Impact of ECMO on neonatal mortality in Michigan (1980-1999)

Although extracorporeal membrane oxygenation (ECMO) is a potentially lifesaving intervention, the effect of ECMO on neonatal mortality has never been evaluated. In this study, we examined the relationship between increased ECMO utilization and its effect on the neonatal mortality rate in the state of Michigan.

Neonatal mortality data were obtained from the Michigan State Department of Community Health. Data included total annual live births, total neonatal deaths, and deaths from respiratory causes in neonates > or =35 weeks' gestational age (ie, potential ECMO candidates). Pooled ECMO patient data from Michigan's 3 ECMO centers were obtained from the Extracorporeal Life Support Organization (ELSO) Registry. Associations between ECMO volume and neonatal mortality rates were assessed using simple linear regression. A scatterplot of ECMO volume and mortality rates was created, and the resulting fitted regression lines were superimposed on the plots. Statistical significance of the associations (ie, difference in slope of the regression line from zero) were based on a standard 2-sided Wald test for the regression slope parameter.

From 1980 through 1999, 1,061 neonates were treated with ECMO in Michigan, and 875 (82.5%) survived. When annual neonatal mortality rate (from all causes) and the rate from respiratory causes versus ECMO volume (cases per year) are superimposed with fitted regression lines, both are significantly different from zero (P =.041 and P =.002, respectively). The model predicts that for every 100 neonates treated with ECMO in Michigan, 38 lives are saved. The model also can be used to predict the annual neonatal mortality rate in Michigan if ECMO had not been utilized.

A strong association exists between ECMO volume and observed reductions in neonatal mortality seen in Michigan over the last 2 decades. This is the first study to show an association between ECMO and neonatal mortality rate at the population level.