Interpreting eGFR: CKD-EPI 2021 vs Cockcroft-Gault vs EKFC

eGFR appears on almost every chemistry panel, yet it is one of the most misread numbers in medicine. Why is the Cockcroft-Gault creatinine clearance so different from the lab's eGFR? Which one should you use to dose a DOAC? And where does the European EKFC equation fit in?

This guide sorts out the common equations, using the KDIGO 2024 chronic kidney disease (CKD) guideline as the reference point [1].

Key points

• eGFR estimates glomerular filtration rate from serum markers (creatinine, cystatin C) and is indexed to 1.73 m² of body surface area [1].
• CKD-EPI 2021 removed the race coefficient. Combined creatinine–cystatin C equations are more accurate than creatinine alone [2].
• EKFC is a European equation covering ages 2 to 90 [3].
• Cockcroft-Gault estimates creatinine clearance in mL/min, not indexed. Many drug labels, including those of the DOACs, still refer to it [4, 5].
• KDIGO recommends using one validated equation consistently within a region and not using race in eGFR [1].

Calculate it: eGFR (CKD-EPI) calculator and Cockcroft-Gault creatinine clearance.

What eGFR does and does not tell you

True GFR can only be measured with exogenous filtration markers such as iohexol (mGFR). In routine care it is estimated from serum creatinine. Creatinine levels depend on filtration, but also on muscle mass, diet and tubular secretion. Equations correct for this using age and sex, which works for the average patient and less well for anyone far from average.

KDIGO lists situations where creatinine-based eGFR is less accurate [1]:

• Unusual muscle mass: eating disorders, bodybuilding, amputation, spinal cord injury, muscle-wasting disease
• Diet: high-protein diets or creatine supplements, and also vegetarian, ketogenic or low-protein diets
• Serious illness: malnutrition, cancer, heart failure, cirrhosis
• Class III obesity

Every eGFR also assumes a steady state. In acute kidney injury, creatinine lags behind the fall in filtration, so eGFR overestimates kidney function while creatinine is still rising. KDIGO advises adapting drug dosing when GFR or its markers are not stable [1].

CKD-EPI 2021: the international default

CKD-EPI equations have long been the most widely used. In 2021 the collaboration published new versions without a race coefficient, on the basis that race is a social rather than a biological construct [2].

Validation results worth knowing [2]:

• The new race-free creatinine equation overestimated measured GFR in non-Black participants by a median 3.9 mL/min/1.73 m² and underestimated it in Black participants by 3.6.
• Creatinine–cystatin C equations were more accurate, with smaller differences between groups.
• For all equations, 85% or more of estimates fell within 30% of measured GFR.

That last point deserves attention: even a good equation is off by more than 30% in a meaningful share of patients.

EKFC: the European option

The European Kidney Function Consortium (EKFC) equation was published by Pottel and colleagues in 2021 [3]. What sets it apart:

• It is a full age spectrum equation, valid from 2 to 90 years, which avoids the jump in eGFR when patients move from paediatric to adult formulas.
• It rescales creatinine using a Q value, the typical creatinine of a healthy reference population.
• In external validation it showed low bias and fewer estimates off by more than 30% than CKiD and CKD-EPI [3].

The caveat: the original study included no Black participants [3]. KDIGO notes that the creatinine-based EKFC equation is recommended for White Europeans, that Q values have since been published for further populations, and that an EKFC cystatin C equation uses only age and cystatin C [1].

KDIGO 2024 does not mandate a single equation. It recommends a validated equation, applied consistently within a region [1]. Your lab report should state which one it uses.

Cockcroft-Gault: old, but still around

The Cockcroft-Gault formula dates from 1976 [4]:

Creatinine clearance (mL/min) = (140 − age) × weight (kg) / (72 × serum creatinine in mg/dL), × 0.85 for women

It is a different kind of estimate:

KDIGO summarises the concerns: the formula predates creatinine standardisation, women were not included, and body weight is distorted by oedema and obesity [1].

It survives because many pharmacokinetic studies behind drug labels used it. The best-known example is the DOACs. The European Heart Rhythm Association (EHRA) practical guide recommends Cockcroft-Gault for assessing kidney function when dosing DOACs and warns that MDRD and CKD-EPI may overestimate kidney function, especially in older and low-weight patients [5].

