Improved Outcomes Associated With Intraoperative Steroid Use in High-Risk Pediatric Cardiac Surgery

Abstract

Corticosteroids are commonly administered perioperatively in pediatric cardiac surgery to reduce cardiopulmonary bypass induced inflammation. However, their effects on outcomes and potential for adverse events are not well defined. A review was undertaken of cardiac operations between September 2004 and December 2007 carrying a comprehensive Aristotle score 10 or greater. A nonrandomized comparison was undertaken comparing those patients having received intraoperative methylprednisolone at anesthesia induction or in the bypass circuit prime with those who did not. To account for nonrandom assignment of steroid use, a propensity model was created to establish each patient's probability of having received steroids (∼150 variables evaluated, 17 in final model, c-stat 0.94, p < 0.001). Associations between postoperative outcomes and intraoperative steroid use were modeled in multivariable linear regression models adjusted for propensity score and relevant surgical characteristics. In 221 identified cases, 134 (61%) patients received intraoperative steroids; of these, 44 (33%) also received preoperative doses. In propensity-adjusted regression models, intraoperative steroid use was associated with lower chest tube volume loss in the first 24 postoperative hours (-5.3 mL/kg, p < 0.001), and shorter durations of stay in intensive care (-2.3 days, p < 0.001) and hospital (-4.1 days, p < 0.001). Use of an additional preoperative dose resulted in further improvements, especially a reduction in duration of mechanical ventilation (-1.7 days versus no steroids, -1.2 days versus intraoperative steroids only, p = 0.002). Steroids were not associated with increased postoperative lactate, creatinine, or glucose levels, or odds of infection. Intraoperative steroid use is associated with improved postoperative outcomes for children undergoing high-risk cardiac surgery, with further benefits associated with a preoperative dose.