eGFR Calculator: The CKD-EPI 2021 Equation Explained

Use an eGFR calculator built on the race-free CKD-EPI 2021 equations, the current US standard. Add cystatin C when creatinine misleads, and check the drug label, because many still ask for Cockcroft-Gault creatinine clearance.

What is the CKD-EPI 2021 equation, and what changed from 2009?

The 2021 equations estimate GFR from serum creatinine, or creatinine plus cystatin C, with age and sex and no race term. The NKF-ASN Task Force recommended immediate implementation of the creatinine refit without race.

In the NEJM 2021 validation, the old race-based equation overestimated measured GFR in Black participants by a median of 3.7 mL/min/1.73 m2. Simply deleting race underestimated it by 7.1. The new creatinine equation lands between: median underestimate of 3.6 in Black participants, overestimate of 3.9 in non-Black participants. Creatinine-cystatin C versions were more accurate, with smaller differences between race groups.

Run it in the NKF calculator, which is valid only with standardized creatinine and cystatin methods, or in CliniWiz's CKD-EPI creatinine-cystatin C 2021 calculator.

When should I add cystatin C to an eGFR?

Add cystatin C when creatinine-based eGFR is likely inaccurate and the result will drive a decision. KDIGO 2024 recommends eGFRcr-cys in that situation (Recommendation 1.2.2.1, grade 1C).

• Start with creatinine-based eGFR. If cystatin C is available, estimate the GFR category from the combination (Recommendation 1.1.2.1, grade 1B).
• Expect distorted creatinine with eating disorders, extreme exercise or bodybuilding, above-knee amputation, and spinal cord injury with paralysis. For class III obesity, KDIGO states eGFRcr-cys has been shown to be most accurate.
• Order creatinine on the same sample as cystatin C, because the combined equation needs both.
• If even eGFRcr-cys seems unreliable and the stakes justify it, KDIGO Practice Point 1.2.2.2 points to measured GFR with an exogenous filtration marker.

How do eGFR values map to KDIGO CKD categories?

KDIGO 2024 splits GFR into G1 through G5, with G3 divided into G3a and G3b. CKD needs abnormalities of kidney structure or function for at least 3 months, so a single value cannot assign a stage.

Source: KDIGO 2024 CKD guideline, Table 2. Without evidence of kidney damage, neither G1 nor G2 meets CKD criteria. Document cause, GFR category, and albuminuria category (A1 to A3) together.

KDIGO says its March 2024 guideline remains the current global standard. Its March 2026 focused update targets Chapter 3 on delaying progression and says nothing about GFR evaluation. No newer GFR-estimation guidance turned up as of 2026-10-10.

Should I dose drugs by eGFR or Cockcroft-Gault creatinine clearance?

Follow the product label for the drug in front of you. FDA's March 2024 guidance recommends eGFR over Cockcroft-Gault estimated creatinine clearance in pharmacokinetic studies, yet labels written around creatinine clearance remain in use.

The rivaroxaban label shows the split. Its atrial fibrillation dosing uses CrCl thresholds (above 50 mL/min versus 50 mL/min or less) and says to calculate CrCl using actual weight. An eGFR on the chemistry panel does not substitute.

• FDA says adult dosage recommendations should use eGFR in mL/min. Convert indexed eGFR by multiplying by BSA and dividing by 1.73.
• FDA says ideal or adjusted body weight in Cockcroft-Gault is likely more accurate than actual weight in overweight or obese individuals. Check which weight the label specifies.
• KDIGO 2024 Practice Point 4.2.2 calls validated creatinine-based eGFR appropriate for most dosing. Practice Point 4.2.4 suggests non-indexed eGFR at extremes of body weight, especially for narrow therapeutic range drugs.
• The NKF Cockcroft-Gault page notes the 1976 equation came from 249 white men, and creatinine standardization lowered average values by about 12%.

Where does this fall short?

eGFR is an estimate, and error varies by patient. KDIGO tells clinicians to weigh the limits of eGFR, account for dietary intake when reading creatinine, and assess error potential when judging change over time. Cystatin C has non-GFR determinants too (KDIGO lists smoking and calls the data minimal), and even the creatinine-only equation leaves group-level over- and underestimation. This article does not cover pediatric equations, pregnancy, dialysis, or transplant, and in acute kidney injury KDIGO Practice Point 4.2.5 says to adapt dosing when GFR is not stable. FDA's guidance is nonbinding, so the specific label, your institution's protocol, and clinical judgment decide.