In this episode I’ll discuss a long time pet peeve of mine – the Cockcroft Gault formula can be used to predict – not calculate – creatinine clearance.
40 years after it’s publication, the Cockcroft Gault formula remains the most appropriate method of estimating kidney function for the purpose of renal dose adjustment of medications. I use the formula dozens of times each day.
Even though in 2010 the FDA changed it’s guidance for industry to allow using MDRD estimated GFR for drug dosing, I’m not aware of any medications that use the MDRD formula in place of Cockcroft-Gault for renal dose adjustment.
The Cockcroft-Gault formula is: (((140-age)*weight in kg)/(serum creatinine*72))*0.85 if female
The full text of the 1976 study by Cockcroft and Gault is required reading for my pharmacy students and residents. I haven’t found it available for free online so you’ll need to check with your medical librarian to get a copy.
Rapid, non-invasive bedside prediction of kidney function is essential for patient care and medication dosing. 24 hour urine creatinine clearance calculations are impractical for many reasons including being time-consuming and error-prone.
History
Donald W. Cockcroft is an asthmatologist who teaches at the University of Saskatchewan.
In 1973 he was finishing a 3 month nephrology rotation with M. Henry Gault. As part of his nephrology rotation, he completed a research project to verify the accuracy of a nomogram that predicted creatinine clearance based on age, weight, and serum creatinine. Cockcroft and Gault were reviewing the data and the negative linear correlation between age and creatinine clearance when Cockcroft realized this could be turned into a formula to predict creatinine clearance. In a 1992 interview Cockcroft stated:
This formula yields results which are in reasonably good agreement with measurements obtained from 24-hour urine collections. It thus allows a quick and easy assessment of creatinine clearance as long as serum creatinine (thus renal function) is in a steady state.
I find that the qualifiers from the original study “as long as serum creatinine is in a steady state” are often forgotten. Because it is presented as a formula, it can be tempting to think of use of the formula as “calculating” creatinine clearance.
Study Characteristics
Let’s examine the characteristics of the patients in the original Cockcroft Gault study:
Creatinine at steady state
All patients had two 24 hour urine creatinine measurements that did not differ by more than 20%. This means all patients had steady-state serum creatinine and renal function.
Male
The 249 patients evaluated in the study were male. The idea of reducing the value from the formula by 15% in female patients is “expert opinion”.
Body weight
Total body weight was used in the study, but essentially all patients were very close to their ideal body weight.
Age
The range of patient ages in the study was 18-92, but 75% of the patients in the study were less than 70 years old.
With all of these caveats, the equation predicted creatinine clearance to +/- 17% of the actual creatinine clearance. This means if the formula predicts a creatine clearance of 50 mL/min, the actual value is between 42 and 58 – which could easily straddle a decision point for reducing a medication dose due to renal insufficiency.
Summary
To summarize, the Cockcroft-Gault equation was validated primarily in male patients under the age of 70, with stable renal function who were close to their ideal body weight. Adjustments to the weight used in the formula, rounding the serum creatinine, and reducing the value empirically for female patients are based on expert opinion and not validated in the original study.
Application to clinical practice
How many of your hospitalized patients match the characteristics of the patients in the original Cockcroft-Gault study?
Most of my ICU patients would not have met the inclusion criteria for the study based on their creatinine not being at steady state.
The Cockcroft-Gault formula remains the best at determining the need for dose adjustment in renal insufficiency, but how should it be applied to patients?
I think an assessment of the risk vs benefit of adjusting a medication for renal insufficiency should be made. I discussed making patient focused risk:benefit assessments in episode 11.
Many critically ill patients are not at steady state renal function. In such patients I will take the trend in serum creatinine and combine that with the results of the Cockcroft-Gault formula.
For example, in a patient whose creatinine has increased from 1 to 1.5 mg/dL and Cockcroft-Gault estimates a creatine clearance of 40 mL/min I will think “The creatinine clearance is worse than 40 mL/min” since the patient’s creatinine is trending up.
Likewise in a patient whose creatinine has decreased from 2 to 1.5 mg/dL and Cockcroft-Gault estimates a creatine clearance of 40 mL/min I will think “The creatinine clearance is better than 40 mL/min” since the patient’s creatinine is trending down.
There are two reasons to adjust a medication based on renal insufficiency:
1. To prevent toxicity
2. To save money
In a critically ill patient, I usually err on the side of not reducing a medication dose unless the goal of reducing the dose is to prevent toxicity.
