Creatinine Clearance & eGFR Comparison
Cockcroft-Gault is linear in body weight and CKD-EPI ignores weight entirely — so which reads higher is decided by the patient size.
A study aid, not a dosing tool. Drug labels specify which estimate they were validated against — many older drugs are labelled against Cockcroft-Gault specifically, so substituting a reported eGFR is not a like-for-like swap. Neither number is a measurement, both fail when creatinine is not at steady state, and nothing here knows anything about a patient. Do not let this be the last thing you check before prescribing.
Which weight?
The same patient at different body weights
And at different ages
They are not two roads to one answer
Cockcroft-Gault is linear in actual body weight. CKD-EPI does not use weight at all. On identical age, sex and creatinine a 140 kg patient scores 2.80 times a 50 kg one on Cockcroft-Gault and exactly the same on CKD-EPI — which means the two equations are not approximations of each other, they answer different questions. Which one reads higher is decided by body weight, and monotonically: across a sweep of ages, creatinines and sexes, Cockcroft-Gault is the higher figure in 3 per cent of cases at 50 kg, 64 per cent at 70, 95 at 90, and 100 per cent at 110 and above. The crossover sits near average body weight, which is exactly why they agree well enough in textbook patients for the problem to stay hidden until it matters.
The indexing trap is the one people miss. CKD-EPI reports per 1.73 square metres of body surface area; Cockcroft-Gault does not index at all, so comparing the raw numbers compares different units. De-indexing CKD-EPI for a 190 cm, 140 kg patient moves it from 86 to 131 mL/min — a difference larger than most of the thresholds anybody cares about. Sweeping plausible patients, the two land on opposite sides of the 30 mL/min cut 15.9 per cent of the time and of 60 in 17.2, and where they disagree at 30, Cockcroft-Gault is the higher number 82 per cent of the time. The disagreement is biased toward the permissive direction.
Two limits worth carrying. The (140 − age) term is an artefact of a 1976 curve fit: Cockcroft-Gault falls linearly to zero at 140 years while CKD-EPI decays exponentially and never reaches it, so by age 100 Cockcroft-Gault reads 58 per cent of CKD-EPI — in precisely the group being dosed most cautiously. And neither equation is valid in acute kidney injury, because both assume creatinine is at steady state. When function changes quickly creatinine lags by a day or more, so both will report reassuring numbers about a situation that is not.
How to use
- Enter age, creatinine, sex, weight and height.
- Compare like with like — the tool de-indexes CKD-EPI before comparing.
- Check whether the two land on opposite sides of a dosing threshold.
- Use whichever estimate the drug label was validated against, not a preference.
Frequently asked questions
What is the difference between creatinine clearance and eGFR?
They estimate different things by different means. Cockcroft-Gault estimates creatinine clearance in millilitres per minute using age, weight, sex and creatinine. CKD-EPI estimates glomerular filtration rate per 1.73 square metres of body surface area, using age, sex and creatinine but no weight at all. They are not approximations of each other.
Which equation should I use for drug dosing?
Whichever one the drug label was validated against, which is a question about the label rather than about the equations. A great many older drugs carry renal dosing guidance derived from Cockcroft-Gault specifically, so substituting a reported eGFR from the laboratory is not a like-for-like swap. Where the label is explicit, it decides; where it is not, pharmacy and local protocol do.
Why does Cockcroft-Gault give a higher number for heavier patients?
Because it is linear in actual body weight — the weight sits directly in the numerator. On identical age, sex and creatinine a 140 kg patient scores 2.80 times a 50 kg one, while CKD-EPI gives exactly the same figure for both because it does not use weight. That is a structural difference between the equations rather than a discrepancy to be reconciled.
Which equation reads higher?
It depends on body weight, and monotonically. Across a sweep of ages, creatinines and sexes, Cockcroft-Gault is the higher figure in about 3 per cent of cases at 50 kg, 64 per cent at 70, 95 at 90, and 100 per cent at 110 and above. The crossover sits near average body weight, which is why the two agree well enough in textbook patients for the difference to stay invisible.
Do the two estimates ever change the dose?
Often enough to matter. Sweeping plausible patients, they land on opposite sides of the 30 millilitre per minute cut about 15.9 per cent of the time and either side of 60 in 17.2 per cent. Where they disagree at 30, Cockcroft-Gault is the higher number 82 per cent of the time, so the disagreement is biased toward the permissive direction.
What does "per 1.73 square metres" mean?
It means the figure has been normalised to an average adult body surface area so that patients of different sizes can be compared. Cockcroft-Gault is not normalised, so comparing the two raw numbers compares different units. De-indexing CKD-EPI for a 190 centimetre, 140 kilogram patient moves it from 86 to 131 millilitres per minute, which is larger than most thresholds anyone cares about.
Should I use actual, ideal or adjusted body weight?
This is a genuine and unsettled argument rather than a solved question. The original 1976 paper used actual weight; a great deal of practice uses ideal or adjusted weight in obesity because fat contributes little to creatinine production; and some drug labels specify. Using actual rather than ideal weight can inflate the estimate by well over a third in a heavy patient.
Why does Cockcroft-Gault use 140 minus age?
Because it was fitted to data in 1976 and that linear term described the sample well. It is an artefact of that fit rather than physiology, and it has the odd consequence that the estimate falls to exactly zero at 140 years. CKD-EPI instead decays exponentially with age and never reaches zero, which is why the two diverge most in the very old.
How far apart do they get in elderly patients?
By age 100 Cockcroft-Gault reads about 58 per cent of CKD-EPI on the same creatinine. That matters more than it sounds, because the very old are precisely the group in whom doses are reduced most cautiously — so the equation that reads lowest is being applied where it is least well supported, and the choice between them has the largest consequence.
Are these estimates valid in acute kidney injury?
No, and this is the failure that most often matters. Both assume serum creatinine is at steady state. When kidney function changes quickly, creatinine lags by a day or more — so a patient whose kidneys have just failed still has a near-normal creatinine, and both equations will report reassuring numbers about a situation that is not reassuring at all.
What else makes creatinine estimates unreliable?
Anything that changes muscle mass, because creatinine is a muscle breakdown product. Cachexia, amputation, paralysis, advanced liver disease and simple old age all lower creatinine without any improvement in kidney function, so both equations read high. A muscular person produces the opposite error. Neither equation can see body composition, only weight and age.
Can I use this to prescribe?
No. This is a study and teaching aid for people learning how these equations differ. It knows nothing about any patient, it cannot see body composition or whether creatinine is at steady state, and a calculated number never replaces assessment, protocol or pharmacist review. It should not be the last thing consulted before prescribing anything.
🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.