GFR vs CrCl for Drug Dosing

Your laboratory reports eGFR. Your drug label specifies creatinine clearance. The two rarely agree, and at the extremes of body size they can differ enough to change a dose.

Which one should you dose on?

The short answer

Dose on the measure the product label specifies. For most drugs on the market that is creatinine clearance, calculated with Cockcroft-Gault.

That is not a claim that Cockcroft-Gault is the better equation. CKD-EPI 2021 is more accurate. It is a claim about where the dosing evidence came from.

Why labels still say creatinine clearance

Renal dose adjustments come from pharmacokinetic studies that stratified subjects by kidney function. The great majority of those studies — including essentially all the older ones that established thresholds still in use — used Cockcroft-Gault creatinine clearance.

The threshold on the label is therefore a statement about a Cockcroft-Gault number, not about kidney function in the abstract. "Reduce the dose if CrCl < 30 mL/min" means: patients whose Cockcroft-Gault CrCl was below 30 in the study had exposures requiring a lower dose.

Substituting a differently derived figure into that threshold breaks the correspondence. The regulatory guidance reflects this: newer studies increasingly report both measures, but the labels of established drugs still say what their studies said.

Where they diverge most

Three sources of divergence, and they can compound.

Body surface area indexing. eGFR is per 1.73 m². CrCl is absolute. A patient with a BSA of 2.3 m² has considerably more absolute clearance than their indexed eGFR suggests; a patient at 1.4 m² has considerably less.

| Patient | BSA | eGFR (mL/min/1.73 m²) | De-indexed (mL/min) | | --- | --- | --- | --- | | Large adult | 2.30 | 55 | 73 | | Standard | 1.73 | 55 | 55 | | Small, elderly | 1.40 | 55 | 45 |

Same eGFR, a 28 mL/min spread in absolute clearance. All three sit either side of thresholds that change doses.

Tubular secretion. Creatinine is secreted as well as filtered, so creatinine clearance exceeds true GFR — typically by 10–20%, and by more as function declines. CrCl is systematically the larger number, by design.

Different equations. Cockcroft-Gault and CKD-EPI were fitted to different cohorts to predict different things. They disagree even after unit alignment.

When they disagree, which is right?

Neither, exactly — both are estimates. The useful question is which is more likely to be misleading for this patient.

Trust the eGFR more when the patient is of average build and you are asking about kidney disease rather than a dose; when body weight is unreliable or unmeasured; or when cystatin C is available, making the combined CKD-EPI 2021 equation the most accurate option on the table.

Trust the CrCl more when the label specifies it; when the patient is at the extremes of body size, where indexing distorts most; and when the drug has a narrow therapeutic index and the study evidence used CrCl.

Trust neither when the patient is not in a steady state. In acute kidney injury both are invalid, because serum creatinine lags the true filtration rate.

A decision sequence

1. Check the label. Which measure, which threshold? Do not work from a remembered general rule.
2. Calculate that measure properly. For CrCl, that means deciding the weight basis deliberately — see obese patients.
3. If the label gives eGFR, use eGFR — de-indexed if the patient is far from average build.
4. Where the two straddle a threshold, treat it as a signal to be careful rather than a coin toss: consider cystatin C, consider a timed urine collection, and plan to measure drug levels if the drug allows it.
5. Document which measure and which value you used. The next clinician needs to know what the dose was based on.

Special cases

Direct oral anticoagulants. The registration trials used Cockcroft-Gault. Substituting eGFR has been shown to reclassify a meaningful proportion of patients across dose-reduction thresholds — in both directions. This is one of the clearest cases for calculating CrCl specifically.

Aminoglycosides and vancomycin. Narrow therapeutic index, and dosed against CrCl historically. Calculate CrCl for the starting dose, then let levels guide you. The calculation is a starting point, not the answer.

Chemotherapy. Carboplatin dosing via the Calvert formula requires GFR, not creatinine clearance, and the choice of GFR estimate materially affects the dose. Follow the local protocol precisely — this is one area where substituting measures has well-documented consequences.

Dialysis patients. Neither measure applies in the usual way. Residual renal function and the dialysis prescription both matter, and dosing follows dialysis-specific references.

The underlying point

The question is not "which equation is better". It is "which number was the dosing recommendation built on" — and then reproducing that number faithfully.

Use eGFR to answer questions about kidney disease. Use creatinine clearance to answer questions about doses, unless the label says otherwise. And when the patient's body size is far from average, know that the gap between the two is at its widest exactly where the decision matters most.

Both figures, the BSA conversion, and the KDIGO category are shown together on the eGFR and CrCl calculator.

Adjacent question: what happens at the threshold boundary?

Renal dosing thresholds are cliffs, but the estimates feeding them are not precise. Cockcroft-Gault carries an error of roughly ±10–30%, so a patient calculated at 32 mL/min could genuinely be anywhere from about 22 to 42.

That matters because the threshold does not care. At 31 mL/min the label says one thing; at 29 it says another.

A practical rule: treat any result within ~20% of a threshold that changes management as uncertain, and resolve it rather than rounding it. Resolving means one of:

• Recheck the weight — an estimated weight is the most common source of avoidable error.
• Confirm the patient is in a steady state and the creatinine is current.
• Check for drug interference — trimethoprim and cimetidine raise creatinine without changing filtration.
• Escalate to a measured clearance or a cystatin C-based estimate.
• Plan therapeutic drug monitoring where the drug allows it.

Documenting the decision

Whichever measure you use, record it. "CrCl 34 mL/min (Cockcroft-Gault, adjusted body weight 78 kg, Scr 1.4 mg/dL, 12 Mar)" is auditable; "renal function reduced" is not, and the next clinician cannot tell whether to recalculate or trust it.

This matters most where two measures disagree. If the laboratory eGFR says 41 and your calculated CrCl says 33, the note should say which one the dose was based on and why — otherwise the next person to review the chart has no way to reconstruct the reasoning.