Comment: Medicare needs to conduct own Alzheimer’s drug trial

By Peter B. Bach and Rita F. Redberg / Bloomberg Opinion

Since last summer, Medicare has been evaluating whether to pay for a newly approved, exorbitantly priced drug to treat Alzheimer’s disease. Its decision was always going to be fraught:

The data on the drug’s potential benefits are ambiguous at best, and its risks are considerable. About 40 percent of patients who have taken it have suffered swelling or bleeding in the brain. Others have experienced disabling nausea, dizziness, headaches and confusion. Biogen, the manufacturer, is investigating a patient death.

Then there is the price: $28,200 a year for the average patient. (This is half the drug’s original cost, which Biogen cut to deflect widespread anger.) It’s one reason that Medicare increased its annual premium for 2022 by more than $250.

As the current and immediate past chairpersons of the committee that advises Medicare on which treatments it should cover, we believe it should limit access to this drug, aducanumab, to beneficiaries who volunteer to participate in a new clinical trial. The results of such a trial would reveal whether Medicare should cover the drug more broadly.

Clinical trials are needed anytime a treatment’s benefits are unclear. So far, it seems, aducanumab (brand name Aduhelm) provides none. Last spring, the Food and Drug Administration’s expert advisers voted 10-0 (with one abstention) against approving the drug. Advisers to the Institute for Clinical and Economic Review, an independent organization that evaluates medical treatments, reached a similar conclusion. The European Medicines Agency recorded a “negative trend” vote in November, an early sign the drug will not be approved in the European Union. Japan has also refused to approve it.

The FDA itself agreed that aducanumab’s data “left residual uncertainties regarding [its] clinical benefit.” And so did Biogen: The company stopped two pivotal aducanumab studies early when they indicated that the drug wasn’t working.

Nevertheless, FDA regulators gave the drug “accelerated approval” — a mistake, in our view — with the condition that Biogen conduct further trials to clearly demonstrate that it can slow the course of Alzheimer’s disease. But the company was given nine years to get this research done. Even if Biogen meets the deadline — hardly a certainty — that’s far too long to wait to learn whether aducanumab works. Medicare needs to run its own trial sooner.

This Medicare trial would randomly assign all patients who are considering aducanumab to either the active drug or a placebo. It would include only voluntary subjects who have discussed the treatment with their doctors, given aducanumab’s potential to cause harm. And it would enroll the types of patients Medicare serves; something Biogen’s studies failed to do.

Alzheimer’s disease is twice as common among Black Americans as among non-Hispanic Whites (18 percent versus 10 percent), yet less than 1 percent of patients in Biogen’s primary studies were Black. Between 85 percent and 92 percent of patients suffering from Alzheimer’s have one or more health conditions that would have excluded them from Biogen’s studies, but such individuals would be included in the Medicare trial. The Biogen studies also excluded patients over age 85, and Medicare would not. Biogen’s choice made no sense, because the prevalence of Alzheimer’s rises with age.

There is precedent for Medicare to limit coverage to beneficiaries enrolled in a randomized trial, and it is a time-tested approach to managing the kinds of uncertainties and risks presented by aducanumab. In the late 1990s, Medicare limited a promising but risky surgery that removed parts of the lungs of emphysema patients to clinical trial volunteers. Ultimately, this important trial revealed that the surgery works for patients with one type of lung abnormality, but not those with several other kinds, sparing the latter group an unnecessary and dangerous treatment.

Requiring a clinical trial would also remind other trial sponsors that it is essential to generate clinically meaningful data that are relevant to Medicare’s beneficiaries. As committee chairpersons we have witnessed a steady erosion in the quality of evidence underpinning the drugs and devices that Medicare considers. Study “endpoints” — the changes in patient status that are examined to see whether a treatment is working — are more likely to be laboratory markers or imaging tests that are less relevant or not relevant to patients’ well-being. And study “generalizability” — the extent to which patients in trials are similar to those who would receive the treatment — has fallen from limited to nonexistent.

The best way to help Medicare beneficiaries who develop Alzheimer’s is to make sure the treatments physicians offer are more likely to help than to harm them. But it’s impossible to know whether aducanumab is such a treatment; the data gathered so far don’t indicate that the drug would have any clinically meaningful benefit. The only way for Medicare to safely learn whether the drug works is to limit coverage to patients who volunteer to receive it in the context of a randomized clinical trial.

Peter B. Bach is the chief medical officer of Delfi Diagnostics. He was previously director of the Drug Pricing Lab at Memorial Sloan Kettering Cancer Center.

Rita F. Redberg is a cardiologist and professor of medicine at the University of California, San Francisco.