Study Alleges Mortality Miscount in FOURIER Trial; TIMI Group Disagrees
Clinical trialists, including Marc Sabatine, as well as evolocumab’s manufacturer reject the reanalysis as flawed and sensationalistic.
A “reanalysis” of mortality data from a large PCSK9 inhibitor cardiovascular outcomes trial suggests that some deaths may have been misclassified, leading the authors to conclude that the potential harm from treatment is higher than initially reported, a claim roundly rejected by the trialists themselves.
Based on their readjudication of outcomes using the trial’s original clinical study report (CSR), authors of this new paper claim that the risk of cardiovascular mortality is actually higher with evolocumab (Repatha; Amgen) than what was initially reported in the FOURIER trial, although the difference compared with placebo was still not statistically significant. Nonetheless, researchers led by Juan Erviti, PharmD, PhD (Navarre Health Service, Pamplona, Spain), say their findings “indicate that complete restoration of all clinical outcomes from the FOURIER trial is required” to fully understand the mortality signal.
“Meanwhile, clinicians should be skeptical about benefits versus harms of prescribing evolocumab for patients with established atherosclerotic cardiovascular disease,” they conclude.
Their reanalysis, published this week in BMJ Open, is the work of the Restoring Invisible and Abandoned Trials (RIAT) initiative, an international collaborative aimed at addressing bias in research trials.
Needless to say, the new analysis didn’t sit well with the FOURIER investigators.
“We stand 100% by our results,” said Marc Sabatine, MD, MPH (Brigham and Women’s Hospital, Boston, MA), who led the TIMI-run trial testing evolocumab, adding that this readjudication of events is flawed and characterizing the RIAT group’s conclusions as “highly erroneous.”
“Moreover, there’s concern that people will only see the top-line conclusions and won’t take time to read through the study’s methodology,” he stressed.
“It just sows misinformation,” Sabatine told TCTMD, citing the wealth of data from large cardiovascular outcomes trials with agents that inhibit PCSK9, including more to come with inclisiran (Leqvio; Novartis), such as ORION-4 and VICTORION-2P. “We’re going to have multiple trials with long-term follow-up that will be high-quality studies to clearly cement not only the efficacy but also the safety of PCSK9 inhibition,” he said. “In the meantime, I think papers like this are ultimately a disservice to not only clinicians but also to patients.”
Steven Nissen, MD (Cleveland Clinic, OH), who wasn’t involved in FOURIER, believes it to be a rigorous, carefully conducted trial by highly respected investigators and academic research organization.
“The cardiovascular outcomes in the FOURIER trial were adjudicated by the clinical events committee [CEC] at the TIMI Group,” Nissen told TCTMD. “The CEC members were blinded to treatment assignment. The CEC at TIMI is one of the world’s most experienced adjudication teams, and their process has been audited by many external groups. I have great confidence in the integrity of the TIMI group and find it inconceivable that bias was present in their adjudication.”
Nissen, who has been involved in numerous studies testing various LDL-lowering therapies, said he is absolutely not concerned about any mortality harms suggested by the reanalysis. None of the LDL-lowering medications, such as statins, ezetimibe, and now the PCSK9 inhibitors, are associated with increased risk of mortality, he said. Adjudication of clinical outcomes is a challenging process, he agreed, noting that it’s always possible another group will arrive at a different conclusion. This happens because the exact standards for determining the cause of death may differ slightly amongst the adjudicating teams.
“The scientific and ethical rationale for establishment of groups like TIMI arises for the need to have unbiased conduct of clinical trials by an academic team independent from the trial sponsor,” said Nissen. “TIMI has conducted hundreds of clinical trials over many decades that have been broadly accepted by the scientific and regulatory communities.”
Erviti and the RIAT investigators have not yet responded to TCTMD’s requests for an interview.
Bump in Hazard Ratio, but Still Not Significant
Evolocumab was first PCSK9 inhibitor approved by the US Food and Drug Administration in 2015 for lowering LDL-cholesterol levels in people unable to get to goal with maximally tolerated statin therapy. Following FOURIER, the drug was indicated to prevent MI, stroke, and coronary revascularization in patients with atherosclerotic cardiovascular disease (ASCVD).
