Estimation of Glomerular Filtration Rate Based on the Modification of Diet in Renal Disease Equation in Patients with Chronic Renal Failure

Abstract

The list of indications for initiating regular dialysis treatment includes residual glomerular filtration rate (GFR). Under the current European Best Practice Guidelines for Hemodialysis, residual GFR (and the presence of one or more symptoms of uremia) should not decrease below 15 ml/min. The present article seeks to determine to what extent the modification of diet in renal disease (MDRD) equation enables the detection of this decrease in GFR. We tried to answer this question using a more detailed analysis of the relationship between MDRD and renal inulin clearance (C_in). Residual GFR based on C_in (under conditions of stable plasma levels and water loading) and GFR calculated using the MDRD equation was measured in 79 individuals with chronic renal failure (with mean C_in = 19.1 ± 10.1 ml/min/1.73 m²). Statistical evaluation was performed using regression analysis, the interchangeability of both methods (Bland-Altman) and receiver-operating characteristic (ROC) curve analysis. Regression analysis demonstrated a significant correlation between MDRD and C_in (r = 0.892; p < 0.001). However, the regression equation line for the correlation differs significantly from the identity line (p < 0.001). The value of the regression coefficient (0.722) is significantly lower than 1.0 (CI₅₀ 0.63; 0.81). The mean MDRD – C_in difference was 3.26 ± 4.46 ml/min/1.73 m² and the value was significantly different from zero (p < 0.001). The mean difference +2 SD was 12.2 ml/min/1.73 m², and the mean – 2 SD was –5.7 ml/min/1.73 m². ROC curve analysis (for a cutoff C_in = 15 ml/min/1.73 m²) indicates an area under the curve (AUC) of 0.954 ± 0.023. The best combination of sensitivity and specificity was obtained for a MDRD of 19.7 ml/min/1.73 m², with a sensitivity of 90.5% and specificity of 87.5%. For cutoff value of C_in = 10 ml/min/1.73 m², the AUC was 0.939 ± 0.026 (CI₉₅ 0.863–0.890). A combination of maximum sensitivity and specificity was obtained with an MDRD of 16.5 ml/min/1.73 m². With this value, MDRD sensitivity was 100% and specificity 81.5%. A significant correlation between the MDRD equation and the measured creatinine clearance (C_cr) was found (r = 0.883, p < 0.001). The mean difference of MDRD – C_cr was –7.2 ± 6.5 ml/min/1.73 m². This is significantly different from that of MDRD – C_in (p < 0.001). Our results suggest that MDRD and C_in in individuals with chronic renal failure are not interchangeable methods for a GFR <15 ml/min/1.73 m² determination. However, MDRD may furnish valuable information in terms of detecting a critical decrease in GFR; but, the MDRD equation for this decrease in GFR (15 ml/min/1.73 m²) will provide a somewhat higher value (19.7 ml/min/1.73 m²).