GFR Worked Examples: Six eGFR Calculations Shown Step by Step
Six fully worked CKD-EPI 2021 examples — real ages, creatinines, and every factor shown — so you can see exactly how an eGFR calculator turns three inputs into a stage.
Equations become trustworthy the first time you watch one work. This post runs six realistic cases through the CKD-EPI 2021 race-free equation — the same arithmetic a GFR calculator performs in milliseconds — with every intermediate factor written out so nothing is hidden. A healthy 55-year-old man, a 40-year-old woman, an older man drifting toward stage 3a, a woman sitting on the G3a border, a severely reduced case in G4, and a young woman whose age does surprising work. Each example ends with the stage band and a one-line interpretation. All values are educational illustrations: estimates, not diagnoses, and no set of numbers here replaces a lab, a trend, or a clinician.
CHAPTER 01The Equation in One Box
Everything below uses one formula: eGFR equals 142, times the larger of creatinine divided by kappa or 1 raised to the power negative 1.200, times 0.9938 to the age, times 1.012 for women. Kappa is 0.7 for women and 0.9 for men. When creatinine divided by kappa is below 1, that first factor is simply 1, which is why mild creatinine values barely register in younger people. Creatinine is in milligrams per deciliter, age in years, and the result is milliliters per minute per 1.73 square meters.
Two reading rules keep the examples honest. First, the numbers are rounded at each step for display, so your own recomputation may land within a point of the printed answer — that is rounding, not error. Second, the stage bands are wide: G1 is 90 and above, G2 is 60 to 89, G3a is 45 to 59, G3b is 30 to 44, G4 is 15 to 29, and G5 is below 15. A value of 88 and a value of 61 are the same stage, and the stage is the unit of interpretation. The gfr calculator at /gfr-calculator.html runs this same arithmetic; the examples below exist so you can check it, not just trust it.
CHAPTER 02Example 1: A 55-Year-Old Man, Creatinine 1.0
Inputs: male, age 55, creatinine 1.0 mg/dL. Kappa is 0.9, so creatinine divided by kappa is 1.0 divided by 0.9, which is 1.111. The larger of that and 1 is 1.111, raised to the power negative 1.200: that factor works out to about 0.881. The age term is 0.9938 to the 55th power, about 0.710.
Now multiply: 142 times 0.881 is 125.1; 125.1 times 0.710 is 88.8, which rounds to about 89 mL/min/1.73m². No female multiplier applies.
Interpretation: 89 sits at the very top of G2 (60–89), one point below G1. For a 55-year-old this is a completely ordinary reading — the kind that gets filed without comment — and it illustrates how a one-point difference has no clinical meaning at all. The stage band, not the digit, is the message.
CHAPTER 03Example 2: A 40-Year-Old Woman, Creatinine 0.8
Inputs: female, age 40, creatinine 0.8 mg/dL. Kappa is 0.7, so creatinine divided by kappa is 0.8 divided by 0.7, about 1.143. Raised to the power negative 1.200, that factor is about 0.852. The age term is 0.9938 to the 40th, about 0.780.
Multiply: 142 times 0.852 is 121.0; 121.0 times 0.780 is 94.4; times the female factor 1.012 gives 95.5, so roughly 95 mL/min/1.73m².
Interpretation: 95 is G1 territory. A woman in her forties commonly carries less muscle mass than a man the same age, so her creatinine is lower and the equation rewards it. The example shows the female multiplier and the lower kappa working in the same direction — the reason identical creatinine values produce different estimates by sex.
CHAPTER 04Example 3: A 70-Year-Old Man, Creatinine 1.6
Inputs: male, age 70, creatinine 1.6 mg/dL. Creatinine divided by kappa is 1.6 divided by 0.9, about 1.778; raised to the power negative 1.200 that factor is about 0.501. The age term is 0.9938 to the 70th, about 0.647.
Multiply: 142 times 0.501 is 71.1; 71.1 times 0.647 is 46.0, so approximately 46 mL/min/1.73m².
Interpretation: 46 lands just inside G3a (45–59). Notice how much work the age term does — the same 1.6 creatinine in a 30-year-old would compute far lower, because the equation expects creatinine to rise as filtration falls with age. The borderline placement is also the lesson: 46 versus 44 versus 59 changes the band, so borderline values are exactly the ones clinicians retest.
CHAPTER 05Example 4: A 65-Year-Old Woman, Creatinine 1.2
Inputs: female, age 65, creatinine 1.2 mg/dL. Creatinine divided by kappa is 1.2 divided by 0.7, about 1.714; to the power negative 1.200 gives about 0.524. The age term is 0.9938 to the 65th, about 0.667.
Multiply: 142 times 0.524 is 74.4; 74.4 times 0.667 is 49.6; times 1.012 gives 50.2, so about 50 mL/min/1.73m².
