Predicting the glomerular filtration rate from serum creatinine, serum cystatin C and the Cockcroft and Gault formula with regard to drug dosage adjustment.

OBJECTIVES: To compare the accuracy of the classification of the degree of decrease in glomerular filtration rate (GFR), measured exactly (as inulin clearance) on the basis of serum concentrations of creatinine (Scr), cystatin C (Scyst) and creatinine clearance predicted according to Cockcroft and Gault (CG), and to establish whether any of the above methods is more accurate than the other 2.
SUBJECTS: The study was conducted in 126 patients (52 men, 74 women) aged 18 to 64 years with various chronic renal diseases (predominantly various forms of glomerulonephritis and tubulointerstitial nephritis). The study subjects were divided into 3 subgroups according to GFR levels. Subgroup A (n = 41) included individuals with GFR >50 ml/min/1.73 m2, subgroup B (n = 56) was made up by individuals with GFR of 20-50 ml/min/1.73 m2, while subgroup C (n = 29) comprised individuals with GFR <20 ml/min/1.73 m2.
METHODS: GFR was determined on the basis of renal inulin clearance (Cin) under conditions of stable plasma levels and water loading. Each individual had his/her Scr, Scyst values measured and CG was calculated. Results were evaluated using discrimination analysis.
RESULTS: Mean values and SD of the monitored markers in the subgroups were as follows. Subgroup A: Scr 102.4 (38.3) micromol/l, Scyst 1.46 (0.42) mg/l, CG 80.0 (19.2) ml/min/1.73 m2. Subgroup B: Scr 161.2 (45.6) micromol/l, Scyst 2.01 (0.55) mg/l, CG 46.1 (16.7) ml/min/1.73 m2. Subgroup C: Scr 314.9 (58.3) micromol/l, Scyst 3.41 (0.96) mg/l, CG 24.8 (7.6) ml/min/1.73 m2. The percent of correct classifications and the respective confidence intervals (95%) for the methods used were as follows. Subgroup A: Scr 79.3 (64.6, 94.0), Scyst 75.9 (60.3, 91.5), CG 86.2 (73.6, 98.8). Subgroup B: Scr 51.8 (35.5, 68.1), Scyst 57.1 (41.5, 72.7), CG 64.3 (48.6, 80.0). Subgroup C: Scr 90.2 (81.0, 99.2), Scyst 80.5 (68.1, 92.9), CG 87.8 (77.8, 97.8). The percent of correct classifications established on the basis of Scr, Scyst and CG in subgroup B is significantly lower than that of correct classifications in subgroups A and C (p < 0.05-0.001). The percent of correct classifications using Scr, Scyst and CG, estimated separately for each subgroup (A, B, C) does not differ significantly.
CONCLUSIONS: The findings support the assumption that estimation of the decrease in GFR using Scr, Scyst and CG is, as regards their utilization in everyday practice, suitable for individuals with severely decreased GFR (<20 ml/min/1.73 m2) and for individuals with a decrease in GFR to levels >50 ml/min/173 m2. Estimation of the decrease in GFR using the above subgroups did not demonstrate significant differences among Scr, Scyst and CG. Using the above markers, estimation of the decrease in GFR is the least reliable with GFR values in the range of 20-50 ml/min/1.73 m2.