In FOURIER, published in the New England Journal of Medicine in 2017, treatment with evolocumab reduced the study’s primary endpoint (cardiovascular death, MI, stroke, hospitalization for unstable angina, or coronary revascularization) by 15% compared with placebo, but there was no reduction in the risk of cardiovascular mortality alone (HR 1.05; 95% CI 0.88-1.25). In the evolocumab and placebo treatment arms, there were 251 and 240 cardiovascular deaths, respectively.
For their analysis, Erviti and colleagues turned to the original CSR, which they obtained from Health Canada and European Medicines Agency. The FOURIER CSR is a 25,000-page document that provides information about the trial’s protocol, amendments, inclusion and exclusion criteria, outcome definitions and measurements, efficacy and safety results, and statistical plan.
In the CSR, there are reports for 870 deaths, including a brief summary of the clinical record as well as the cause of death declared by the local investigator and FOURIER clinical-events committee. Overall, the cause of death adjudicated by the CEC differed from the local investigators in 41.4% of cases. This may occur when a death is adjudicated as sudden cardiac death but initially reported as MI or when a noncardiovascular death is later adjudicated by the CEC to be an MI.
The researchers focused on these discordant reports and reviewed the CEC’s final decisions to either confirm or readjudicate the cause of death. They identified 11 more deaths from MI in the evolocumab-treated patients and three fewer deaths from MI in the placebo group than reported in the original NEJM manuscript. After readjudication, there was a slightly higher hazard of cardiovascular mortality with evolocumab than with placebo (HR 1.20; 95% CI 0.95-1.51). Instead of 251 and 240 cardiovascular deaths in the evolocumab and placebo arms, they counted 150 and 125, respectively. These cardiovascular deaths were readjudicated to be from noncardiovascular or unknown causes.
The degree of inconsistencies, according to the researchers, “puts into question the validity of the cause of death adjudication by the FOURIER clinical-events committee.” While discrepancies affected both treatment arms, the “misadjudications improperly reduced cardiovascular deaths in the evolocumab group,” they write.
Responding to ‘Reanalysis’
In response to the study, Sabatine highlighted the different ways the TIMI study group collected and analyzed data compared with the RIAT researchers.
“Anytime there is a cardiovascular event, that triggers the collection of a dossier containing all the relevant and available source documents,” he said. “If somebody had a heart attack, we need to see discharge summary, the lab data for the troponin values, the ECGs, the angiograms—we get all the source documents. If there’s been a death, we get the autopsy reports.”
These dossiers are reviewed independently by two board-certified clinicians—a cardiologist for the cardiovascular outcomes and a neurologist for the cerebrovascular outcomes—who are blinded to treatment. The criteria for clinical outcomes, including the type of death, are based on standard FDA definitions, said Sabatine. At the end of a large cardiovascular outcomes trial, the FDA even audits the adjudication process and results, which they did in FOURIER. The agency, he said, found nothing of concern.
Clinical trials also require the CSR narrative, which is generated 24 hours after a patient has a serious adverse event, such as death. These narratives describe the event but may be incomplete, said Sabatine, noting they may be based solely on information in the death certificate. For this reason, it’s not surprising to have the initial cause of death listed in the CSR narrative by the local investigator differ from the CEC-adjudicated cause of death. The CEC has much more information on which to base their decision, said Sabatine.
“That process will naturally result in reclassification,” said Sabatine. “That’s the very purpose. There’s nothing nefarious that it doesn’t perfectly agree with the investigators.”
Sabatine said that among the deaths initially deemed to be from unknown causes by local investigators, the CEC classified roughly half to be the result of cardiovascular disease, which makes sense given the ASCVD patient population. As for missed MIs as the cause of death, as the RIAT group alleges, Sabatine argued that the patient dossiers used by the CEC do not support such an attribution. He also pointed out that because of the lack information in the CSR, the RIAT researchers readjudicated numerous deaths from cardiovascular disease to undetermined causes. Because fewer deaths were shifted from cardiovascular to undetermined causes in the evolocumab arm, this resulted in the higher hazard ratio than reported in the NEJM paper (HR 1.05 vs 1.20).
“It’s highly selective reporting on their part,” said Sabatine.
