Estimated GFR (MDRD 4-Variable, IDMS-Traceable)

The Modification of Diet in Renal Disease study equation came out of a 1999 analysis by Andrew Levey's group, refit in 2005 for IDMS-traceable creatinine assays, which is the 175 coefficient this page uses. It is a straight power-law regression: creatinine to the −1.154, age to the −0.203, times 0.742 for women. A 55-year-old man with a creatinine of exactly 1.0 mg/dL scores $175 \times 55^{-0.203} = 77.6$ mL/min/1.73 m².

Read the units carefully, because they are the most misunderstood thing about eGFR. The result is already normalised to a body surface area of 1.73 m², the average American adult of the 1920s. It is not this person's filtration rate; it is what their filtration rate would be if they had an average-sized body. For population staging that normalisation is the point, since it makes people comparable. For drug dosing it is a trap, because a 45 kg woman and a 120 kg man with the same eGFR do not clear a drug at the same absolute rate. Multiplying back by the patient's own BSA and dividing by 1.73 recovers the un-normalised figure.

MDRD has two well-documented limits. It was derived in people with chronic kidney disease, so it systematically underestimates in healthy kidneys and should not be reported as a number above 60; laboratories that follow the guidance print "greater than 60" instead. And like every creatinine-based estimate it assumes a steady state, so it is not valid in acute kidney injury, in pregnancy, or at the extremes of muscle mass. The CKD-EPI equations, and the 2021 race-free revision in particular, track measured GFR better across the normal range and have replaced MDRD in most laboratories. MDRD is here because a great deal of published dosing guidance still quotes it.