Since 2005, international guidelines propose a stadification for chronic renal failure based on the glomerular filtration rate (GFR) value. The performance of the creatinine-based equations allowing the estimation of GFR and the bias of the creatinine measurements is, more than ever, a crucial issue. The consequences for the clinical biologists are of importance. First, the Cockcroft-Gault formula must be replaced by the four variable-MDRD equation. Second, the biologists must chose from the "175" and the "186" versions of the MDRD equation. The first one fits the creatinine methods which are traceable to the reference method (liquid or gas chromatography coupled to mass spectrometry). The second equation must be used for creatinine methods, which are not traceable to the reference method. Today, only some enzymatic methods can prove that they are traceable to the reference method. For the colorimetric methods, future is inclear.