Background:The purpose of this study was to describe our experience in the application of extracorporeal membrane oxygenation (ECMO) in patients requiring mechanical assistance after cardiac surgical procedures, and in to determine the main predictors factors of successful in new cases using a mathematic statistical method.
Methods: Data from 34 pediatric patients that requiring ECMO support after cardiac operation were retrospectively analyzed from our records between January 1999 and December 2008. The main variables recorded were:demographic data, survival,cardiac physiology, indication, length of support, renal failure, bleeding. We used the Data Mining algorithm to develop a database of hypothetical 340 patients with a similar distribution of the variables in order to estimate the survival of the new cases.
Results: 34 patients (20/14 male /female) required ECMO support. Patients ranged in weight from 3 to 25 kg (mean 3,4 Kg) and age from 1 day to 72 months (mean 5,8 months). Eleven patients (32,4%) had a single ventricle physiology, 16 (48%) developed acute renal failure and 14 (41%) required reexploration for surgical bleeding. The global survival to discharge was 32,4% (11/34); in infant operated before 2005 was 3% (1/34) as compared with 43,5% (10/23) after 2005.
The mathematical model reveals the year of the operation, the weight and the presence of renal failure as the main risks factors for mortality.
Conclusion: The mathematical model based in Data Mining algorithm is a successful tool to predict the survival and mortality risk in patients managed with ECMO in the pediatric cardiac intensive care unit.