How this instrument works
Estimated glomerular filtration rate (eGFR) approximates how many millilitres of blood the kidneys clear each minute, standardized to a body surface area of 1.73 square metres. Measuring true filtration directly requires an invasive clearance study, so clinicians instead estimate it from a routine blood test — serum creatinine — combined with age and sex, run through an equation fitted to thousands of paired lab and clearance results. The output, reported in mL/min/1.73m², sits at the center of how chronic kidney disease gets staged and tracked over time.
The 2021 version used here raises the ratio of creatinine to a sex-specific constant (κ) to a power that itself switches depending on whether creatinine sits above or below that constant — a steeper exponent past the breakpoint, a shallower one before it — then multiplies by an age-decay term and, for women, a small upward adjustment. It's a piecewise regression fit to data, not one clean line: the breakpoint sits at 0.9 mg/dL for men and 0.7 for women, chosen because that's roughly where the real relationship bends. Earlier versions of this formula included a separate multiplier applied only to patients identified as Black, built on the premise that they carry systematically higher average muscle mass and therefore produce more creatinine at any given filtration rate. A task force convened by the National Kidney Foundation and American Society of Nephrology reviewed that premise in 2021 and concluded a race-based coefficient wasn't a scientifically sound way to estimate organ function, recommending it be dropped — a change that reclassified the calculated stage of disease for a meaningful share of patients.
Muscle mass, not just filtration, drives creatinine levels, so accuracy narrows at the extremes: bodybuilders, amputees, and anyone with unusual bulk for their age and sex will see the equation drift. A single elevated reading isn't a diagnosis on its own, either — sustained kidney disease is generally confirmed by a result under 60 mL/min/1.73m² on two occasions at least three months apart, since dehydration, a big meal of cooked meat, or certain drugs can nudge creatinine up temporarily.
- Set Sex — it decides which breakpoint (0.9 or 0.7 mg/dL) and which pair of exponents the equation reaches for.
- Enter Age in years.
- Enter Serum creatinine (Scr) in mg/dL from a recent blood draw.
- Read eGFR in mL/min/1.73m²; the working shows exactly which branch fired.
Worked example — two branches side by side
A 50-year-old man with a creatinine of 1.0 mg/dL. Since 1.0 exceeds his 0.9 breakpoint, the steeper exponent applies: 1.0 ÷ 0.9 ≈ 1.111, raised to −1.200 ≈ 0.881. Multiply by 142 for about 125.1. The age-decay term, 0.9938 raised to the 50th power, comes to roughly 0.733. Multiplying those together lands the result at 91.7 mL/min/1.73m², a normal reading.
A 65-year-old woman with a creatinine of 0.6 mg/dL sits below her 0.7 breakpoint, so the shallower exponent kicks in: 0.6 ÷ 0.7 ≈ 0.857, raised to −0.241 ≈ 1.038. Multiply by 142 for about 147.4. The age term for 65 years works out to roughly 0.667, and the female multiplier of 1.012 nudges the total up slightly at the very end. 147.4 × 0.667 × 1.012 lands at 99.5 mL/min/1.73m² — comfortably normal, even 15 years older, because her creatinine sits so far under her breakpoint.
Questions
Why did the 2021 equation remove race from the calculation?
Earlier CKD-EPI versions applied a coefficient that raised the calculated result for patients identified as Black, based on studies suggesting higher average muscle mass and therefore higher average creatinine at any given filtration rate. A task force convened by the National Kidney Foundation and American Society of Nephrology reviewed that evidence in 2021 and concluded that using a social category as a stand-in for biology wasn't scientifically justified — and that it risked delaying diagnosis or nephrology referral for patients whose true kidney function was worse than their inflated number suggested. Dropping the coefficient shifted the assigned disease stage for a meaningful number of people.
Why does the formula switch exponents above and below a breakpoint?
Because the true relationship between creatinine and measured filtration isn't one straight line on a log scale — it bends. Below the breakpoint (0.9 mg/dL for men, 0.7 for women), creatinine climbs only gradually as filtration falls, so a shallow exponent fits well; above it, filtration is dropping faster for the same creatinine rise, so a steeper exponent fits better. Rather than force a single curve across the whole range, the CKD-EPI investigators fit two connected segments to the data — a piecewise regression, not one tidy expression.
Why do men and women use different breakpoints and constants?
Average muscle mass — and with it, average creatinine output at a given filtration rate — differs by sex, so the equation was fit separately for each group: a 0.9 mg/dL breakpoint for men versus 0.7 for women, plus a 1.012 multiplier applied only to women. These are population averages baked into a regression, not corrections tailored to any single patient's actual build.
Is one eGFR result enough to diagnose kidney disease?
No. A single low reading can reflect dehydration, a recent high-protein meal, hard exercise, or certain medications rather than true, lasting kidney disease. Guidelines generally call for a result under 60 mL/min/1.73m² on at least two occasions three months apart, usually alongside a urine albumin test, before chronic kidney disease is diagnosed.
What do the different eGFR ranges mean?
Roughly: 90 and above with no other sign of damage is considered normal; 60-89 can be normal for an older adult or a mild reduction depending on other findings; 30-59 is moderate reduction; 15-29 is severe; below 15 is kidney failure. Staging always weighs trends over months and other markers of damage, not one isolated figure.
How accurate is CKD-EPI against a directly measured filtration rate?
It's a strong population-level estimate but an imperfect individual one. Creatinine-based equations typically land within about 30% of a measured clearance study (iohexol or inulin) for roughly 85-90% of people, per the equations' original validation work. Accuracy narrows at the extremes — very muscular people, amputees, and anyone whose build is unusual for their age and sex.
References
- Inker LA et al. 2021, N Engl J Med — original study (PubMed)
- National Kidney Foundation — Changes to eGFR Calculation
Read this first: This instrument computes a screening figure from population formulas — it is not a diagnosis, and it cannot see the whole picture a clinician can. Use it to inform a conversation, not to replace one.