ABSTRACT
To identify a radionuclide-free, SCI (spinal cord injury) population-specific method for estimating renal function utilizing the total body clearance of gentamicin as a measure of creatinine clearance (CL(cr)) and glomerular filtration rate (GFR). To incorporate SCI population-specific patient variables into a predictive model for estimating CLcr and GFR. A retrolective study of gentamicin clearance as a measure of GFR in patients with SCI. Each patient had received a single, intravenous dose of gentamicin. Simultaneously, a 24-hour creatinine clearance (CL(cr-24)) was obtained. Experimentally measured total body clearance of a single intravenous administration of gentamicin (CL(gent-iv)) was estimated from PK analyses of gentamicin disposition in patients with SCI. Comparisons were made between CL(gent-iv) and predictors of GFR derived from a SCI population-specific nomogram (SCI-psn), the method of Cockcroft and Gault (C-G), CL(cr-24), and a multiplicative statistical model. A multiplicative model, calculated gentamicin clearance (CL(gent-calc)), that incorporates SCI population-specific characteristics to predict CL(cr-24) and GFR was derived from a multivariate nonlinear regression analysis, and the strengths of association between CL(gent-iv) and CL(gent-calc), and estimates of GFR derived from the method of C-G, SCI-psn, and CL(cr-24) were tested. The best predictive performance was demonstrated for CL(gent-iv) and the multiplicative model, CL(gent-calc), when tested against CL(cr-24) or the SCI-psn, supporting our premise that CL(gent-iv) can be used to estimate GFR in SCI. CL(gent-iv) and CL(gent-calc) outperformed CL(cr-24), the method of C-G, and the SCI-psn as predictors of GFR. The multiplicative model produced estimates of CL(gent-iv) which were more precise and less biased than CL(cr-24), the standard radiation-free predictor of GFR in patients with SCI, the SCI-psn, and the C-G equation. In this preliminary study, the clearance of gentamicin administered intravenously, CL(gent-iv), correlated highly with simultaneously measured CL(cr-24). CL(gent-iv) and our predictive model, CL(gent-calc), are more precise and less biased predictors of GFR in SCI than CL(cr-24). CL(gent-iv) and CL(gent-calc) seem to be clinically useful alternatives to CL(cr-24), the method of C-G, and the SCI-psn.
