gms | German Medical Science

29. Wissenschaftlicher Kongress der Deutschen Hochdruckliga

Deutsche Hochdruckliga e. V. DHL ® - Deutsche Hypertonie Gesellschaft Deutsches Kompetenzzentrum Bluthochdruck

23. bis 25.11.2005, Berlin

Predictive performance of Cockroft-Gault and MDRD equations to assess renal function in a population with normal serum creatinine

Wertigkeit der Cockroft-Gault und MDRD Formel zur Beurteilung der renalen Funktion in eine Population mit normalem Serum Kreatinin

Meeting Abstract

  • M.P. Schlaich - Universitätsklinikum Erlangen-Nürnberg (Erlangen, D)
  • B. Schmidt - Universitätsklinikum Erlangen-Nürnberg (Erlangen, D)
  • I. Fleischmann - Universitätsklinikum Erlangen-Nürnberg (Erlangen, D)
  • R.E. Schmieder - Universitätsklinikum Erlangen-Nürnberg (Erlangen, D)

Hypertonie 2005. 29. Wissenschaftlicher Kongress der Deutschen Hochdruckliga. Berlin, 23.-25.11.2005. Düsseldorf, Köln: German Medical Science; 2006. Doc05hochP91

The electronic version of this article is the complete one and can be found online at:

Published: August 8, 2006

© 2006 Schlaich et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License ( You are free: to Share – to copy, distribute and transmit the work, provided the original author and source are credited.



Background: Accurate estimation of renal function is important for the diagnosis of kidney disease and assessment of cardiovascular risk. Various equations have been proposed to predict glomerular filtration rate (GFR), the most frequently applied being the Cockroft-Gault (CG) and the MDRD2 equations. The predictive performance of these equations in a population with normal renal function has not yet been validated sufficiently.

Methods: Inulin-Clearance was performed in a total of 302 subjects with serum creatinine <1.3mg/dl to measure GFR. CG and MDRD2 equations were applied to predict GFR in these subjects. Accuracy of each equation was assessed by comparing its results with those of the inulin-clearance, according to the formula: (predicted value - inulin-clearance)x(100/inulin-clearance). Precision (R²statistics) and bias (mean prediction error) were also calculated.

Results: The majority of the 302 study participants were male (n=214) with a mean age of 46±15 years. The study population included patients with hypertension (n=72), diabetes (n=103), hypercholesterolemia (n=22) and healthy volunteers (n =105). None of the participants had a serum creatinine above 1.3 mg/dl (mean: 0.85±0.16; range: 0.43-1.3mg/dl). Body mass index was 26.8±4.4kg/m². Measured GFR (inulin-clearance) was 112±20ml/min/1.73m². Estimated GFR was 106±25 ml/min/1.73m² with CG and 98±22 ml/min/1.73m² with MDRD2 equations, respectively. The number of subjects with predicted GFR values within 30% and 50% of the inulin GFR was 74.6% and 92.4% with CG and 78.2% and 96.6% with MDRD2 equation, respectively. Precision was higher for MDRD2 (0.08) compared to CG (0.02). Bias was greater for MDRD 2 than for CG (14.9 vs 7.35).

Conclusions: Analysis of the predictive performance of renal function equations in our large study cohort suggests that the MDRD2 equation is more accurate and precise but also more biased than the CG formula in predicting GFR in subjects with normal serum creatinine.