A Comparison of the 2009 CKD-EPI and EKFC Equations for Estimating GFR in a Large Italian Outpatient Cohort, and the Potential Consequences for Patients’ Management

Abstract

Introduction
An accurate estimation of the glomerular filtration rate (eGFR) is essential for chronic kidney disease (CKD)
management and clinical decision-making. The CKD-EPI 2009 has long been the standard but it is currently being
replaced by new equations designed to correct some of its shortcomings. Specifically, in 2021 the European Kidney
Function Consortium (EKFC) developed a full age spectrum equation aimed at improving accuracy by introducing agespecific
parameters. Unlike the CKD-EPI equation, which applies a single fixed age coefficient, the EKFC equation
theory supports a non-linear decline of kidney function, where GFR remains quite stable until 40 years of age before
progressively decreasing with ageing.
Objectives
To understand how eGFR estimates obtained by CKD-EPI and EKFC differ according to age and sex, and to assess the
impact of CKD categories reclassification and subsequent clinical decision making induced by shifting from the CKD-EPI
2009 to the EKFC equation.
Methods
Data were retrospectively collected from the Emilia-Romagna regional PIRP registry of individuals with chronic kidney
disease, which was established in 2004 and includes data from patients attending the 12 hospital nephrology units of the
region. For this analysis, we selected patients who entered PIRP from 2009 onwards and medical records with creatinine
range 0.45-5.54 mg/dl to ensure compatibility with creatinine calibration methods used by EKFC. After excluding records
with known data quality issues, we obtained a cohort of 127,876 medical records from 30,549 patients. Calculated eGFR
was grouped into categories (G1 to G5) according to the KDIGO CKD guidelines. Statistical analyses were all
descriptive and included gender-specific scatterplots to compare the delta between formulae (CKD-EPI - EKFC) and
patients’ age.
Results
The EKFC equation provided lower eGFR estimates than CKD-EPI in 80.7% of records. This was particularly evident in
younger and older age groups, while higher estimates from EKFC concentrated around 40 years of age. Grouping
according to the baseline eGFR placed 87.9% of patients in the same category. EKFC estimation using baseline data
classified 254 patients (0.8%) in a less advanced CKD group and 3418 patients (11.2%) in a more advanced CKD group.
Dialysis initiation was disproportionately lower (2.5% vs. 8.9% overall) and higher (36.2%) in patients assigned to a more
or less advanced CKD group by EKFC, respectively. Deaths were disproportionately lower when EKFC assigned
patients to a less advanced CKD group (10.2% vs. 30.6% overall).
Conclusions
Consistent with previous literature, our results showed that EKFC equation tend to attribute lower eGFR values than
CKD-EPI equation. However, this was not uniformly distributed across all age brackets, due to the age dependency of
GFR hypothesized in the EKFC equation. For 10% of patients, switching to the EKFC estimate would have resulted in a
change of CKD grouping, which could have impacted disease management. Assuming that the hypothesis of age-GFR
relationship made in EKFC is more accurate than that underlying CKD-EPI, this figure indicates that in absolute terms a
great number of patients would benefit from the switch. Nevertheless, the observed discrepancies in clinical outcomes
warrant careful further investigation.