Should Cockcroft-Gault Take A Hike?

When estimating a patient’s renal function to determine an appropriate medication dose, many clinicians turn to the Cockcroft-Gault (CG) equation. But is this equation the best method to quickly assess and characterize renal function?   Several studies have compared various methods to predict renal function.^1,2,3  The most accurate way to measure creatinine clearance (CrCl) requires a 24-hour urine collection; however this method is not routinely used in clinical practice because it is cumbersome, more expensive, and time consuming.¹  Equations, such as Cockcroft-Gault (CG), that estimate creatinine clearance based on readily available laboratory data are well established, simple, and widely used in practice.  However, there is much debate about the ability of these equations to accurately assess renal function.  CG, the most widely used equation, has been validated but it has several limitations.  The serum creatinine (Scr) measurement itself is affected by age, sex, race, weight, muscle mass, diet, drugs, and the laboratory methods used.  It has long been known that the CG equation can overestimate or underestimate renal function depending on the weight use in the equation (e.g. actual body weight, lean body weight, or adjusted body weight).   

In addition to CG, there are several other equations that can be used to estimate renal function. These include the Modification of Diet in Renal Disease (MDRD) and the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation.²  The difference between these equations and the CG equation is that the MDRD and CKD-EPI estimate the glomerular filtration rate (GFR) whereas the CG equation was designed to estimate CrCl.  The Salazar-Corcoran equation was developed specifically to calculate CrCl in obese patients.¹  However, the Salazar-Corcoran equation is not widely used in practice nor has it been validated in pharmacokinetic studies.

Winter, Guhr, and Berg conducted a retrospective analysis to determine the impact of various body weights on the accuracy of the CG equation compared with measured 24-hour CrCl.  Patients were classified as underweight (BMI <18.5 kg/m2), normal weight (BMI 18.5-24.9 kg/m2), overweight (BMI 25-29.9 kg/m2), obese (BMI 30-39.9 kg/m2), or morbidly obese (BMI ≥ 40 kg/m2).  In addition, the authors evaluated the effect of rounding up SCr values to either 0.8mg/dL or 1mg/dL — a common practice when the patient’s SCr is lower than the specified threshold.  Patients were included if they were at least 18 years old and had a 24-hour urine collection performed between 1996 and 2010.  Patients were excluded if they had unstable renal function (>20% change in serum creatinine during their hospital admission), if they had a limb amputation during the hospitalization, or if they were pregnant (or admitted to a hospital location that suggested they might be pregnant).  There were 3678 patients evaluated, one-fourth of the patients were male; 38% were obese; roughly 20% were normal weight, overweight, and morbidly obese; and a mere 2% were underweight. 

The authors defined “equation bias” as the mean difference between the estimated CrCl for each calculation (using various weight and SCr values) and the measured CrCl.¹ Unbiased results were those where the mean calculated CrCl was not significantly different than the measured CrCl.  The authors found that using actual body weight in the CG equation for underweight patients yielded an unbiased estimate of CrCl when compared to the measured CrCl (p=0.898).  In normal weight patients, ideal body weight yielded in an unbiased estimate of CrCl compared to the measured 24 hour urine collection (p=0.544).  For the overweight, obese, and morbidly obese patients, an unbiased measurement of CrCl could not be calculated, regardless of the weight value used.  In the overweight patients, using adjusted body weight with a factor of 0.4 yielded the least biased estimate and the difference was unlikely to be clinically important.  In the obese and morbidly obese populations, all weights (ideal, actual, and adjusted) yielded significantly biased estimates.  Although the results indicate there were very significant statistical differences between the estimated CrCl using CG and the measured CrCl, the clinical significance of these differences was not evaluated.  In patients of all ages, including those 65 years and older, using actual SCr values resulted in an unbiased CrCl estimated.  Rounding up SCr values did not improve the accuracy of the CrCl estimate. 

This study has several limitations – including, most importantly, a retrospective design.  24-hour urine collection is cumbersome and subject to inaccuracies; however, most patients’ urine was collected either by indwelling catheter or under the supervision of a health care professional.  Finally, there were a relatively small number of patients aged 65 and older with a low SCr value.

Other studies support the findings in this study.  Wilhelm and Kale-Pradhan conducted a meta-analysis of published trials to determine which body weight most accurately predicts measured CrCl as well as the impact of rounding up SCr values to 1mg/dL when the value is less than 1mg/dL.² The authors concluded that the CG equation with no body weight (assuming a body weight of 72 kg) and actual SCr value most closely estimated measured CrCl.  In obese patients, actual body weight with a correction factor of 0.3 or 0.4 was also reasonable.  These authors recommend against the use of actual body weight, ideal body weight, and rounding serum creatinine value in the CG equation.  Recently, Dowling and colleagues conducted a prospective study to evaluate the performance of the MDRD and CKD-EPI when compared to measured 24-hour urine clearance.³ The authors concluded that the MDRD and CKD-EPI equations overestimated CrCl in elderly individuals and this may lead to dose calculation errors for many drugs.

These data indicate that the CG equation remains the most useful, easiest, and most accurate equation for estimating renal function and making dosing adjustments in patients with renal dysfunction.  The biggest question that remains is what value to use for body weight in overweight and obese patients.  Based on the data, we believe it is reasonable to use an adjusted body weight using a factor of 0.4.  In our opinion, more evidence is needed using the Salazar-Corcoran equation before it can be recommended over CG to estimate CrCl for overweight and obese patients.   And finally, clinicians should stop rounding up SCr values – don’t do it!  What do you do in your practice?