Development of a Highly Accurate Nomogram for Prediction of the Need for Exploration in Patients With Renal Trauma

Abstract

To develop and validate a nomogram for predicting the need for renal exploration after renal trauma. From 1995 through 2004, 419 consecutive patients presented to our institution with traumatic renal injury. All were randomly divided into a development (50%, n = 210) and a split sample validation cohort (50%, n = 209). Logistic regression models were used to develop a nomogram for prediction of the need for renal exploration after renal trauma. Internal (200 bootstrap resamples) and 50% split sample validations were performed. Overall, 89 patients (21.2%) underwent renal exploration, from which 60.7% (54 of 89) underwent nephrectomy and 39.3% (35 of 89) underwent renorrhaphy. Nine percent of patients with grade II injury underwent renal exploration, 16% with grade III injuries, 41% with grade IV injuries, and 100% of grade V injuries. The kidney injury scale, the mechanism of injury, the need for transfusion, blood urea nitrogen level, and serum creatinine represented the most informative predictors and were included in the nomogram. The split sample accuracy of the nomogram for prediction of the need for renal exploration was 96.9%. It significantly (p < 0.001) exceeded the accuracy of each of its components including the American Association for the Surgery of Trauma kidney injury scale (87.7%). The nomogram generates highly accurate and reproducible predictions of the probability for renal exploration according to our decision-making. It could help standardize the management of patients with renal trauma (i.e., inclusion criteria for clinical trials) and serves as a proof-of-principle that predictive tools can be applied to the trauma setting. Its use may improve the management of renal trauma patients at institutions with limited trauma experience.