December 6th, 2011
FDA Reviewers Raise Questions About CardioMEMS CHAMPION Trial Ahead of Advisory Panel
FDA reviewers have raised serious questions about the results of the CHAMPION trial of the implantable CardioMEMS HF Pressure Measurement System (HF System), which provides daily pulmonary artery pressure measurements for the purpose of guiding heart failure treatment. On December 8, the FDA’s Circulatory System Devices Panel is scheduled to review the PMA for the device. (Click here for the FDA documents.)
The main FDA review concludes that the trial “met the primary effectiveness endpoint and both primary safety endpoints in the population studied, based upon their pre-specified hypotheses.” But, the review continues: “Although statistically significant, FDA has questions about the clinical significance of the findings, in particular related to the primary effectiveness and several of the secondary effectiveness endpoints (Proportion of subjects hospitalized for heart failure, Days Alive Outside of the Hospital and Quality of Life).” The FDA said it “would like panel input on the overall risk/benefit profile and approvability of the device.”
An addendum to the main FDA review raises additional new concerns about the trial. An earlier report about these concerns was refuted by the company’s CEO and founder, cardiologist Jay Yadav. However, with the release of the FDA documents, it now appears that the earlier rumors were at least partially correct, though there is no evidence to show that CardioMEMS had knowledge of the new FDA concerns at the time.
The addendum to the FDA review discusses information “not present in the PMA submission… based on inspections conducted by” FDA monitors. The FDA reviewers write that the trial may have been biased because the company and the national principal investigators provided additional medical support to patients in the treatment group but not to control patients. It should be noted that the purpose of the single-blind design of the trial was to test whether information received from the device could improve HF treatment. The crux of the difference between the FDA and CardioMEMS therefore appears to be whether the company’s efforts to enforce protocol-mandated treatments fall outside the standard practice of a clinical trial.
Here is the text of the additional question the FDA will ask the panel on Thursday:
FDA is concerned that the conduct of the clinical trial may have biased the study results. Inspections coordinated by FDA’s Division of Bioresearch Monitoring (BIMO) identified evidence that the sponsor, who had knowledge of the randomization assignment, or the principal investigators routinely contacted investigational sites and made specific therapeutic recommendations for some Treatment Group study subjects, including
- titration of medication doses,
- addition or discontinuation of medications,
- recommendations for outpatient intravenous medication administration,
- the addition of medications that were not included in the protocol, and
- sleep study evaluations not included in the protocol.
FDA is concerned that the study results may be biased by these recommendations because investigators did not receive similar treatment recommendations from the sponsor for Control Group study subjects. FDA’s interpretation of the protocol is that therapeutic changes were to be made by the study site investigators.
The information obtained through inspections coordinated by FDA’s Division of Bioresearch Monitoring raises concerns for FDA that specific therapeutic recommendations may have minimized heart failure hospitalizations for Treatment Group study subjects without an equivalent impact for Control Group study subjects. FDA is concerned that the study results may be biased and that the ability to interpret study results may be compromised by the trial conduct.
a. FDA is concerned that the specific treatment recommendations made by the sponsor may have introduced bias into the study results. Please discuss whether or not you agree with this concern.
b. If the specific treatment recommendations introduced bias into the study results, please discuss how this impacts the ability to interpret the study results.
c. FDA is concerned that the measures taken by the sponsor would not be duplicated in a post-market setting, if the device were to be approved. Please discuss how the difference between how the study was conducted and how the device would likely be used in a post-market setting should be taken into account when interpreting the study results.
d. The information identified in the BIMO inspections was not submitted by the sponsor as part of the PMA. Please discuss whether this information is relevant to an evaluation of the safety and effectiveness of the device.
In its response to the FDA concerns, CardioMEMS said that “the FDA has not provided to date any evidence from the study to support a finding that either inappropriate contact was made with study site personnel or that legitimate contact with study site personnel, as outlined in the protocol, resulted in bias being introduced into the study results obtained.”
CardioMEMS further explained:
As prespecified within the clinical protocol, when the PA pressures were outside of values that were setup in advance for individual patients and modifiable by investigators, each investigator was required to consider medication changes per the hemodynamic monitoring treatment guidelines in the protocol. Follow-up email contact after pressure alerts, with investigators by CardioMEMS clinical staff, enforced compliance with protocol requirements and ensured that investigators considered all treatment options to lower PA pressures, including the treatment guidelines prespecified in the clinical protocol. The FDA is concerned that similar emails were not sent regarding Control group subjects but as is evident from the study design, emails directed towards reducing PA pressure could not have been sent regarding the Control group subjects. These PA pressure reduction recommendations could have been made for Control patients only if the pressures for Control patients were monitored during the trial.
In addition, the company performed an analysis showing that the emails from the company had at most a negligible impact on treatment.
Two well-known cardiologists who have consulted for CardioMEMS and are familiar with the documents discussed the documents and the CardioMEMS device. Michael Domanski said that he believes “the device has passed the legal test the FDA has to apply: I think it is safe and effective for its intended purpose.” Domanski was not bothered by the additional concerns raised by the FDA, given the single-blind design of the study, since the company advice “falls under the general heading” of getting advice from a colleague. He felt that the FDA would likely mandate post-approval studies which would help settle the question of the role of the device in a real-world setting.
Milton Packer had a somewhat more nuanced perspective about the FDA’s additional concerns. “By definition, the fact that the FDA has raised it makes it an issue,” he said. The company can respond by stating that it will market both the device and a case management system to go with it, he speculated. “In other words,” he said, “what it is commercializing is exactly how the trial was conducted.” Another option is to demonstrate, as the company already has attempted to do, that the communications that occurred didn’t have any impact on the outcome of the trial. The two options are not mutually exclusive, he noted.