Hyperoxia During Cardiopulmonary Bypass Is Associated With Mortality in Infants Undergoing Cardiac Surgery*

OBJECTIVES: Patients undergoing cardiac surgery using cardiopulmonary bypass have variable degrees of blood oxygen tension during surgery. Hyperoxia has been associated with adverse outcomes in critical illness. Data are not available regarding the association of hyperoxia and outcomes in infants undergoing cardiopulmonary bypass. We hypothesize that among infants undergoing cardiac surgery, hyperoxia during cardiopulmonary bypass is associated with greater odds of morbidity and mortality. DESIGN: Retrospective study. SETTING: Single center at an academic tertiary children's hospital. PATIENTS: All infants (< 1 yr) undergoing cardiopulmonary bypass between January 1, 2015, and December 31, 2017, excluding two patients who were initiated on extracorporeal membrane oxygenation in the operating room. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The study included 469 infants with a median age of 97 days (interquartile range, 14-179 d), weight 4.9 kg (interquartile range, 3.4-6.4 kg), and cardiopulmonary bypass time 128 minutes (interquartile range, 91-185 min). A Pao(2) of 313 mm Hg (hyperoxia) on cardiopulmonary bypass had highest sensitivity with specificity greater than 50% for association with operative mortality. Approximately, half of the population (237/469) had hyperoxia on cardiopulmonary bypass. Infants with hyperoxia were more likely to have acute kidney injury, prolonged postoperative length of stay, and mortality. They were younger, weighed less, had longer cardiopulmonary bypass times, and had higher Society of Thoracic Surgeons and the European Association for Cardio-Thoracic Surgery mortality scores. There was no difference in sex, race, preoperative creatinine, single ventricle physiology, or presence of genetic syndrome. On multivariable analysis, hyperoxia was associated with greater odds of mortality (odds ratio, 4.3; 95% CI, 1.4-13.2) but failed to identify an association with acute kidney injury or prolonged postoperative length of stay. Hyperoxia was associated with greater odds of mortality in subgroup analysis of neonatal patients. CONCLUSIONS: Hyperoxia occurred in a substantial portion of infants undergoing cardiopulmonary bypass for cardiac surgery. Hyperoxia during cardiopulmonary bypass was an independent risk factor for mortality and may be a modifiable risk factor. Furthermore, hyperoxia during cardiopulmonary bypass was associated with four-fold greater odds of mortality within 30 days of surgery. Hyperoxia failed to identify an association with development of acute kidney injury or prolonged postoperative length of stay when controlling for covariables. Validation of our data among other populations is necessary to better understand and elucidate potential mechanisms underlying the association between excess oxygen delivery during cardiopulmonary bypass and outcome.