PURPOSE: Case reports and series have implicated amiodarone as a risk factor for acute lung injury (ALI) and acute respiratory distress syndrome (ARDS). This study aims to estimate more precisely whether pre-hospital use of amiodarone increases the risk of developing ALI among patients admitted to a hospital.
METHODS: The US Critical Illness and Injury Trials Group recently studied 5584 adult patients admitted to 22 centers with at least one risk factor for developing ALI. In a secondary analysis of this cohort, the pre-hospital use of amiodarone was determined at admission, and the patients subsequently followed for the primary outcome of ALI. A propensity score for being on amiodarone was determined in a standard logistic regression model, and used to match cases of ALI to controls without ALI (1:4). Using this propensity-matched, case-control design to best adjust for the inherent bias introduced in the observational cohort, the adjusted risk of developing ALI from being on amiodarone was then computed from a conditional logistic regression model.
RESULTS: Forty-nine of 5584 patients were on amiodarone at the time of hospitalization. (Patient characteristics have been previously reported.) 12.2% of patients on amiodarone developed ALI (versus 6.7% among those not on amiodarone), resulting in a trend towards an increased risk for ALI when taking amiodarone (odds ratio=1.94, p=0.14). In the propensity-matched, case-control model, all 377 ALI cases were successfully matched by their propensity scores. The final adjusted risk of developing ALI when on amiodarone diminished, with an odds ratio of 1.05 (p=0.92, 95% confidence interval: 0.40-2.74).
CONCLUSIONS: Pre-hospital amiodarone therapy does not significantly increase the risk of developing ALI among patients requiring acute hospitalization.
CLINICAL IMPLICATIONS: The concern for ALI should not preclude the appropriate use of amiodarone in the outpatient setting, but the effect of acute inpatient use of amiodarone on the development of ALI/ARDS requires further study.
DISCLOSURE: The following authors have nothing to disclose: Lioudmila Karnatovskaia, Emir Festic, Augustine Lee
No Product/Research Disclosure Information