Once upon a time, doctors had to estimate GFR (glomerular filtration rate, which relates to kidney function) from the creatinine using an app or even a calculator. Now almost all labs provide estimated GFR in blood test results, but this is both a positive and a negative. I rarely find a resident (or attending physician) who understands the principles behind eGFR, and thus too often they interpret the number incorrectly and use it wrong. Here is my quick guide for understanding eGFR.
Back in the day, before the turn of the century, we measured 24-hour urines. We would measure the volume over 1440 minutes and measure the urine creatinine level. Given these numbers and the serum creatinine level, we would plug the numbers into the classic formula UV/P.
This method had the major problem of actually collecting an accurate 24-hour urine.
Cockroft and Gault published this article in 1976: Cockcroft, D.W. and M.H. Gault. Prediction of creatinine clearance from serum creatinine. Nephron. 1976. 16(1):31-41.
The formula Creatinine clearance = (140-age)*wt kg / Scr * 72 for men and 85% of that number for women was the commonly used for many years.
Two other formulas based on the same reasoning are likely familiar to many readers – the MDRD formula (used by most labs) and the CKD-EPI equation.
What are these formulas all doing? They are estimating daily creatinine production! The numerator of the creatinine clearance formula, while the creatinine is stable, represent daily creatinine production. In steady state situations, virtually all the creatinine produced each day from muscle breakdown gets excreted in the urine.
We can predict muscle mass in several ways. Muscle mass estimation allows us to estimate daily creatinine production and hence the numerator of our formula.
MDRD and CKD-EPI both estimate muscle mass from age, gender and race. They assume a reasonable estimation from those demographics.
So we have two constraints on the use of eGFR:
• Serum creatinine has not changed and is not changing
• The patient has a “normal” muscle mass
So here are the cautions. If the serum creatinine is either increasing or decreasing then eGFR is meaningless and should not be reported or considered. Here is the thought experiment. Mr. Jones has a serum creatinine of 1. We remove both his kidneys. Tomorrow his creatinine is 2. We know that he has no GFR, yet the formula would suggest a GFR 1/2 of the previous days. So just report that the serum creatinine is increasing or decreasing, but do not report eGFR.
We have two situations in which patients clearly do not have “normal” muscle mass. The first occurs in athletes with increased muscle mass – body builders, many football players, etc. Our formulas underestimate the numerator. The denominator has increased because they do make more creatinine each day. Thus, we will underestimate their creatinine clearance with the eGFR formula. Sometimes this would label these individuals as having CKD. In this situation, you will need to measure a 24-hour urine creatinine collection.
The second occurs in patients with decreased muscle mass. Some such situations include spinal cord injury patients, major amputations, muscle diseases, and severe cachexia. We recently had a patient with a spinal cord injury and a serum creatinine of 0.4. The formula suggested that his eGFR was approximately 200. The formula was wrong. His numerator was actually much lower than the formula would predict.
Before you use the serum creatinine and the eGFR to make clinical decisions, you should understand these issues. As clinician educators, we must remain vigilant and make certain that our learners understand the reasoning behind eGFR. It really bugs me that all our students and residents do not know this. This speaks to inadequate education.
Robert M. Centor is an academic general internist at the Alabama School of Medicine and the associate dean for the Huntsville Regional Medical Campus of UASOM. He uses his off-the-cuff, oftentimes funny doctor’s blog as a sounding board for ethical debates and daily musings. Check out his blog at http://www.medrants.com/