A worked example

An 80-year-old woman, 50 kg, 160 cm, serum creatinine 1.0 mg/dL (illustrative only):

Depending on the formula she sits in G3a or well below it, which matters for any drug with a dose threshold at 50 or 30 mL/min. In young, muscular men the relationship often reverses: a 40-year-old man weighing 80 kg with creatinine 1.0 mg/dL has a CKD-EPI 2021 eGFR of about 98 mL/min/1.73 m² and a Cockcroft-Gault estimate of about 111 mL/min.

eGFR and drug dosing

KDIGO 2024 offers several practice points [1]:

• Consider GFR when dosing drugs cleared by the kidneys (Practice Point 4.2.1).
• For most patients, validated creatinine-based eGFR equations are appropriate for dosing (4.2.2).
• Where accuracy matters, for example with a narrow therapeutic index, use creatinine plus cystatin C or measured GFR (4.2.3).
• At extremes of body weight, use non-indexed eGFR (4.2.4): indexed eGFR × body surface area / 1.73.

KDIGO acknowledges that this guidance does not always match drug labels [1]. In practice, the summary of product characteristics for a given drug defines how it should be dosed. For DOACs that is usually Cockcroft-Gault creatinine clearance [5].

When to add cystatin C

KDIGO recommends combined creatinine–cystatin C eGFR when creatinine-based eGFR is likely to be inaccurate and GFR will affect clinical decisions (Recommendation 1.2.2.1, grade 1C) [1]. The muscle-mass scenarios above are the classic cases. When even more accuracy is needed, for example when assessing a potential living kidney donor, measure GFR directly [1].

Cystatin C has its own confounders; KDIGO lists smoking among them [1]. A large gap between eGFRcr and eGFRcys is itself informative and is worth explaining before acting on either number.

KDIGO staging in one place

CKD is defined by abnormalities of kidney structure or function present for more than 3 months, such as GFR <60 mL/min/1.73 m² or albuminuria [1]. It is classified by Cause, GFR category and Albuminuria category (CGA):

Without markers of kidney damage, neither G1 nor G2 meets the definition of CKD [1]. A single eGFR of 55 in an 85-year-old is not a diagnosis on its own; trend and albuminuria decide.

A quick checklist

1. Which equation does the report use? CKD-EPI, EKFC or something else?
2. Is the patient in a steady state? If not, eGFR is unreliable.
3. Does the body fit the equation? Think of cystatin C with extreme muscle mass, cachexia or amputation.
4. Is this about dosing? Check the drug label. For DOACs use Cockcroft-Gault; at extreme weights use non-indexed eGFR.
5. Is this about diagnosing CKD? Look at the 3-month trend and the ACR.

Bottom line

No single kidney equation is right for every question. CKD-EPI 2021 and EKFC estimate GFR for diagnosis and staging. Cockcroft-Gault estimates creatinine clearance and remains relevant wherever drug labels rely on it. Knowing what each formula measures and when it fails lets you interpret disagreements instead of being surprised by them.

For drug- or guideline-specific follow-up questions, Medelix gives answers with citations you can open and check.

Sources

1. Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease. Kidney Int. 2024;105(4S):S117–S314. PDF at kdigo.org
2. Inker LA, Eneanya ND, Coresh J, et al. New Creatinine- and Cystatin C–Based Equations to Estimate GFR without Race. N Engl J Med. 2021;385(19):1737–1749. doi:10.1056/NEJMoa2102953 · PMC8822996
3. Pottel H, Björk J, Courbebaisse M, et al. Development and Validation of a Modified Full Age Spectrum Creatinine-Based Equation to Estimate Glomerular Filtration Rate. Ann Intern Med. 2021;174(2):183–191. doi:10.7326/M20-4366
4. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron. 1976;16(1):31–41. doi:10.1159/000180580
5. Steffel J, Collins R, Antz M, et al. 2021 European Heart Rhythm Association Practical Guide on the Use of Non-Vitamin K Antagonist Oral Anticoagulants in Patients with Atrial Fibrillation. Europace. 2021;23(10):1612–1676. PMC11636576