Hospital mortality rates were 55% lower in 2009 influenza A(H1N1) patients with severe acute respiratory distress syndrome who received extracorporeal membrane oxygenation compared with non-ECMO controls, based on data from a cohort study of 80 patients.
Data from a recent randomized trial showed that patients with acute respiratory distress syndrome (ARDS) who were transferred to an ECMO center were more likely to survive without severe disability compared with non-ECMO patients, but the role of the treatment remains controversial because of the increased costs associated with its use, said Dr. Moronke Noah of Glenfield Hospital in Leicester, England, and colleagues.
In this study, the researchers used data from the Swine Flu Triage study (SwiFT) to compare mortality rates in patients with ARDS resulting from the 2009 H1N1 flu who received ECMO with non-ECMO controls. SwiFT was a prospective study of patients with suspected or confirmed 2009 H1N1 flu who were referred for critical care.
The findings were simultaneously presented at the European Society of Intensive Care Medicine and published online Oct. 5 in JAMA (Epub doi:10.1001/jama.2011.1471).
The researchers reviewed data from 80 patients who were referred for ECMO. They used three different matching techniques: individual matching, propensity scoring, and GenMatch matching. A total of 59 patients were matched with non-ECMO controls using individual matching, 75 were matched with non-ECMO controls using propensity scoring, and 75 were matched with non-ECMO controls using GenMatch matching. GenMatch "combines propensity score matching with multivariate matching," the authors explained.
The hospital mortality rates were significantly lower among ECMO patients than among non-ECMO patients in each of the three matching techniques. Hospital mortality rates were 24% in ECMO patients and 53% in non-ECMO patients when individual matching was used; 24% and 47%, respectively, when propensity scoring was used; and 24% and 51%, respectively, when GenMatch was used.
The findings were limited by the possible role of unobserved confounding variables, and by the lack of data on the exact treatment protocols of non-ECMO patients, the researchers noted.
However, "the unique value of this study lies in the homogeneity of the patients with H1N1-related ARDS and the matching methods used," they said. The consistency of the results across all three matching methods strengthens the role of ECMO in reducing hospital mortality in these patients, they said.
The SwiFT study was supported by the U.K. National Institute for Health Research. Dr. Noah had no financial conflicts to disclose, but many of the study coauthors have received reimbursement or grant support from multiple pharmaceutical companies and institutions.