SAN FRANCISCO – A proposed new definition of acute respiratory distress syndrome describes categories based on mild, moderate, or severe hypoxemia that correlate increasing severity with significantly increased mortality or increased time on mechanical ventilation among survivors.
The draft definition, created under consensus process by an international panel of experts, was refined by empirical testing in a meta-analysis of data on 4,457 patients in two large data sets from seven centers. The risk of mortality from acute respiratory distress syndrome (ARDS) was 27% with mild disease, 32% with moderate ARDS, and 45% with severe ARDS, Dr. Niall D. Ferguson and Dr. Gordon D. Rubenfeld reported at an international conference of the American Thoracic Society.
Sherry Boschert/IMNG Medical Media
Dr. Niall D. Ferguson
The median duration of mechanical ventilation in survivors was 5 days in patients with mild ARDS, 7 days with moderate ARDS, and 9 days with severe ARDS, said Dr. Ferguson, director of critical care medicine at the University of Toronto. The study was published online May 21 in JAMA (doi: 10.1001/jama.2012.5669).
The European Society of Intensive Care Medicine convened the panel of experts in Berlin in 2011 to draft a new definition of ARDS in hopes of improving upon the 1994 definition from the American-European Consensus Conference (AECC), said Dr. Rubenfeld, professor of medicine at the University of Toronto and chief of the program in trauma/emergency and critical care at Sunnybrook Health Sciences Center, Toronto. Since the AECC definition was adopted widely, issues of reliability and validity have emerged. The American Thoracic Society and the Society of Critical Care Medicine endorsed the 2011 consensus effort.
Under the new Berlin Definition, patients with mild ARDS have mild hypoxemia, defined as a ratio of arterial partial pressure of oxygen to fraction of inspired oxygen of 201-300 mm Hg (PaO2/FIO2 = 201-300 mm Hg). Moderate hypoxia (PaO2/FIO2 = 101-200 mm Hg) defines moderate ARDS, and severe hypoxia (PaO2/FIO2 = 100 mm Hg or less) defines severe ARDS.
The initial draft of the Berlin Definition included four ancillary variables for severe ARDS that were dropped after the meta-analysis found that they did not improve the predictive value for mortality, the speakers and their associates in the study reported. The abandoned variables were radiographic severity, respiratory system compliance, positive end-expiratory pressure, and corrected expired volume per minute.
Dr. Rubenfeld stressed that these variables still are important for clinicians to measure and for understanding ARDS, but they were not included in the definition of severe ARDS because they made the definition more complex while not adding anything to the predictive value of the definition.
He also cautioned that neither the Berlin Definition nor the AECC definition is designed to be a prognostic model; the end point of mortality was used to hone the Berlin Definition.
The Berlin Definition had better predictive validity for mortality than the AECC definition in an analysis using the area under the receiver operating curve (AUROC) in logistic regression models. The Berlin Definition had an AUROC of 0.577, compared with 0.536 for the AECC definition, a statistically significant improvement.
The data for the meta-analysis came from four multicenter clinical studies and three single-center physiological studies, the investigators reported.
Twenty-two percent of patients met the Berlin Definition criteria for mild ARDS, 50% had moderate ARDS, and 28% had severe ARDS. Median ventilator-free days declined with severity of disease, from 20 days with mild ARDS to 16 days with moderate ARDS and 1 day with severe ARDS.
Among patients with mild ARDS at baseline under the Berlin Definition, 29% progressed to moderate disease and 4% progressed to severe disease within 7 days. Among patients with moderate ARDS at baseline, 13% progressed to severe disease within 7 days.
Sherry Boschert/IMNG Medical Media
Dr. Gordon D. Rubenfeld
The investigators suggested that this approach of combining consensus discussions with empirical evaluation might be a model for creating more accurate, evidence-based definitions for critical illness syndromes. Previous ARDS definitions relied on expert consensus alone.
Without the empirical evaluation that led to deleting the ancillary variables, a needlessly complex ARDS definition would have been proposed, Dr. Rubenfeld said.
The study was sponsored by the European Society of Intensive Care Medicine, the National Institutes of Health, and the Canadian Institutes of Health Research (CIHR). CareFusion provided in-kind support for the study. Dr. Ferguson was supported by a CIHR New Investigator Award. Dr. Rubenfeld reported having financial relationships with Ikaria, Faron, and Cerus. Some of his associates in the study reported financial relationships with Maquet Medical, Hemodec, Faron, AstraZeneca, U.S. Biotest, Sirius Genetics, Sanofi-Aventis, Immunetrics, Abbott, Eli Lilly, Ikaria, GlaxoSmithKline, Tarix, Apeiron, and/or Novalung.