BACKGROUND Plasma creatinine concentration and creatinine-based equations are most commonly used as markers of glomerular filtration rate (GFR). The abbreviated MDRD formula is considered the best available formula. Altered renal handling of creatinine, which may occur in the nephrotic syndrome, will invalidate creatinine-based formulas. We have evaluated the abbreviated MDRD formula in a large cohort of patients with proteinuria. METHODS Data on a cohort of patients with glomerular diseases were available from a large database. We have studied the relationship between estimated GFR (MDRD formula), and plasma cystatin C (CysC) and plasma beta-2-microglobulin (β2m) as markers of GFR. RESULTS The final analysis included 142 patients (93 M/49 F), median age 48 years (±15), plasma creatinine 101 μmol/L (42-368), plasma albumin 28.0 g/L (10.0-47.0), proteinuria 6.4 g/day (0.03-37.9), eGFR-MDRD4 64 mL/min/1.73 m2 (15-165), β2m 3.43 mg/L (0.7-13.8) and CysC 1.14 mg/mL (0.56-4.00). As expected, we observed a hyperbolic relationship between eGFR and both β2m and CysC. In multivariable analysis, plasma albumin concentration proved to be the most important predictor of the relationship between eGFR and both CysC and β2m. In the presence of hypoalbuminaemia, eGFR was ~ 30-40% higher at equal levels of plasma CysC or β2m. Conclusions were similar when using the recently developed CKD-EPI formula. Plasma albumin concentration did not effect the relationship between eGFR estimated by the six-variable original MDRD formula and β2m. CONCLUSIONS Our data point to discrepancies between eGFR using the six-variable MDRD formula and eGFR using the abbreviated MDRD formula as well as the CKD-EPI formula in patients with hypoalbuminaemia. One should be aware of possible limitations of creatinine-based eGFR formulas in patients with a nephrotic syndrome.