Study finds 70-85% of atypical Alzheimer's patients ineligible for new treatments

A new study has exposed a troubling gap in who gets access to the latest Alzheimer's treatments. Researchers at Mayo Clinic found that between 70 and 85 percent of patients with atypical forms of the disease would be turned away from anti-amyloid therapies—drugs designed to slow cognitive decline—even though many of them are still in the early stages of illness.

The problem lies not with the patients or the drugs, but with how eligibility is being measured. Anti-amyloid treatments like lecanemab and donanemab were tested and approved based on criteria that work well for the most common form of Alzheimer's, the kind that steals memory first. But atypical Alzheimer's doesn't follow that script. It can begin by impairing vision, scrambling language, or eroding the ability to plan and solve problems. When these patients take the standard cognitive screening tests used to determine who qualifies for treatment, they often score poorly—not because they're further along in disease, but because the tests weren't designed to measure what's actually breaking down in their brains.

The study, published in Neurology on August 5, 2026, examined 184 people with confirmed atypical Alzheimer's disease. These patients had one of four distinct presentations: posterior cortical atrophy, which damages vision and spatial awareness; logopenic variant primary progressive aphasia, which affects language and communication; dysexecutive Alzheimer's disease, which impairs planning and executive function; or corticobasal syndrome, which disrupts movement and cognition. Researchers then checked whether each patient would have qualified for the new treatments based on the eligibility standards used in major clinical trials.

The findings were stark. Half to two-thirds of exclusions came down to a single cause: poor performance on cognitive screening tests, particularly the Mini-Mental State Examination, an eleven-question assessment that has become a gatekeeper for treatment access. Yet here was the paradox: many of these same patients were still managing their daily lives reasonably well. They could dress themselves, prepare meals, handle basic tasks. The tests were flagging them as too cognitively impaired to treat, even though their functional abilities suggested otherwise. Dr. Dror Shir, the study's lead author from Mayo Clinic's Jacksonville campus, noted that patients with atypical Alzheimer's may score poorly on tests of visual-spatial reasoning or executive function precisely because those are the domains being attacked by their disease—not because they've progressed further than they actually have.

Other barriers to treatment eligibility also emerged. Brain imaging findings excluded 22 to 27 percent of participants, while disease severity classifications—with 19 to 23 percent deemed to have moderate or severe dementia—blocked another group. But the screening test remained the primary obstacle. "Current tools might overestimate the functional stage of the disease," Shir said, pointing to a fundamental mismatch between how severity is being measured and how atypical Alzheimer's actually presents itself.

The implications are significant. Atypical Alzheimer's disease has long been underrecognized and underrepresented in clinical research. These patients already face diagnostic delays and fewer specialists trained to identify their condition. Now, even when they do get diagnosed early enough to potentially benefit from disease-modifying treatments, the eligibility criteria designed for a different disease presentation are locking them out. The study was limited by its retrospective design—researchers estimated eligibility based on existing medical records rather than evaluating patients directly—but the signal is clear. Treatment guidelines may need to be rethought to account for how these less common forms of Alzheimer's unfold, ensuring that patients aren't excluded from potentially beneficial therapies simply because their disease doesn't fit the standard template.