Testing the predictive ability of the “spinal cord injury equation” in estimating vancomycin clearance

Abstract

A new method of estimating drug clearance in patients with spinal cord injury (SCI) was tested against other methods through a retrospective analysis of its predictive ability in estimating vancomycin clearance. A retrospective chart-based investigation was conducted at a hospital with a large SCI population. Using data collected from electronic patient records, 179 cases were reviewed, and 87 cases met the inclusion criteria: a diagnosis of longstanding SCI, administration of vancomycin via i.v. infusion during the 18-month study period, and at least one documented steady-state vancomycin concentration. Using the "spinal cord injury equation" (i.e., CL(SCI) equation), pharmacokinetic analysis was performed to determine patient-specific clearance values, which were compared to values resulting from five widely used methods of estimating the glomerular filtration rate in spinal cord-injured patients. The primary outcome measures were bias and precision, as indicated by statistical analyses to determine the mean prediction error (ME) and the square root of the mean squared prediction error (RMSE) of each method. Compared with the other evaluated methods, the CL(SCI) equation was found to be less biased and more precise, with the smallest calculated ME and RMSE values (p < 0.05). The five alternative methods significantly overestimated vancomycin clearance, by 45-92% (p < 0.05). The CL(SCI) method underestimated vancomycin clearance (by 6%) but not to a significant degree (p = 0.06). The study results suggest that the CL(SCI) equation for predicting vancomycin clearance was unbiased and may be more precise relative to other frequently used methods in the study population of patients with long-term SCI.