Interpretation: 50 is mid-G3a. Compare this with Example 3: a man of 70 at creatinine 1.6 scored 46, while this woman at 65 and 1.2 scores 50 — different inputs, similar destination. That is the equation doing its job, mapping very different people into a shared scale. It is also why you cannot eyeball creatinine alone and guess the stage.
CHAPTER 06Example 5: A 45-Year-Old Man, Creatinine 2.8
Inputs: male, age 45, creatinine 2.8 mg/dL. Creatinine divided by kappa is 2.8 divided by 0.9, about 3.111; to the power negative 1.200 gives about 0.256. The age term is 0.9938 to the 45th, about 0.756.
Multiply: 142 times 0.256 is 36.4; 36.4 times 0.756 is 27.5, so roughly 27 mL/min/1.73m².
Interpretation: 27 sits inside G4 (15–29). Values in this band are where the estimate is most trusted — closer to the data it was fitted on — and where clinical involvement is typically already established rather than discovered by calculator. For an educational example, the takeaway is how steeply the negative exponent penalizes rising creatinine: from 1.6 to 2.8, the same man's age profile drops from 46 to 27.
CHAPTER 07Example 6: A 30-Year-Old Woman, Creatinine 0.9
Inputs: female, age 30, creatinine 0.9 mg/dL. Creatinine divided by kappa is 0.9 divided by 0.7, about 1.286; to the power negative 1.200 gives about 0.740. The age term is 0.9938 to the 30th, about 0.830.
Multiply: 142 times 0.740 is 105.1; 105.1 times 0.830 is 87.2; times 1.012 gives 88.3, so about 88 mL/min/1.73m².
Interpretation: 88 is the top of G2, and it surprises people — a young, healthy creatinine can still compute below 90 because the age term at 30 is doing less lifting than intuition expects. The lesson is that 90 is a border, not a wall, and single digits around it are noise. Run these six cases yourself at /gfr-calculator.html and confirm each result; the arithmetic agreement is the point.
CHAPTER 08Cross-Checks and Cautions
Sanity-check any calculation against three anchors. Direction: higher creatinine must lower the estimate, all else equal. Age: older must lower it, all else equal. Sex: at the same creatinine, women score a touch higher than men because of the 1.012 multiplier and lower kappa. If your hand calculation or a web tool violates any of those, something is wrong with the inputs — units above all, since micromoles per liter must be divided by 88.4 before entering.
And the standing caution: every number in this post is an educational illustration of an equation, not a clinical result. Real interpretation adds albuminuria, trend, medications, and the specific human — work that belongs to a clinician. Use worked examples to understand the machine, not to diagnose the patient; that distinction is the difference between numeracy and self-diagnosis. And if the arithmetic sold you, one hand-checked example against /gfr-calculator.html will close the sale honestly.
🔑 Key takeaways
- One formula drives all six examples: 142 times a creatinine-power factor, times 0.9938 to the age, times 1.012 for women, with kappa 0.7 for women and 0.9 for men.
- Example 1 (man, 55, creatinine 1.0) computes to about 89 — one point below G1, and clinically identical to 91.
- The age term does heavy lifting: the same creatinine scores very differently at 30 versus 70, as Examples 3 and 6 show.
- Example 5 shows the steep penalty of rising creatinine: from 1.6 to 2.8, an estimate falls from 46 to 27.
- Round-off at display precision means hand recomputation can differ by a point; that is rounding, not error.
- Check direction, age, and sex anchors on any tool, convert micromole units by dividing by 88.4, and treat every output as an estimate — not a diagnosis.
❓ Frequently asked questions
Why do my hand calculations differ slightly from your numbers?
The examples round each factor to two or three decimals for readability, so a full-precision computation can land a point away. That is expected rounding drift, and it is far smaller than the biological noise between blood draws.
Which stage is an eGFR of 88?
G2, which runs 60 to 89. Because G2 borders G1, values in the high 80s are usually treated as reassuring in the absence of other markers, and no single digit at a border carries meaning by itself.
Do the examples apply to children?
No. The CKD-EPI 2021 adult equation was fitted on adult data. Children use different pediatric equations, which is why a calculator built on the 2021 formula should ask for adult inputs only.
Can I use a creatinine in micromoles per liter?
Divide it by 88.4 first to convert to milligrams per deciliter — 80 micromoles per liter is about 0.9 mg/dL. Entering international units directly is the single most common way to get a nonsense result.
What if creatinine divided by kappa is below 1?
Then that power factor is simply 1, and the estimate comes almost entirely from the age term and sex. This is why mild creatinine differences between healthy young people barely move the output.
Is an estimate of 27, like Example 5, something to self-manage?
No. G4-range values are firmly clinician territory — real patients there typically have established care plans. The example exists to show the arithmetic, not to suggest that arithmetic is care.
The free Gfr Calculator on Toolfyra runs everything in your browser — no signup, nothing uploaded.
Open the Gfr Calculator →📚 More in the Toolfyra blog · or browse all free online tools.