Lancet opens investigation into anti-clotting drug trial after BMJ report

My latest article, originally published in The BMJ on 30 Dec 2022, reports on the “Record4 trial,” which tested the anti-clotting drug, rivaroxaban (XARELTO®).

The drug was initially approved in the US and the EU for the prevention of deep vein thrombosis following hip & knee replacement, but regulators later expanded its use as a first-line therapy for stroke prevention in people with a common heart rhythm disorder known as atrial fibrillation.

In 2019, the FDA also approved rivaroxaban to help prevent blood clots in acutely ill patients without a high risk of bleeding, during and after hospitalisation.

Rivaroxaban has now been suggested as a prophylaxis therapy in COVID-19 patients who’ve been discharged from hospital and are at high risk of developing blood clots.

The drug, dubbed the “warfarin alternative,” has been prescribed more than 80 million times in the US alone. In 2020, it was rated 10th in the global sales rankings and generated almost US$7 billion that year.

Lancet opens investigation into anticlotting drug trial after BMJ report

The Lancet has said it will open an investigation into a trial of the anticlotting drug rivaroxaban, after an investigation by The BMJ into the regulatory oversight of clinical trials by the US Food and Drug Administration.

During an inspection audit of the Record4 study into rivaroxaban, the FDA identified numerous and serious data integrity deficiencies at eight of the study’s 16 clinical trial sites. But when the trial was published in the Lancet in 2009 the investigators made no mention of the data integrity issues, and since then the paper has been cited more than 1100 times by other researchers unaware of the problems.

All Record4 authors claimed that they had “full access to all of the data and analyses, and confirm[ed] the accuracy and completeness of the data reported.” But when The BMJ raised the data integrity issues with them in October this year [2022], the study’s lead author, Alexander Turpie, emeritus professor of medicine at McMaster University, Ontario, said that he was unaware of the FDA violations. Turpie and the Lancet told The BMJ that they would investigate the matter.

Then, on 9 December 2022, more than a decade after the Record4 trial was published, the Lancet published a correction by Turpie. It read, “On behalf of the Record4 Steering Committee and my coauthors, I regret and apologise for the errors in the 2009 article. . . We are disappointed to learn this information, particularly at this late date, and it is most important that the medical community be informed of the inaccuracies.”

Underreporting of adverse events

Turpie outlined some of the issues identified by the FDA’s audit report, such as the need to exclude data from 652 (21%) of the 3148 trial participants, because they might have been randomised postoperatively, rather than preoperatively, as stated in the protocol and the Lancet article. Turpie also wrote that adverse events and serious adverse events had been underreported at 9.9% of the sites audited for Record4. “Adverse events and serious adverse events in Table 4 [“Safety outcomes”] should be disregarded because we cannot verify concerns about reporting in the remaining 90.1% of non-audited sites,” he wrote. But Turpie claimed that “there was no concern regarding the primary efficacy and safety outcomes described in the FDA and sponsor audits.”

In the week beginning 12 December, when The BMJ made initial inquiries about whether Turpie or the Lancet was considering retracting the study in light of the data integrity issues, Turpie wrote, “We feel the publication in the Lancet correspondence is sufficient.” The Lancet agreed, saying, “Retractions are issued when the magnitude of the error is such that a research paper is no longer reliable. In this instance, the primary efficacy and primary safety outcomes are not affected, so in line with our corrections policy, a correction has been issued alerting readers to the concerns about certain data in the study.”

However, the Lancet has since reconsidered its position after The BMJ presented it with details of the FDA review, which has been available on the FDA website since at least 2011. The review concluded that the data in the Record4 study were in fact “unreliable.” Specifically, the FDA’s Division of Scientific Investigations found that the unreported serious adverse events in Record4 brought into question the “adequacy and completeness of the Record4 safety data submitted to the agency” and that “sites audited by or inspected by FDA ended with an evaluation that the data from the sites was not reliable, reflective of drug accountability deficiencies and other violations of good clinical practice, including postoperative randomization, falsification, missing records, and improper study drug storage.”

The Lancet has responded to this new information by saying, “The Lancet journals take issues relating to scientific misconduct extremely seriously and follow best practice guidelines as set by the Committee on Publication Ethics (COPE). The corrections to the Record4 study made were after correspondence with the authors. Now that we have had sight of the FDA report, we will continue to investigate further and provide updates as necessary.”

Peter Wilmshurst, a consultant cardiologist at the Royal Stoke University Hospital and a whistleblower who has investigated conduct in research trials, has also analysed the FDA’s review document. He disputed Turpie’s claim that there was no concern regarding the trial’s primary safety and efficacy outcomes. “That’s not what the FDA concluded. In fact, the FDA said that it did not accept the Record4 data because they were not reliable, and the problems identified included falsification. At that point, the FDA and the drug company should have told the investigators,” said Wilmshurst.

Poor documentation

Another major concern about the trial was the poor documentation of serious adverse events. Wilmshurst said, “Turpie himself said that table 4 can’t be relied on, which is the safety outcomes data. Safety is a big part of a trial, and they’ve missed two thirds of the adverse events. If they missed those, what else have they missed? “The other thing that struck me was the people who died in the study period . . . The investigators said some deaths were unexplained. If someone in a clinical trial has an operation and they die within the 17 day treatment period, how can the cause of death be allowed to go unexplained?”

In response to Wilmshurst’s points made, Turpie said, “FDA is not the US national arbiter of what data from controlled studies is reliable, or not. Plenty of controlled studies are submitted to journals and rejected as unreliable, then resubmitted elsewhere, where they may face the same fate, or ultimately be accepted and published. We think the Record4 data is in that category: unreliable for FDA but sound in its primary efficacy and safety conclusions.”

He added, “FDA doesn’t report its conclusions to investigators in the US. It does report to sponsors. That the study chair and investigator colleagues did not know of the FDA review’s conclusions until you prompted us to seek them out in 2022 speaks for itself. It’s very disappointing if not, frankly, deplorable.”

Wilmshurst, who recently called for the retraction of another Lancet study, has welcomed the journal’s move to investigate the issue, adding that the “investigation should be done swiftly.”

The trial’s sponsor, Janssen and Bayer (formerly Johnson and Johnson), said that the Record4 study was not the basis for any approval of rivaroxaban in Europe or the US for the prophylaxis of deep vein thrombosis after total hip and knee replacement. “We continuously assess the safety and efficacy of [rivaroxaban] in routine clinical use and study after study continues to confirm the positive benefit-risk profile observed in the clinical trials,” it said in a statement to The BMJ.

Maryanne’s post-article notes:

The drug company defends the FDA’s approval of rivaroxaban for clot prevention after hip and knee surgery because, it says, the decision did not rely upon data from the Record4 trial. Instead, it relied upon data from the previous Record trials (1, 2 and 3).

That is true. However, the FDA also carried out inspections of sites for Record 1,2 and 3 trials and found multiple violations and flaws, arguably casting doubt about the integrity of all four Record trials that investigated rivaroxaban.

See this [LINK] for a more detailed look at the FDA’s review.

---