Characterization of Extracorporeal Membrane Oxygenation for Pediatric Cardiac Arrest in the United States: Analysis of the Kids’ Inpatient Database

Adam W. Lowry, David L. S. Morales, Daniel E. Graves, Jarrod D. Knudson, Pirouz Shamszad, Antonio R. Mott, Antonio G. Cabrera, Joseph W. Rossano

To characterize the overall use, cost, and outcomes of extracorporeal membrane oxygenation (ECMO) as an adjunct to cardiopulmonary resuscitation (CPR) among hospitalized infants and children in the United States, retrospective analysis of the 2000, 2003, and 2006 Kids' Inpatient Database (KID) was performed. All CPR episodes were identified; E-CPR was defined as ECMO used on the same day as CPR. Channeling bias was decreased by developing propensity scores representing the likelihood of requiring E-CPR. Univariable, multivariable, and propensity-matched analyses were performed to characterize the influence of E-CPR on survival. There were 8.6 million pediatric hospitalizations and 9,000 CPR events identified in the database. ECMO was used in 82 (0.9 %) of the CPR events. Median hospital charges for E-CPR survivors were $310,824 [interquartile range (IQR) 263,344-477,239] compared with $147,817 (IQR 62,943-317,553) for propensity-matched conventional CPR (C-CPR) survivors. Median LOS for E-CPR survivors (31 days) was considerably greater than that of propensity-matched C-CPR survivors (18 days). Unadjusted E-CPR mortality was higher relative to C-CPR (65.9 vs. 50.9 %; OR 1.9, 95 % confidence interval 1.2-2.9). Neither multivariable analysis nor propensity-matched analysis identified a significant difference in survival between groups. E-CPR is infrequently used for pediatric in-hospital cardiac arrest. Median LOS and charges are considerably greater for E-CPR survivors with C-CPR survivors. In this retrospective administrative database analysis, E-CPR did not significantly influence survival. Further study is needed to improve outcomes and to identify patients most likely to benefit from this resource-intensive therapy.