Estimation of glomerular filtration rate in type II (non-insulin dependent) diabetes mellitus patients

The aim of this research was to develop an estimation of glomerular filtration rates (GFRs) from a combination of simple parameters in a large group of type II diabetic patients. We selected 122 newly presenting, previously untreated, type II patients whose GFR was determined from the plasma clearance of 51Cr-ethylenediamine tetraacetic acid (51Cr-EDTA) and simultaneous measurements of demographic variables, including fasting plasma glucose concentration, HbA1c, blood pressure, lipids, age, weight, body-mass index, body surface area, urea, and plasma creatinine concentration. The actual GFR values were compared with estimated values obtained from multiple regression and the Cockroft-Gault equations. Out of all the demographic variables, only plasma creatinine concentration (r = -0.56, p < 0.001), age (r = -0.50, p < 0.001), urea (r = -0.28, p < 0.01), and systolic blood pressure (r = -0.21, p < 0.05) showed significant correlations with the actual GFR values, for which the mean and standard deviation were 117.5 +/- 22.0 ml min-1 x 1.73 m-2. The estimated values are highly correlated with the actual values (r = 0.70), having an identical mean value of 117 +/- 15.3 and an unbiased regression relation (y = 0.000 + 1.000x). As standard measurements of the GFR are very time consuming and expensive, the use of the simple equation GFR1 = 218.1 - 0.916 x Age - 0.635 x Creatinine is recommended. The classification of GFR values into three ranges has also revealed the nonlinear characteristics of GFR in relation to other demographic variables: age and creatinine are the dominant variables in the middle GFR range, while the body-mass index and urea are dominant in the high and low ranges, respectively.