CardioMEMS Postapproval Study Linked to Reduction in Heart Failure Hospitalizations at 1 Year

NEW ORLEANS, LA—The performance of the CardioMEMS HF System (St. Jude Medical/Abbott) in the after-market setting is associated with decreased heart-failure hospitalizations across all patient sex, race, and ejection fraction categories with few device-related complications at 1 year, according to a postapproval study.

The implantable system, which uses unique technology to gauge pulmonary artery pressure but relies on patients to transmit their data daily, is US Food and Drug Administration-approved for patients with NYHA class III heart failure symptoms and who were hospitalized for decompensated heart failure within the last year. The CHAMPION trial showed a 37% reduction in hospitalization for heart failure associated with CardioMEMS, but MAUDE database findings have indicated that the system might be associated with an increased rate of pulmonary artery injury.

David Shavelle, MD (University of Southern California, Los Angeles), who presented the findings here this week at the American College of Cardiology (ACC) 2019 Scientific Session, stressed that the device is best used in a select cohort of class III heart failure patients who are going to be compliant in submitting their data, already receiving appropriate guideline-directed medical therapy, perhaps still having issues with volume overload, and are very comfortable communicating with their healthcare providers.

“When I call a patient, 'It's Dr. Shavelle calling. I want you to increase your diuretic,' if they say, 'Who's Dr. Shavelle?' that's not the right patient. You have to have a relationship with the patient,” he said in a media briefing, adding that obese patients and those who live far away from the hospital might also benefit from this technology.

Postapproval Study Results

For the study, Shavelle and colleagues enrolled 1,214 patients who met the CHAMPION inclusion criteria to be implanted with the CardioMEMS device between September 2014 and March 2018.

Compared with 1 year before implant, patients who received the system were 58% less likely to be hospitalized for heart failure (primary endpoint) in the subsequent year (HR 0.42; 95% CI 0.38-0.47). Additionally, a 44% reduction of the combined endpoint of heart-failure hospitalization and death (HR 0.56; 95% CI 0.51-0.62) and a 28% reduction in all-cause hospitalizations (HR 0.72; 95% CI 0.67-0.77) were observed.

The benefit in the primary endpoint was seen across all EF cohorts with 54%, 63%, and 61% reductions in hospitalization for heart failure observed in the EF < 40%, 41-50%, and > 50% cohorts, respectively. Also, the results were maintained across prespecified subgroups of gender, ischemic or nonischemic cardiomyopathy status, with versus without an implantable cardioverter defibrillator, and race.

Freedom from device-related complications at 1 year was 99.7%, and sensor failure was only observed in a single patient.

Select Patients, Safety

In a panel discussion following the presentation of the results, Anita Deswal, MD, MBBS (Baylor College of Medicine, Houston, TX), said, “The totality of data of the CaridoMEMS device, including this trial, tells us that the device is very effective, [has a] very high success rate of implantation, low complication rate, and with good adherence to transmission and effective action, it can reduce hospitalizations.”

Thinking ahead, she said she agreed with Shavelle’s assessment that only select patients who are “not too sick and not too healthy” should be considered for implantation of the device as “NYHA class III can be very broad.” Additionally, Deswal asked about the hospital set-up required for effective ongoing monitoring and treatment changes in routine clinical practice outside of a clinical trial setting.

Shavelle replied that things might be different for his practice “because I'm at a big academic medical center where we have a lot of infrastructure. I think it can be done in academic medical centers, but I do believe it can be done in commercial settings, private practices,” he said. “I like to say you can either go to emergency room and admit your patients, or you can follow the patient over time and keep them out of the emergency room. So it's just a cost balance there.”

Deswal also asked whether there was any difference in the benefit seen for patients who had just one versus several hospitalizations. “I don't have the data in front of me, but there was a benefit of all three of those groups,” Shavelle responded. “If you were a so-called frequent flier, you still had a benefit. The ones that had more frequent admissions were the ones that tended to die, and you could argue that you probably shouldn’t be putting this device in if you're class IIIb moving into class IV because we won’t derive benefit.”

Referencing prior public concern over safety of this device, Gurusher Panjrath, MD (George Washington University Hospital, Washington, DC), said in the media briefing that it is important to note that the safety signal in this study was “really good.” Outside of a clinical trial and in the commercial setting, there’s a concern that the rigor of implantation might change, he said, but “this data shows that the safety is still there.”

Real-World Pragmatism

Based on his experience working with the FDA, in which postapproval studies were “always the challenge when we dealt with devices,” panelist Clyde Yancy, MD (Feinberg Northwestern University School of Medicine, Chicago, IL), said he was “very impressed that a successfully executed postapproval study was accomplished.”

But he questioned the pragmatism of the CardioMEMS device in the real world. “I noted 76% accomplished daily transmissions out a very large denominator of patients—90% at least weekly. That’s a fairly intense amount of oversight and data aggregation,” Yancy said. “Do you have any sense that the weekly transmission was sufficient? Was a daily transmission necessary? Because it has direct implications on infrastructure and resource burden.”

Shavelle said that while he is not exactly sure of the optimal transmission frequency, “I can tell you in clinical practice we are following about 120 patients, and we like [to tell] patients to transmit daily knowing that they are going to probably transmit several times a week. I personally think weekly is not enough. For a stable patient, weekly is enough. For an unstable patient, weekly is not enough. So, if we can get our patients to go every 2 to 3 days, we think we're helping them.”

On a related note, “one of the areas of tension with your first data set had to do with how many times there was a communication between a nurse coaching a patient about best care,” Yancy said. “So we realize that the device per se has no therapeutic benefit. It is actually the actions that are prompted by the device—action number one being a telephone conversation. Do you have records in this postapproval study of how many times that kind of exchange occurred with the patients, where a nurse or another kind of provider picked up the phone and communicated with the patient?”

Unfortunately not, Shavelle answered. “This was a postapproval study with certain selected criteria and we don't have that information,” he said. “I would just suggest that the significant decline in [pulmonary artery] pressures over time would indicate that something was working, but I don't have individual variability between the sites on how often patients transmitted, what the healthcare provider at bedside did, [and] if they acted on those action items.”