Did Regional Differences Muddy Results in COMMANDER-HF?

Geographical differences in treatment uptake among participants in the COMMANDER-HF trial may muddy the interpretation of the study, which ultimately saw no impact of rivaroxaban on the primary outcome in the overall trial, a new analysis suggests.

“These differences were most marked in Eastern European countries that presented lower event rates and treatment adherence,” write João Pedro Ferreira, MD, PhD (Université de Lorraine, Nancy, France), and colleagues in a paper published online February 3, 2021 in JACC: Heart Failure.

Ferreira and colleagues believe the findings have implications for future global drug trials, which they say “should require objective inclusion criteria that are less prone to investigator interpretation, ensure that financial incentives to recruitment are carefully controlled, and confirm that participants are aware of the consequences for others if they are nonadherent without good reason.”

Regional differences in medication use and trial outcomes became a flash point for discussion in the PLATO trial, as reported by TCTMD, that formed part of a US Department of Justice investigation into trial conduct. The topic has also reared its head in the setting of COVID-19, where clinical trials have been halted, slowed, or progressed as normal to varying degrees as different parts of the world grapple with the pandemic, posing potential problems for interpreting the results.

COMMANDER-HF tested 2.5 mg of rivaroxaban (Xarelto; Bayer/Janssen) twice daily in patients from 32 countries who had heart failure and underlying CAD and had been recently treated for an episode of decompensated heart failure. At a median of 21.1 months, there was no difference in the primary outcome of MI, stroke, or all-cause death with rivaroxaban compared with placebo (HR 0.94; 95% CI 0.84-1.05).

In the new analysis, Ferreira and colleagues found that when compared against patients from the other regions, those from Eastern Europe had lower rates of the primary outcome: 11.6 per 100 person-years versus 19.5 in Western Europe and South Africa, 14.2 in North America, 17.7 in Asia-Pacific, and 18.6 in Latin America, with a similar pattern for the composite of MI, stroke, and CV death. Eastern European patients also had less bleeding, including bleeds that required hospitalization. Additionally, whereas approximately 5% of patients in the other regions had undetectable blood concentrations of rivaroxaban at 4 weeks, 21% of Eastern European patients had undetectable levels at that time.

Commenting on the study for TCTMD, Mitchell Psotka, MD, PhD (Inova Heart and Vascular Institute, Falls Church, VA), said it’s also important to note that there was no evidence of treatment-by-region heterogeneity for the primary outcome, meaning while the differences in regional performance are problematic, they were not significant enough to upend the primary outcome of COMMANDER-HF.

“What this speaks to is now you have two large outcome-driven trials in heart failure, this trial and TOPCAT—which was a trial of spironolactone in patients with heart failure and preserved ejection fraction—where these regional differences in behavior, both in the patients who are enrolled and in the blood-measured adherence to the treatment regimen, indicate substantial differences and nonadherence to the treatment regimen,” he added. “That will directly lead to biasing your trial results to the null hypothesis, meaning that you are much more likely to show no benefit because many of your patients are apparently not being treated with the medication under study.”

Addressing Problem Areas

Ferreira and colleagues suspect lower adherence occurred in Eastern Europe for several reasons. One is that investigators, in order to enhance recruitment, may have been less selective about enrolling people they believed would be compliant. Reimbursement for enrolling patients into trials in some countries is made directly to physicians, and in other cases employment or promotion may depend on ability to enroll patients, which Ferreira and colleagues say amounts to “perverse incentives” to include study participants they know are less likely to be adherent.

Another second possibility, they say, is that those investigators “might have had less time to spend on educating their patients about the protocol because of high recruitment rates.” As for the patients themselves, enrolling in a trial may be the only guarantee of accessing healthcare in certain situations, especially in Eastern European countries that lack public funding for treatment and essential medications.

In an editorial, Marvin A. Konstam, MD (Tufts Medical Center, Boston, MA), says the analysis sends an important message to international researchers that accelerating enrollment to boost sample sizes asserts “collateral damage” on the entire trial and should not be allowed.

“If Eastern Europe is a problem, then far from abandoning the region, sponsors or academic  associations should work to educate trial personnel, emphasizing that they will not sacrifice trial integrity for unbridled enrollment,” Konstam writes. “There is a variety of ways to address problematic regional variation after it has been identified through monitoring.” He adds that Eastern Europe is unlikely to be the sole problem, since “irregularities can undoubtedly be found if carefully sought within some high-enrolling sites across all regions.”

So, what can be done to curtail negative influences and poor adherence issues before they impact large, international trials? Ferreira and colleagues say one solution may be mandating collection of blood and urine samples during the follow-up to confirm medication adherence.

Konstam points out that while halting enrollment in an aberrant region is the “ultimate solution,” it also should be considered that “trials may benefit from more rigorous site selection, particularly drawing upon historic site information available to the sponsor or, in the future, compiled more broadly by a consortium of sites or an alternative outside entity.”

The Heart Failure Collaboratory, a public-private partnership with US Food and Drug Administration, is attempting to help in this effort. Psotka, one of the Collaboratory’s project leads, said the group is increasingly focused on refining global clinical trial infrastructure to make findings more reliable and usable.

“We also are working on creating global networks and national networks of sites that are essentially vetted in terms of their ability to provide high-quality data for clinical trials and have a history of providing that,” he said. “The other thing that we're trying to do from a different vantage point is really try to figure out how to leverage all of our digital health and electronic health records to better assist with this process and make these data collections more streamlined, but also more complete.”

Psotka said there’s no doubt that the issues are multifactorial and complicated.

“It's probably also important to recognize that there are cultural differences between the patients in these areas and [with regard to] their underlying beliefs about chronic disease management,” he concluded.