How Meaningful Are Heart Failure Hospitalizations in Today’s RCTs?

Heart failure (HF) hospitalization is an endpoint worthy of inclusion in clinical trials but not the definitive solution for challenges in study design, according to a new commentary.

The paper, published online Wednesday in JACC: Heart Failure, explores the many options available going forward.

“There’s always been a discussion about: what are the optimal endpoints for heart failure clinical trials? And heart failure hospitalization has always been a cornerstone in that,” Abhinav Sharma, MD, PhD (McGill University Health Centre, Montreal, Canada), one of the commentary’s authors, told TCTMD. While HF hospitalization continues to be important, “there is increasing recognition that there could be other endpoints that are also very patient-centric and can capture different elements of the patient experience with heart failure.”

One reason to consider creative approaches is that differences in mortality, the most definitive outcome of all, are becoming harder to detect in heart failure trials as treatment improves, said Sharma. “We now have more therapies that have shown evidence of reducing morbidity and mortality amongst patients with heart failure and also those with risk factors for heart failure, such as type 2 diabetes with atherosclerotic cardiovascular disease.”

This is not to say that HF hospitalization isn’t of value, he specified. “We know that if a patient with heart failure is hospitalized, then their future risk of mortality is inherently much higher than someone who does not have hospitalization.” But, he added, it also takes “a huge amount of effort and time to identify heart failure hospitalizations and adjudicate these events.”

Moreover, phase III heart failure trials are “very expensive to conduct,” he said. “So are there are other types of endpoints that we could start to look towards that could provide information that . . . has meaningful impact on patients and the healthcare environment, and actually indicates patients are improving their underlying biology and severity of heart failure?”

There is increasing recognition that there could be other endpoints that are also very patient-centric and can capture different elements of the patient experience with heart failure. Abhinav Sharma

Sharma wrote the paper along with Pedro Marques, MD (University of Porto and Centro Hospitalar Universitário de São João, Porto, Portugal; McGill University Health Centre), and João Pedro Ferreira, MD, PhD (University of Porto; Centro Hospitalar de Vila Nova de Gaia/Espinho, Portugal; Centre Hospitalier Régional Universitaire, Nancy, France).

Their commentary lays out evidence for and against HF hospitalization, arriving at a list of pros and cons.

HF hospitalization, to its credit, falls within the causal pathway for mortality and carries weight as it presents a financial burden for healthcare systems, the co-authors note. It’s also a patient-centered outcome and occurs often enough to be adequately analyzed in randomized controlled trials.

But there are downsides as well. These include conflicting data on the link between HF hospitalization and mortality seen in observational versus randomized studies, inconsistencies due to geographical variations in HF management, the endpoint’s limited ability to fully capture HF-associated morbidity, and the aforementioned need for adjudication.

More Creative Trial Design

Sharma, speaking with TCTMD, outlined the options going forward.

To start, researchers can think of inventive methods to leverage the information trials are already collecting. “Using the endpoints we have, can we think of different ways of analyzing the data? This includes using different definitions, such as hierarchical endpoints and win ratios, which allow you to encompass many elements such as heart failure hospitalizations and other types of very patient-centered outcomes” like symptoms and time spent out of the hospital, he suggested. “That allows you to potentially reduce the sample size compared to a more traditional time-to-event analysis.”

A second approach is to use other “endpoints that also inform the causal pathway of heart failure progression,” Sharma continued. “For example, intensification of diuretic therapies, or use of intravenous therapies, as an outpatient until more recently was not used as a common endpoint in clinical trials. But as a practicing clinician and cardiologist I know that . . . these are patients that don’t do as well and they have higher risk of bouncing into the hospital with heart failure and potentially passing away.”

Novel tools, too, might track with worsening prognosis.

“Given the increasing prevalence of HF and associated comorbidities, the need for continuous and close follow-up of these patients will overwhelm an already stressed healthcare system,” the authors observe. “In this regard, the use of digital health technologies can potentially allow for continuous patient monitoring, identify novel outcomes for use in clinical trials, anticipate clinical decompensation, and reduce the rate of hospitalizations.”

Actigraphy, for example, could detect early signs that patients aren’t moving as much as before, Sharma said. As reported by TCTMD, researchers are also testing things like smartphone apps that use voice recognition to identify HF patients at risk of impending decompensation that might lead to hospitalization. Other technologies monitoring hemodynamic changes and pulmonary artery pressure are also helping physicians to better anticipate HF hospitalizations and adverse events.

Sharma predicted clinical trialists are on the cusp of adopting additional endpoints, like outpatient intensification of diuretics, when designing heart failure studies. “We’ve already seen this with some of the SGLT2 inhibitor trials that were published over the past few years, where they’ve now incorporated these types of [patient-centered] endpoints in the primary endpoint itself,” he said.

Pending validation of these novel endpoints and approaches, however, “HF hospitalization remains the de facto clinical event for use in randomized trials of patients with HF,” the commentary concludes.