Last week we reported on the failure of CardioMEMS to gain FDA expert panel approval for its implantable haemodynamic monitoring system due to questions relating to the CHAMPION study. At the time there was a sense of disappointment expressed by at least some of the 10 person panel who clearly believed in the technology, but were dismayed by the company’s apparent involvement in the study, which may have influenced the outcome. Our original article can be found here, and a subsequent article from Medical Device Daily casts further light on the scene with some interesting comments from the various parties.
The hypothesis of the CHAMPION trial was that heart failure management using monitoring of pulmonary artery pressures reduces the rate of heart failure hospitalizations. The unique nature of an implanted CardioMEMS device allows intermittent and remote assessment of pulmonary artery systolic, diastolic and mean pulmonary artery pressures.
The trial was designed to demonstrate superiority at six months as determined by the number of hospitalizations. In it 550 patients were randomised at a roughly 1:1 ratio and met this endpoint as well as a dual endpoint of no sensor failure and no device-related complications at six months. However the difference in rates of hospitalization of women between the two arms was statistically indiscernible. The rates of deaths (5.6% study, 7. 1% controls) and serious adverse events (44.8% study, 55.4% controls) both favored the device.
FDA Panel concerned that bias was introduced
According to an addendum to the FDA Panel Meeting (held last week); “inspections conducted by the Division of Bioresearch Monitoring (BIMO) revealed an issue that FDA believes is important to bring to the Panel’s attention. The issue pertains to the sponsor’s and the national Principal Investigators’ involvement in the management of individual enrolled subjects. This issue may have bearing on the interpretation of the results of the study.”
In the study protocol the sponsor (CardioMEMS) was permitted to review pulmonary artery pressure readings and alert investigators if pressure was elevated, although the investigators were to be responsible for management of the enrolled subjects, making medication changes according to recommendations provided in the protocol. It appears that nurses employed by CardioMEMS had made therapy recommendations on 193 occasions on treatment group patients (note none were permitted nor occurred on control group patients), thereby effectively rendering the trial one of treatment plus therapy against medial therapy alone, which “was not FDA’s understanding of the intent of this trial.”
FDA’s executive summary contains numerous specific examples of occasions on which recommendations were made by CardioMEMS nurses in which medications were varied, on occasions outside the parameters laid down in the protocol. In its concluding statement FDA notes that; “While FDA acknowledges that the sponsor bears responsibility for ensuring compliance with the protocol, the level of interaction between the sponsor and the clinical investigators regarding individual subjects’ treatment plans was inconsistent with FDA’s expectations based on the protocol.”
Furthermore; “FDA is concerned that the heart failure recommendations by the sponsor and National PIs for individual study subjects in the Treatment (investigational) arm only may bias the study results because these efforts may minimize hospitalization for Treatment group subjects without a comparable effort for Control group subjects.”
On the basis of the potential for bias in the results the panel voted against recommending the device for approval, but not without a number of dissenting voices at the public meeting. The sense of disappointment was best expressed by a nurse who stated; “in heart disease management, I rely on communication” with patients, who may live hundreds of miles away from where she practices. She said the CardioMEMS device “provides the objective data that is often missing in our current labor-intensive attempts” to track a patient’s condition. Thanks to the remote telemetry provided by the device, she said, treatment decisions can be made remotely, “allowing a patient who lives hundreds of miles away to receive quality treatment” without undergoing the stress of making a long trip for consultation.
One panelist, Valluvan Jeevanandum, MD, emphasised the problem of the additional treatment bias saying; “Clearly the recommendations from the CardioMEMS site to the investigational sites helped manage these patients,” adding that he would have no problem with the influence of this practice on the trial should the firm opt to offer such help in a real-world setting. “If they’re only going to go with the monitoring [provided electronically], then that becomes an issue.”
Clearly it’s not the end of the road for the device as CardioMEMS is in discussion with FDA about where to go from here. The sentiment from at least some of the panel members was that the device is “incredibly compelling” and “very safe”, which suggests that what looks like an effective real world monitoring solution may well get there eventually even if FDA takes on board its panel recommendations this time around.
Source; FDA, Medical Device Daily, medlatest staff