The 2021 Race-Free eGFR: Why Kidney Function Testing Changed and What It Means for You

The 2021 race-free eGFR is the CKD-EPI creatinine equation re-fitted without a race term. It uses only serum creatinine, age and sex. The older 2009 CKD-EPI equation multiplied the result by 1.159 when a patient was recorded as Black, and the MDRD equation used a multiplier of 1.210. Removing that step means the number printed on a lab report can change at the switch even though the kidneys did not. In our calculations, a Black patient typically sees a result about 5 to 14 percent lower under the 2021 equation, while a non-Black patient sees a result between about 0 and 11 percent higher.

What actually changed in the equation

The coefficients below come from the NIDDK and NKF pages for the 2021 equation and from the published 2009 equation. Creatinine is in mg/dL, and results are in mL/min/1.73 m2.

MDRD, the older equation, carried a 1.210 multiplier for African American patients. Because the 2021 equation was re-fitted rather than simply having a multiplier deleted, the effect of the change is not a flat percentage.

How big is the change? Computed examples

We computed both equations for the same lab values. The 2009 columns are shown with and without the race multiplier so you can see both sides of the change. The NKF categories are G1 90 or above, G2 60 to 89, G3a 45 to 59, G3b 30 to 44.

One of these rows crosses a category line. The 55-year-old woman moves from 65.3 (G2) to 59.3 (G3a) if she was recorded as Black, because the old multiplier no longer applies. A 65-year-old non-Black woman with a creatinine of 1.0 moves in the other direction, from 58.9 (G3a) to 62.5 (G2). Across ages 20 to 90 and creatinine values giving an eGFR between 15 and 120, the 2021 result was 4.5 to 13.9 percent lower than the multiplied 2009 result and 0.3 percent lower to 10.7 percent higher than the unmultiplied one. The effect is largest at higher eGFR and smaller at lower eGFR, which matches the NKF laboratory guidance that values may differ by more than 10 percent particularly at higher eGFR and younger ages.

Why the race term was removed

The 2021 development paper in the New England Journal of Medicine (Inker and colleagues) opens with the reason: race in eGFR equations is a social and not a biologic construct. The NKF and the American Society of Nephrology convened a joint task force that recommended the 2021 CKD-EPI creatinine equation, without a race variable, for use in all laboratories, together with wider use of cystatin C.

The paper also quantified the trade-offs. Using the 2009 creatinine equation with its race term, measured GFR was overestimated in Black participants by a median of 3.7 mL/min/1.73 m2. Simply deleting the race term from that equation would have underestimated measured GFR in Black participants by a median of 7.1. The refitted 2021 equation underestimated it in Black participants by 3.6 and overestimated it in non-Black participants by 3.9, and 85 percent or more of estimates for both groups fell within 30 percent of measured GFR for all equations. Tufts, which hosts the equations, adds that the 2021 version is somewhat less accurate than 2009 in both groups against measured GFR, though acceptable for clinical use. In other words, the change was chosen to avoid using race as a biological proxy, and it comes with a small accuracy cost that cystatin C is intended to reduce.

A published review of the change reported a projected national effect of reclassifying roughly 584,000 Black adults to more advanced CKD stages and expanding nephrology referral for about 41,800. That is a modelled estimate, not a measurement, and it should be read as such.

What the change means for your own result

• Your kidneys did not change on the day your lab switched. Only the arithmetic did.
• If your old and new results sit on different sides of 60, compare them using the same equation before concluding anything. You can recompute an older value from the original creatinine if your report lists it.
• Staging requires more than one number. Under KDIGO 2024, CKD is abnormalities of kidney structure or function present for more than 3 months, shown by GFR below 60 or by markers of kidney damage such as albuminuria. NKF classification also says G1 and G2 are not CKD without evidence of kidney damage.
• The NKF laboratory guidance recommends cystatin C confirmation when eGFR from creatinine is 45 to 59 and the urine albumin-to-creatinine ratio is under 30 mg/g, because creatinine alone can be misleading in that zone.
• Creatinine-based equations are affected by muscle mass, diet, pregnancy and some medicines, regardless of race.

Common mistakes

• Subtracting an old eGFR from a new one to measure change over time when they came from different equations.
• Applying a race multiplier by hand. The 2021 calculation has none; entering race adds nothing.
• Assuming a lower number means disease. The same patient simply has a different estimate; disease is decided from the trend, urine tests and clinical context.
• Treating the 2021 equation as exact. It is still an estimate with error.

Cystatin C: the other half of the 2021 update

The 2021 CKD-EPI creatinine-cystatin C equation uses both markers and is stated by NIDDK to be preferred and more accurate than creatinine alone. In the Inker paper, equations that combined creatinine and cystatin C without race were more accurate and gave smaller differences between Black and non-Black participants than creatinine-only or cystatin-only versions. Access to cystatin C testing varies by lab and region, so ask your clinician whether it is appropriate for you.

Frequently asked questions

Did my kidney function drop when my lab updated its equation?

No. The blood test was the same. The equation changed. A higher or lower printed number after the update is a calculation effect, and your clinician will interpret it with your history.

Does our calculator ask for race?

No. It uses the 2021 CKD-EPI creatinine equation, which has no race input.

Is the 2021 equation more accurate than the 2009 one?

Against measured GFR it is slightly less accurate on average according to Tufts, but it avoids using race and is recommended by the NKF and ASN task force. The creatinine-cystatin C version is the most accurate of those tested.