Speaking more broadly, Sabatine said the survival benefit of lowering LDL cholesterol takes time to emerge. In the 4S trial with simvastatin, for example, the mortality curves didn’t begin to separate until 2 years. “We never expected to see a mortality benefit in FOURIER, which only had a median follow-up of 2.2 years,” he said.
Longer follow-up data from the FOURIER-Open Label Extension (OLE) study, which was published last year in Circulation, suggested that with enough time evolocumab could reduce the risk of cardiovascular mortality (HR 0.77; 95% CI 0.60-0.99). In ODYSSEY, a large cardiovascular outcomes trial testing the other FDA-approved PCSK9 inhibitor, alirocumab (Praluent; Sanofi/Regeneron), was associated with a 15% lower risk of all-cause mortality.
To TCTMD, Nissen said the Cleveland Clinic’s academic research organization, known as C5Research, which he heads up, has a professional, experienced team of adjudicators, just like the TIMI group, with all clinical outcomes analyzed according to rigorous standards. An unblinded post hoc analysis introduces its own significant bias, he said.
“I think we have to discount this kind of second-guessing,” said Nissen. While he praised BMJ Open for publishing the data, noting that while full transparency in clinical research is important, “anybody can go in, post hoc, and arrive at a slightly different answer.” The implication from the research, though, is that the FOURIER researchers, or the drug’s sponsor, are manipulating data, said Nissen, when that is not the case.
Sabatine made a similar point, stating that the language in the paper, even its title, is highly sensationalistic and rife with innuendo.
“‘Restoring mortality data?’ What does that even mean?” he said. “And then, ‘A reanalysis based on regulatory data?’ By definition, the FDA has all the regulatory data and they agreed with our analysis for cardiovascular deaths. It’s not a reanalysis based on regulatory data.”
On social media, some cardiologists pointed out that it’s very difficult to ascertain the exact cause of death in clinical trials and said for that reason all-cause mortality might be a better endpoint to use.
Amgen, the maker of evolocumab, also discounted the study findings. In a statement to TCTMD, an Amgen spokesperson said they “stand behind the integrity and validity of the FOURIER trial and results.
“We stand confident in the prospectively defined clinical trial event adjudication process that was applied to FOURIER and believe that the most rigorous classification process for adjudication was used by the TIMI Study Group,” according to Amgen. “This process was based on the uniform Standardized Data Collection for Cardiovascular Trials Initiative established by the Food and Drug Administration, as well as guidance from other regulatory agencies and was consistent with the adjudication process used in other cardiovascular outcomes trials.”
The company added that treatment guidelines and recommendations around the world recommend PCSK9 inhibitors because the treatment is a safe and effective way to reduce cardiovascular risk in patients with ASCVD who need to intensity lipid-lowering therapy. They state that evolocumab has been used in 1.7 million patients globally and tested in more than 50,000 patients in clinical studies, including the FOURIER-OLE study with more than 8 years of follow-up.
Erviti J, Wright J, Bassett K, et al. Restoring mortality data in the FOURIER cardiovascular outcomes trial of evolocumab in patients with cardiovascular disease: a reanalysis based on regulatory data. BMJ Open. 2022;12:e060172.
- Amgen funded the FOURIER trial.
- Erviti and Nissen report no relevant conflicts of interest.
- Sabatine reports institutional grants from Abbott, Amgen, Anthos Therapeutics, Aralez, AstraZeneca, Bayer, Daiichi Sankyo, Eisai, GlaxoSmithKline, Intarcia, Ionis, Medicines Company, MedImmune, Merck, Novartis, Pfizer, Poxel, Quark Pharmaceuticals, Takeda, and additional grant support through Brigham and Women’s Hospital from ARCA Biopharma Inc, Janssen Research and Development, Siemens Healthcare Diagnostics Inc, Softcell Medical Limited, Regeneron, Roche, and Zora Biosciences. He reports consulting fees from Althera, Amgen, Anthos Therapeutics, AstraZeneca, Beren Therapeutics, Bristol Myers Squibb, CVS Caremark, DalCor, Dr Reddy's Laboratories, Dyrnamix, Esperion, IFM Therapeutics, Fibrogen, Ionis, Intarcia, Janssen, MedImmune, Merck, Moderna, Novartis, Novo Nordisk, and Silence Therapeutics.