Extracorporeal Membrane Oxygenation After Congenital Heart Surgery: Does One Database Fit All?*
The Extracorporeal Life Support Organization (ELSO) registry remains the largest international database incorporating roughly 310 centers. In the most recent ELSO report (2016), the survival rate for neonates after cardiac ECLS was shown to be roughly 62%, compared with respiratory ECMO survival of 84% (6). Although this database contains detailed standardized information regarding perisupport period one of its major limitations is lack of data on preoperative risk factors, operative characteristics, individual congenital heart lesions, and interhospital variation. Most importantly, it also fails to capture patients who were considered for ECMO support but ultimately did not receive this therapy, limiting our ability to study the true benefit of ECMO.
The Society of Thoracic Surgeons Congenital Heart Surgery Database (STS-CHSD) collects perioperative data on all children undergoing cardiac surgical repair at more than 100 North American centers. The perioperative data include information regarding the primary procedures of index cardiovascular operation of the admission which is analyzed individually and categorized using the International Pediatric and Congenital Cardiac Code risk stratification system (7). Mascio et al (8) used this database to describe the patterns of mechanical circulatory support (MCS) use and outcomes across STS-CHSD participating centers. They revealed that MCS (> 95% ECMO) was used in 2.4% cardiac operations at 80 centers with the greatest number used in those with single ventricle palliative procedures and complex biventricular repairs. More than half of those treated with MCS (53%) did not survive to hospital discharge with mortality reaching greater than 70% for the more complex operations and a significant variation in MCS rates between high- and low-volume hospitals. Although STS-CHSD captures detailed data on perioperative variables with excellent completion and agreement rates, it fails to incorporate the details regarding indication for MCS, omitting detailed information regarding timing of the initiation, type and duration of perioperative MCS as well as related complications.
In this issue of Pediatric Critical Care Medicine, Bratton et al (9) embark on yet another quest to shed more light on ECMO support in the setting of CHS. In their retrospective cross-sectional study, they choose the Pediatric Health Information System (PHIS) database. The PHIS database includes administrative and billing data from more than 40 U.S. children’s hospitals, providing generalized reflection of surgical experience in the country. Diagnoses and procedures for all children hospitalized at these institutions are coded by billing personnel using International Classification of Diseases, 9th Edition (ICD-9) codes. Using the combination of ICD-9 diagnosis and procedure codes, one is able to identify and group surgical mortality risk using the Risk Adjustment in CHS (RACHS)-1 system (10).