Take for example piperacillin-tazobactam in a patient with sepsis and acute renal insufficiency. The risk of dosing high for 24 hours is lower than the risk of dosing low. So I’ll often wait to renally adjust the dose until the 2nd serum creatinine result is back and I can get a better idea of how the patient’s kidneys are functioning.
If the goal of reducing the medication dose is to prevent toxicity, then I usually err on the side of reducing the dose or monitoring when possible.
Take for example vancomycin in the same patient with sepsis and acute renal insufficiency. Since vancomcyin is nephrotoxic, I may give single doses and check levels before the next dose to avoid making assumptions that could lead to further kidney injury.
I’ve created a simple, mobile responsive creatinine clearance estimator at pharmacyjoe.com/clcr based on the Cockcroft-Gault formula. I haven’t included multiple equation variations or body weight adjustments because in the end, it is just an estimate and should be treated as such.
If you like this post, check out my book – A Pharmacist’s Guide to Inpatient Medical Emergencies: How to respond to code blue, rapid response calls, and other medical emergencies.
It is important to realize that Cockcroft-Gault is in essence a calculation of the creatinine generation rate. Dividing that by the serum creatinine gives you the clearance in steady state conditions.
Estimating the crea generation rate (by whatever method) is helpful for calculating clearance in the non-steady state too. This is nicely described by Winter in Basic clinical pharmacokinetics for example.
For an excel based calculator of crea clearance and GFR in non-steady state conditions check out my blog post at
http://www.swissnephro.org/blog/2015/3/23/estimating-the-dynamic-gfr-when-creatinine-concentrations-are-changing-rapidly
Greetings, Florian
I would like to voice a different view regarding the use of Cockroft Gault for drug dosing.
I believe following eGFR equations, such as MDRD, is usually at least as reasonable as using CG for dose adjustment. This approach is endorsed by the National Kidney Disease Education Program (NKDEP), and I will mention a few especially salient points from their summary of the subject:
– Studies have demonstrated relatively high concordance rates between dose adjustment requirements that would arise from using the different estimation methods (Stevens et al, Used data from 5500 patients demonstrating that concordance rates between the CG and CG(IBW) equations and MDRD Study equation were 89% and 88%). (I understand the remaining 11-12% variation can be a cause of great concern)
– Both the CG formula, as well as the vast majority of currently available drug dosage recommendations are based on serum creatinine results prior to standardization. There was previously a large laboratory dependent variability in serum creatinine results, and the relationship between creatinine results before and after standardization will be different for each specific method and instrument used. The CG equation as well cannot be re-expressed for IDMS-traceable creatinine values. The creatinine method used in the development of the equation is no longer in use and samples from the study are not available.
– eCrCl has more variability than eGFR. Cockcroft-Gault eCrCl has only 50 to 70 percent of results within 30 percent of measured GFR (vs. 83 percent for eGFR).
See more: http://www.niddk.nih.gov/health-information/health-communication-programs/nkdep/a-z/ckd-drug-dosing/Pages/CKD-drug-dosing.aspx
KDIGO guidelines have similarly recommended using either of the equations for the purpose of drug dosing: http://www.ncbi.nlm.nih.gov/pubmed/21918498
I believe the overall non-inferiority of eGFR for dose adjustment can also be demonstrated with newer medications. Analyses have been published for clinical outcomes with the new direct anticoagulants in relation to renal clearance assessed using different methods, see for example:
http://www.ncbi.nlm.nih.gov/pubmed/24323795
http://www.ncbi.nlm.nih.gov/pubmed/22933567
Regardless of what equation is used, I believe they only provide a relatively crude approximation. I therefor wholeheartedly agree with your emphasis on the need to understand the assumptions of these estimations and make a clinical decision based on the specific situation while taking into account patient characteristics and the risk/benefit of over/under dosing.
p.s. I have no knowledge or opinion regarding the medicolegal ramifications of following the guidelines I’ve mentioned, which sometimes contradict recommendations in the medication labeling
Excellent review Amichai, thank you for sharing!
In Pharmacy school (Oregon State University) we were taught that extremes of age required an estimation of creatinine to be increased to 1mg/dl if creatinine measured lower than 1mg/dl. Although I understand the theory behind it (less muscle mass in the elderly) I have not been able to find actual studies that support this practice. Is this common practice all throughout the word of Pharmacy?
Hi Jessica! My understanding is that the practice of rounding up the serum creatinine in elderly patients results in underestimating creatinine clearance. I do not use rounding in my practice rather I use an individualized risk:benefit assessment each time I consider renally adjusting medications.
Why 140 is used in equation??