A preprint research letter linking the Alzheimer’s drug lecanemab (Leqembi) with excess deaths was withdrawn last week.
The preprint was posted on Research Square and had not been peer-reviewed, but its findings were publicized by articles in Science and the New York Times.
Led by Alberto Espay, MD, MSc, a Parkinson’s disease and movement disorders specialist at the University of Cincinnati (UC), the study aimed to determine whether mortality associated with Alzheimer’s monoclonal antibody treatment — lecanemab or the now-discontinued aducanumab (Aduhelm) — exceeded the background mortality of Alzheimer’s disease alone.
The analysis was based on deaths reported in the FDA Adverse Event Reporting System (FAERS), which states on its website that its records cannot be used to calculate incidence rates due to duplicate reports, unverifiable information, and incomplete data. It also used rough assumptions of how many patients might have been treated with the drugs.
From that, the researchers posted on Research Square that deaths with lecanemab and aducanumab were nearly three or four times higher than expected.
Within a week, several study authors requested that the preprint be withdrawn. The version posted to Research Square did not have the review and permission of all the study’s authors, said Lon Schneider, MD, of the Keck School of Medicine of USC in Los Angeles, who was listed as a co-author and asked to be removed from the study.
“There was a brief 600-word letter that was summarily rejected by JAMA and JAMA Network Open that said much less that I agreed to be on, but that was it,” Schneider told MedPage Today.
What was startling to him wasn’t just that the letter was changed to include “extreme claims” without his knowledge or agreement, but that the New York Times publicized the preprint’s conclusions within 24 hours.
Moreover, because the preprint was assigned a digital object identifier (DOI), it remains a permanent part of Alzheimer’s disease literature. “It’s always up there as they wrote it,” Schneider said.
While the original premise of the research was worthwhile, the way the study was conducted was not, Schneider maintained. “How anyone can believe that there is a three- to four-times higher risk for death that has not previously been identified” is surprising, Schneider noted. “This is absurd, a real misuse of numbers and safety data.”
“There is concern about some individual cases of people who died that they also received other medication such as anticoagulants and TPA [tissue plasminogen activator] that they should not have, or that they had CAA [cerebral amyloid angiopathy],” he continued. “There is an indication of individual deaths related to amyloid antibodies and these medications, of course; all of this has to be followed and studied. Yet, this is not nearly at the rate that the people who wrote this paper were promulgating.”
The number of deaths in the study — 25 associated with lecanemab and 27 with aducanumab — were small and some may have been counted twice, pointed out co-author Eric Widera, MD, of the University of California San Francisco, who also asked to be removed from the study.
“For this reason, and solely for that reason, I didn’t think we could rely on this analysis to accurately quantify the known mortality risk of these drugs until re-analysis has been done — and even then, there may remain significant concerns around this dataset that may make FAERS unreliable,” Widera told MedPage Today.
“The point of this study was not to argue that there is a mortality risk with these drugs. We know that,” Widera stated. “The point of this study was to attempt to quantify the degree of mortality risk using post-marketing data of individuals outside of clinical trials.”
What happened between the time the study was conceived and the preprint was posted remains unclear. Also unclear is how the preprint was so quickly reported in the New York Times and Science.
Science has now included a disclaimer that the preprint was withdrawn, but its article still has a sticking point for UC: The article suggests that UC does not offer lecanemab treatment and claims that to Espay’s knowledge, “not a single patient has gotten the monoclonal antibody lecanemab” there.
However, Alzheimer’s patients have been receiving lecanemab at UC for over a year.
“We initiated lecanemab treatments in October 2023 and currently have 53 patients undergoing treatment, and six eligible but waiting for insurance approval,” Rhonna Shatz, DO, medical director of the UC Memory Disorders Center, told MedPage Today. Three patients who received lecanemab had asymptomatic amyloid-related imaging abnormalities (ARIA), including two who discontinued treatment. None died.
“Dr. Espay’s comments regarding the anti-amyloid treatments reflect his own opinions and not those of the University of Cincinnati, and especially not those of us who are behavioral neurologists and diagnose and treat these patients,” Shatz emphasized.
In a comment posted on Research Square, Espay said the withdrawn research project also was an effort “intended to incentivize the release of clinical trial data” about mortality and adverse effects.
“As this was the only data source on deaths (no other registry is accessible) and no better data was available, the results could not be trusted,” Espay said in an email to MedPage Today. “As a result, most authors voted to have the paper withdrawn.”
“My intention was to raise a serious concern relevant to my patients,” he added. “I am sorry that the data from which we conducted the analyses wasn’t ready.”
Please enable JavaScript to view the