[Skip to Navigation]
Views 237
Citations 0
Comment & Response
September 21, 2021

Antimicrobial Therapy and Respiratory Hospitalization or Death in Adults With Idiopathic Pulmonary Fibrosis—Reply

Author Affiliations
  • 1New York Presbyterian Hospital/Weill Cornell Medicine, New York, New York
  • 2Duke Clinical Research Institute, Duke University, Durham, North Carolina
  • 3Division of Pulmonary and Critical Care Medicine, University of Virginia, Charlottesville
JAMA. 2021;326(11):1071-1072. doi:10.1001/jama.2021.11813

In Reply Dr Chen and colleagues raise important questions regarding the interpretation of the data published by the CleanUP-IPF investigative group.1 Two specific points require additional discussion.

The first focuses on the overall interpretation of our study, which found—using an ITT analysis—that an antimicrobial therapy strategy does not yield beneficial results in a composite clinical end point in patients with IPF. Chen and colleagues suggest that the preliminary data underlying our study noted a reduction in mortality in patients with fibrotic idiopathic interstitial pneumonia in a per-protocol analysis compared with the ITT analysis.2 This example highlights the limitations of per-protocol analyses, particularly those that are not carefully predefined, in clinical trials; this is especially relevant to pragmatic trials.3 Our investigative group chose a prospective, randomized, effectiveness clinical trial design following pragmatic principles, including broad inclusion criteria, limited exclusion criteria, a wide variety of participating sites, the absence of a placebo control, 2 antimicrobial agents (co-trimoxazole or doxycycline) with preliminary supportive data to provide clinical flexibility, and a relevant end point to enhance the generalizability of the study findings.4 Our study1 confirms the feasibility of such an approach in a rare lung disorder, albeit with a limitation in documenting and securing maximal adherence to the study drug. Importantly, the trial was stopped for futility because estimates for most of the study end points were pointing in an unfavorable direction. Our conclusion regarding the lack of utility of the co-trimoxazole treatment component of our therapeutic strategy, following the same dosing regimen as Shulgina et al,2 is strengthened by the parallel, explanatory, placebo-controlled trial published by Wilson et al.5 Despite 2 very different study designs, the treatment effect and study drug tolerability were remarkably similar in these trials. These results highlight the difficulty of drawing firm conclusions from per-protocol analyses of clinical trials while supporting the conclusion that co-trimoxazole therapy, delivered to alter disease progression in patients with IPF, is unlikely to have a beneficial effect.