Predictors of mortality after short-term ventricular assist device placement.

Background: Risk assessment prior to long-term ventricular assist device (VAD) placement has been shown to be crucial for successful outcomes and efficient resource utilization. Short-term VADs are often used as salvage therapy in acute heart failure when the clinical scenario precludes such thorough preoperative assessment. Our goal was to devise a risk stratification system that may be used shortly after stabilization of hemodynamics with a short-term VAD to predict the likelihood of survival.

Methods: A retrospective study was performed of all patients undergoing Abiomed (Abiomed Cardiovascular Inc., Danvers, MA) or CentriMag (Levitronix LLC, Waltham, MA) placement at our institution or transferred to our institution with the device in place. From January 2001 until August 2009, 93 patients were identified. Preoperative and early postoperative variables were analyzed for their correlation with in-hospital mortality.

Results: Multivariable logistic regression analysis identified factors that were associated with death. A three-point scoring system, utilizing a diagnosis of postcardiotomy shock or graft failure, female sex, and postoperative day 3 total bilirubin greater than 5.2 mg/dL was devised. A score of 2 or greater was associated with an 86% mortality rate, whereas a score of 0 was associated with a 13% mortality rate.

Conclusions: A simple scoring system based on readily available data may predict mortality after short-term VAD placement. Such a scoring system may be of prognostic value for physicians and patient families early in the support period and may help guide decisions.