• Timothy M Maul, Bradley A Kuch, and Peter D Wearden.
• From the *Department of Cardiothoracic Surgery, University of Pittsburgh School of Medicine and Children's Hospital of Pittsburgh of UPMC, Pittsburgh, Pennsylvania; †McGowan Institute for Regenerative Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania; and ‡Department of Pediatric Critical Care Medicine, Children's Hospital of Pittsburgh of UPMC, Pittsburgh, Pennsylvania.
• ASAIO J. 2016 Sep 1; 62 (5): 584-90.

AbstractVenoarterial extracorporeal membrane oxygenation (VA-ECMO) has saved thousands of newborns. Population comparisons for research and quality initiatives require risk-matching, but no indices exist for this population. We sought to create a pre-ECMO risk index using the registry data from the Extracorporeal Life Support Organization. We analyzed 5,455 neonatal (<30 days old) respiratory VA-ECMO patients for the period 2000-2010. Multivariate regression examining the impact of pre-ECMO variables on survival to hospital discharge was performed to create the Pittsburgh Index for Pre-ECMO Risk (PIPER), which was ultimately was based on seven pre-ECMO variables. Each PIPER quartile demonstrated increasing mortality by 15% (R = 0.98) and was associated with increased complications on ECMO. Further modeling to include on-ECMO complications (PIPER), including complications and length of time on ECMO, increased the predictive power of the model, with 21% increases in mortality per PIPER quartile (R = 0.97). Our developed indices provide the first steps towards risk-adjusting patients for meaningful comparisons amongst patient populations. There may be additional clinically relevant measures, both pre- and on-ECMO, which could provide better predictive capability. Future work will focus on finding these additional measures and expansion of our techniques to include other patient populations.

16 keywords...

hide…