Predicting shunt failure in children: should the global shunt revision rate be a quality measure?

Abstract

OBJECT Ventricular shunts for pediatric hydrocephalus continue to be plagued with high failure rates. Reported risk factors for shunt failure are inconsistent and controversial. The raw or global shunt revision rate has been the foundation of several proposed quality metrics. The authors undertook this study to determine risk factors for shunt revision within their own patient population. METHODS In this single-center retrospective cohort study, a database was created of all ventricular shunt operations performed at the authors’ institution from January 1, 2010, through December 2013. For each index shunt surgery, demographic, clinical, and procedural variables were assembled. An “index surgery” was defined as implantation of a new shunt or the revision or augmentation of an existing shunt system. Bivariate analyses were first performed to evaluate individual effects of each independent variable on shunt failure at 90 days and at 180 days. A final multivariate model was chosen for each outcome by using a backward model selection approach. RESULTS There were 466 patients in the study accounting for 739 unique (“index”) operations, for an average of 1.59 procedures per patient. The median age for the cohort at the time of the first shunt surgery was 5 years (range 0–35.7 years), with 53.9% males. The 90- and 180-day shunt failure rates were 24.1% and 29.9%, respectively. The authors found no variable—demographic, clinical, or procedural—that predicted shunt failure within 90 or 180 days. CONCLUSIONS In this study, none of the risk factors that were examined were statistically significant in determining shunt failure within 90 or 180 days. Given the negative findings and the fact that all other risk factors for shunt failure that have been proposed in the literature thus far are beyond the control of the surgeon (i.e., nonmodifiable), the use of an institution’s or individual’s global shunt revision rate remains questionable and needs further evaluation before being accepted as a quality metric.