For the millions of families watching a loved one slip away to Alzheimer’s disease, the arrival of drugs that promised to slow the decline felt like a turning point. Now the most comprehensive independent analysis of those treatments has delivered a blunt verdict: the drugs do not work well enough for patients or families to notice a difference.
The Cochrane Collaboration, widely considered the gold standard for medical evidence reviews, published its findings on April 15, 2026. Researchers examined 17 randomized controlled trials enrolling 20,342 people with mild cognitive impairment or mild dementia due to Alzheimer’s. The drugs studied, a class of monoclonal antibodies designed to strip amyloid protein from the brain, include lecanemab (sold as Leqembi), aducanumab, and donanemab (sold as Kisunla). Across the board, the pattern was the same: tiny shifts on cognitive test scores that fell below the threshold doctors and families can detect in everyday life, paired with serious safety risks.
“These drugs do not slow down the progression of Alzheimer’s disease in a way that patients or their families would notice,” said lead investigator Edo Richard.
A half-point difference on an 18-point scale
The gap between statistical significance and real-world impact is at the heart of the debate. In the landmark CLARITY-AD trial for Leqembi, treated patients scored roughly 0.45 points better than those on placebo on the CDR-SB, a widely used 18-point scale that tracks cognitive and functional decline across domains like memory, judgment, and personal care. That result, published in the New England Journal of Medicine in 2023, was statistically significant. But half a point spread across six domains of daily life is not something a spouse, a daughter, or a neurologist can reliably observe at the bedside.
The Cochrane reviewers found this pattern repeated across the drug class. No anti-amyloid antibody produced a clear improvement in day-to-day functioning. Their conclusion: the overall balance of benefit and harm does not support routine use of these agents for people with early Alzheimer’s.
Safety risks that compound the problem
If the benefits are marginal, the risks are not. All anti-amyloid antibodies carry a significant chance of amyloid-related imaging abnormalities, known as ARIA, which include brain swelling and microbleeds visible on MRI. In clinical trials, these events were common enough to require regular monitoring. In the real world, they have proved more dangerous.
The U.S. Food and Drug Administration issued a Drug Safety Communication recommending earlier and more frequent MRI scans for patients on Leqembi. The agency cited postmarketing pharmacovigilance data, including reports of deaths occurring early in treatment, as the basis for the updated guidance. Causality assessments for those deaths remain ongoing, but the signal was strong enough to prompt action.
The FDA’s move is not a recall. It is a risk-management adjustment that adds cost and logistical burden to an already demanding regimen. Leqembi requires biweekly intravenous infusions at a list price of roughly $26,500 per year, and the Centers for Medicare and Medicaid Services covers it only for patients enrolled in a qualified registry. Each additional MRI scan adds time, expense, and anxiety for patients and the people caring for them.
Unanswered questions the review cannot settle
The Cochrane analysis is powerful because of its scale, but pooled data can obscure variation that matters at the individual level. Whether a specific subgroup of patients, such as those with certain genetic profiles like APOE4 carriers, or people at the very earliest detectable stage of amyloid buildup, might benefit more than the average participant remains an open question. The review was not designed to answer it.
Drug developers have long argued that clearing amyloid plaques is a necessary first step and that longer or larger trials could eventually show clearer benefits. The Cochrane authors pushed back on that framing, noting that decades of amyloid-focused research have not yet produced convincing evidence that targeting the protein alone changes the disease course in ways patients can feel.
Neither Eisai and Biogen, which market Leqembi, nor Eli Lilly, which markets Kisunla, had issued a public response to the Cochrane findings at the time of publication. Both companies have previously maintained that amyloid reduction is a meaningful biological endpoint and that longer follow-up could reveal greater separation between treated and untreated groups.
The Alzheimer’s Association, which has historically advocated for broad access to anti-amyloid therapies, has not yet commented on the April 2026 review. Its response will carry weight with patients, clinicians, and policymakers who look to the organization for guidance.
Why the amyloid hypothesis keeps falling short
Understanding why expectations were so high requires a brief look at the science. For more than three decades, the amyloid hypothesis has dominated Alzheimer’s research. The logic is straightforward: abnormal clumps of amyloid-beta protein accumulate in the brains of people with Alzheimer’s, so removing them should slow or stop the disease. Public health agencies, including the CDC, describe amyloid plaques as a hallmark feature of the condition.
But hallmark does not mean cause. Alzheimer’s involves a cascade of biological processes, including tau protein tangles, neuroinflammation, vascular damage, and synaptic loss, that may be well underway by the time amyloid plaques become detectable. Removing plaques after the downstream damage has started may be too little, too late. The Cochrane findings are consistent with that possibility, though they do not prove it.
What families and clinicians face now
For families weighing whether to start or continue treatment, the Cochrane review sharpens a difficult decision. Caregivers who have spent months arranging biweekly infusion visits, coordinating MRI schedules, and managing the emotional weight of each scan result now confront evidence that, on average, the treatment their loved one endured is unlikely to have produced a change they could see. The calculus is painful: real side effects, real disruption to daily life, and a measured benefit too small to notice at the kitchen table or during a walk around the block.
The FDA’s adverse event data, accessible through the agency’s safety reporting portal, continue to accumulate, though no single summary document has publicly quantified the full scope of post-market harms.
Some specialists may continue offering anti-amyloid drugs to carefully selected patients who understand the limited benefits and accept the risks. Others, particularly in community settings where access to infusion centers and frequent MRI monitoring is limited, may pull back. Professional societies have not yet issued updated practice guidelines incorporating the April 2026 review, leaving room for wide variation in how the evidence is applied across clinics and health systems.
Non-pharmacologic strategies remain central to living with Alzheimer’s: structured support for caregivers, management of coexisting conditions like hypertension and diabetes that accelerate cognitive decline, physical activity, social engagement, and early planning for future care needs. None of these carry the drama of a new drug approval, but none carry the risks either.
The Cochrane review does not close the door on Alzheimer’s drug development. It does, however, make clear that the current generation of amyloid-targeting antibodies has not delivered what patients and families were promised. The next chapter in Alzheimer’s research will likely need to look beyond amyloid alone, and regulators, insurers, and clinicians will need to reckon honestly with what the evidence says right now rather than what future trials might someday show.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
