Finally, the study was unblinded, and eight patients (26%) randomized to anticoagulation therapy withdrew after randomization. Such a high percentage of differential dropouts eliminate the benefit of randomization, as those who withdrew may somehow be different from those who did not (perhaps they were more ill). Consequently, the differences between groups could be due to confounding variables and not a true treatment effect. A better approach would have been to include those patients who withdrew as part of the “treatment group” in an intention to treat analysis. Less robust but still valid, the investigators could have analyzed the data using a Cox proportional hazards model including well-characterized covariates that may influence mortality in IPF patients such as age, smoking status, and baseline FVC. The results of this trial are dramatic but must be interpreted with caution and confirmed by future studies addressing the concerns outlined above before anticoagulation is adopted as a new standard of care in IPF.