Quantitative evaluation of contrast-induced-nephropathy in vascular post-angiography patients: Feasibility study of a semi-empirical model.

In this study, the contrast-induced-nephropathy (CIN) of vascular patients who had undergone angiography was quantitatively evaluated using a semi-empirical model. The model compiled six essential serum readings and biological data from 70 patients in order to develop a 1st-order nonlinear equation with 16 defined terms. The expectation value of the model was used to predict the serum creatinine reading of patients that had been determined to be at high risk of CIN after contrast media (CM) administration. The other five variables included body surface area (BSA), administrated CM, serum creatinine level before CM administration, blood urea nitrogen (BUN), and systolic blood pressure level. A loss function was used to define the difference between the observed and predicted serum creatinine readings after CM administration. The dominant variables were proven to be systolic blood pressure, serum creatinine level before CM administration, and BSA. The cross interaction between the serum creatinine level before CM administration and systolic blood pressure was the decisive term of the model's performance, indicating that both should be specially considered in order to prevent CIN. The BSA, which was usually ignored by medical staff, was also proven to be a significant variable, whereas the BUN reading and amount of injected contrast media were negligible in the semi-empirical model.