August 31, 2020
Thank you for the opportunity to speak today on behalf of the National Center for Health Research. I am Dr. Meg Seymour, a senior fellow at the center. Our center analyzes scientific and medical data to provide objective health information to patients, health professionals, and policy makers. We do not accept funding from drug or medical device companies, so I have no conflicts of interest.
Today, the committee is asked to discuss whether data from the IMPACT trial provide substantial evidence supporting the claim that TRELEGY improves all-cause mortality (ACM) for patients with COPD. The design of the IMPACT trial limits its ability to address this claim. We share the concerns of the FDA scientists that trends shown by the data do not support the claim that TRELEGY improves all-cause mortality.
In the IMPACT trial, the difference in ACM between ICS and non-ICS trial arms was limited to the first 90 days of the trial, and disappeared after that. This clearly suggests that the difference could have been due to how some patients assigned to the non-ICS arm were being treated with ICS prior to the study.
As FDA scientists pointed out, 71% of patients were treated with ICS prior to the study. However, if they were assigned to the non-ICS arm, they had their ICS removed as part of the study design. Of the 29% of patients who were ICS-naive, there was no apparent difference in ACM between those in the ICS and non-ICS arms. Although this analysis was underpowered, the data suggest there was no meaningful difference. Thus, the difference in ACM between the ICS and non-ICS arms appears to be due to this removal of ICS.
It is essential to note that when FDA scientists analyzed the data as if ICS treatment was a control and ICS removal was an active condition, they found a potentially clinically significant 5-fold increased risk of mortality attributable to ICS removal from baseline to 90 days into the trial. This analysis clearly suggests that TRELEGY does not actually reduce ACM. Instead, removing ICS for some patients appears to have led to the difference in ACM between the ICS and non-ICS arms. In other words, the ACM result of the IMPACT trial is due to the design of the trial, rather than the drug itself reducing ACM.
In addition to this apparent effect of the study design, we are also concerned about the limited diversity of the trial sample, which was mostly older white males. It is unclear whether women, people of color, or other age groups would or would not benefit.
The design of the IMPACT trial is not consistent with how patients would be treated in the real world clinical setting. In a typical clinical setting, patients would generally be prescribed TRELEGY because they are moving from double to triple therapy. A TRELEGY prescription would mean adding treatment with ICS to those who are ICS naive. Since the IMPACT study design terminated ICS use in the control arm, the data from IMPACT cannot address whether adding ICS to those who are ICS naive reduces all-cause mortality. We therefore urge the committee to keep this in mind when discussing the proposed labeling claim, and to reject it as not adequately supported by the data.
The Pulmonary-Allergy Drugs Advisory Committee voted 14 to 1 against approval of the proposed labeling.