Creatinine clearance (Cockcroft–Gault)

Clinical background

Creatinine clearance (Cockcroft–Gault) estimates the renal clearance of creatinine and is used primarily for dosing drugs that are renally excreted. The equation serves a specific purpose that has not been fully taken over by newer eGFR formulae: the dosing recommendations in the summaries of product characteristics for the direct oral anticoagulants (DOACs) and many other drugs were based, in the pivotal trials, on Cockcroft–Gault, not on CKD-EPI. This means that dosing decisions for these agents formally presuppose this particular equation, even though laboratories today report eGFR by CKD-EPI as standard [1,2].

The decision the tool serves is therefore not the diagnosis of chronic kidney disease but the concrete question: what dose should this patient receive of this drug, given its renal elimination? Without an estimate of creatinine clearance, dose adjustment becomes guesswork, particularly around the thresholds at which dose reduction takes effect.

Calculating creatinine clearance (Cockcroft–Gault)

The equation estimates creatinine clearance from age, body weight, sex and serum creatinine:

$$\text{CrCl} = \frac{(140 - \text{age}) \times \text{weight}}{72 \times S_{\text{creatinine}}} \times 0{.}85_{\text{(female)}}$$

where CrCl is given in mL/min, age in years, weight in kg and serum creatinine in mg/dL. The factor 0.85 is applied for women and compensates for their lower muscle mass and hence lower creatinine production. Note that serum creatinine must be entered in mg/dL; if the laboratory reports in µmol/L, the value must be divided by 88.4 before being entered into the formula.

The derivation cohort consisted of 249 patients aged 18 to 92 years, in whom the relationship between age and 24-hour creatinine excretion per kg body weight was established. The equation was validated against the mean of two measured 24-hour creatinine clearances in 236 patients, with a correlation coefficient between estimated and measured value of 0.83 [3]. The cohort was predominantly male and consisted of patients with stable renal function on a general medical ward in Halifax, Canada, published in 1976.

Interpretation in practice

Creatinine clearance (Cockcroft–Gault) has no built-in interpretation bands in the sense that risk scores do. The result is a continuous estimate in mL/min that should be matched against the dosing recommendations for the drug in question. Concrete examples:

The thresholds vary between agents and indications. It is the summary of product characteristics, not the equation, that defines the cut-offs. The task of the calculator is to deliver the CrCl value on which the dosing decision is to be based.

Validation and performance

The original correlation of r = 0.83 with measured creatinine clearance [3] has subsequently proved optimistic for several reasons. One systematic weakness is that Cockcroft–Gault was developed before the standardisation of creatinine assays against isotope dilution mass spectrometry (IDMS), introduced worldwide around 2010. Since standardisation, serum creatinine values have fallen somewhat, which systematically inflates the estimated creatinine clearance compared with what applied in the derivation cohort [1].

In a Dutch retrospective cohort study of 455 patients with diabetes and measured 24-hour creatinine clearance as the reference, Cockcroft–Gault performed better than both MDRD and CKD-EPI in overweight and obese patients. The bias of Cockcroft–Gault fell from −13.4 mL/min at a BMI below 25 to −3.2 mL/min at a BMI above 30, while the bias of CKD-EPI increased in absolute terms to −29.6 mL/min in the same group. With an accepted dispersion of 30 per cent, Cockcroft–Gault had an accuracy of 76.8 per cent in obese patients, compared with 37.3 per cent for CKD-EPI [4]. This is an important argument for retaining Cockcroft–Gault specifically for dosing in overweight patients, even though CKD-EPI is otherwise superior for classifying renal function.

In intensive care, performance is poorer. In a multinational multicentre study of 383 intensive care patients with a total of 1,708 measured creatinine clearance values, Cockcroft–Gault performed with moderate correlation (r = 0.33 to 0.39) and moderate accuracy (48 to 58 per cent within 30 per cent of the measured value) in patients without augmented renal clearance. Performance was best with adjusted body weight. In patients with augmented renal clearance (defined as a measured CrCl ≥ 130 mL/min/1.73 m²), the correlation was low (r = 0.24 to 0.28) and the bias high, whichever body weight parameter was used [5].

