iohexol clearance; the eGFR was calculated from serum creatinine concentration alone (eGFRCR) and with cystatin C. All GFR results are presented as mL/min/1.73 m2. Results: The participants' mean age was 59 years; 32% were Black, 55% were women, and the mean mGFR was 68. The population-level differences between mGFR and eGFRCR were small; the median difference (mGFR_eGFR) was -0.6 (95% CI, -1.2 to -0.2); however, the individual-level differences were large. At an eGFRCR of 60, 50% of mGFRs ranged from 52 to 67, 80% from 45 to 76, and 95% from 36 to 87. At an eGFRCR of 30, 50% of mGFRs ranged from 27 to 38, 80% from 23 to 44, and 95% from 17 to 54. Substantial disagreement in chronic kidney disease staging by mGFR and eGFRCR was present. Among those with eGFRCR of 45 to 59, 36% had mGFR greater than 60 whereas 20% had mGFR less than 45; among those with eGFRCR of 15 to 29, 30% had mGFR greater than 30 and 5% had mGFR less than 15. The eGFR based on cystatin C did not provide substantial improvement. Limitation: Single measurement of mGFR and serum markers without short-term replicates Conclusion: A substantial individual-level discrepancy exists between the mGFR and the eGFR. Laboratories reporting eGFR should consider including the extent of this uncertainty to avoid misinterpretation of eGFR as an mGFR replacement.