A retrospective study of DOAC dosing in atrial fibrillation found dosing discordance between Cockcroft–Gault and CKD-EPI in 8 per cent of patients correctly dosed by Cockcroft–Gault for rivaroxaban, and in 3 per cent for dabigatran. Discordance based on CKD-EPI rather than Cockcroft–Gault was associated with an increased risk of thromboembolism for rivaroxaban (OR 2.83, 95 per cent CI 1.02 to 7.79) [2]. This underlines that the dosing criteria in the summaries of product characteristics cannot be arbitrarily replaced with eGFR.

Limitations

Stable renal function is a prerequisite. The equation presupposes a steady state between creatinine production and creatinine elimination. In acute kidney injury, serum creatinine is a delayed marker reflecting the glomerular filtration of hours to days earlier, and Cockcroft–Gault then systematically overstates the actual clearance. In intensive care, augmented renal clearance is also often present, where the equation performs poorly whichever body weight parameter is used [5].

Extremes of body habitus. In obesity, the term (140 − age) × weight overstates clearance, since adipose tissue does not produce creatinine. Adjusted body weight is often recommended at a BMI above 30, and ideal body weight may be considered in marked obesity. In undernutrition or sarcopenia the serum creatinine is low because of reduced muscle mass, which can give a misleadingly high estimate of clearance despite genuinely impaired filtration. This is particularly relevant in older, frail patients.

Creatinine assay standardisation. Because the equation was derived against non-standardised creatinine assays, it can deviate appreciably from the eGFR calculated with CKD-EPI, which is aligned to IDMS-standardised methods [1]. The deviation is not constant but depends on the serum creatinine level: at low creatinine values the overstatement of clearance is more pronounced.

Tubular secretion. Creatinine clearance includes not only glomerular filtration but also tubular secretion of creatinine (20 to 30 per cent), so creatinine clearance systematically overstates the actual GFR by 10 to 20 per cent [1]. This is a property of creatinine clearance as such, not specifically of Cockcroft–Gault, but it explains why dosing recommendations based on Cockcroft–Gault may sit higher than the true GFR would suggest.

The sex factor. The fixed multiplier of 0.85 for women rests on an assumption of 15 per cent lower muscle mass. Individual variation in body composition means that this correction may be either too large or too small.

Cockcroft–Gault alongside reported eGFR

Laboratories generally report eGFR by CKD-EPI as standard, not creatinine clearance by Cockcroft–Gault. Clinicians dosing DOACs or other renally dose-adjusted drugs must therefore often calculate Cockcroft–Gault separately. The summaries of product characteristics (SmPCs) and prescribing information for the DOACs state dosing criteria based on CrCl by Cockcroft–Gault, and these apply notwithstanding that eGFR (CKD-EPI) is the primary measure for classifying chronic kidney disease. Where there is uncertainty about which method underlies a dosing recommendation, the summary of product characteristics should be consulted.

References

1. Sandhu G et al. Aligning kidney function assessment in patients with cancer to global practices in internal medicine. EClinicalMedicine 2025. PMID: 40290845
2. Bhalodia NJ et al. DOAC Dosing Discordance Using Different Estimates of Kidney Function and Outcomes. Journal of Clinical Pharmacology 2023. PMID: 37178305
3. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976. PMID: 1244564
4. Drion I et al. The Cockcroft-Gault: a better predictor of renal function in an overweight and obese diabetic population. Obesity Facts 2011. PMID: 22166760
5. Cucci MD et al. Performance of different body weights in the Cockcroft-Gault equation in critically ill patients with and without augmented renal clearance: a multicenter cohort. Pharmacotherapy 2023. PMID: 36373197