When the first drugs capable of slowing Alzheimer’s disease won FDA approval, the moment was hailed as a turning point for the nearly seven million Americans living with the condition. But as Leqembi (lecanemab) and donanemab move from carefully controlled clinical trials into hospitals and infusion centers, a more complicated picture is taking shape. Post-approval safety warnings, grueling monitoring schedules, and high dropout rates are forcing families and physicians to weigh modest cognitive benefits against significant real-world burdens. Here is what the strongest available evidence, drawn from federal regulators, Medicare policy, and peer-reviewed research through early 2026, actually shows.
Safety signals arrived faster than expected
The most consequential warning came directly from the FDA. After reviewing adverse event reports submitted through its postmarketing surveillance system, the agency issued a Drug Safety Communication calling for additional MRI scans earlier in the course of Leqembi treatment. The reason: amyloid-related imaging abnormalities, known as ARIA, were appearing sooner and, in some cases, more severely than clinical trials had predicted.
ARIA encompasses brain swelling and microbleeds. Both were known risks during trials, but the FDA found that in routine clinical settings, where patients tend to be older and carry more medical complexity than trial volunteers, these events could be dangerous or even fatal. The agency changed its monitoring guidance specifically because the timeline of risk had shifted.
A peer-reviewed cohort study published in Alzheimer’s Research & Therapy, covering patients screened from July 2023 through January 2025, confirmed that the side-effect profile seen in controlled trials carries over into everyday practice. ARIA-related changes on MRI were common enough to require treatment pauses or outright discontinuation. Some patients experienced symptomatic complications, including headaches, confusion, and worsening neurological function. Outside the guardrails of a clinical trial, these risks proved harder to manage.
Getting from diagnosis to infusion is its own obstacle course
That same cohort study documented something trials never capture: the steep drop-off between a referral for anti-amyloid therapy and a patient actually receiving an infusion. Repeated MRI scans, strict eligibility criteria, scheduling logistics for biweekly infusions, and the need for specialized clinical infrastructure all created friction. Many patients never progressed from initial evaluation to a first dose. Among those who did start, a substantial fraction later stopped because of side effects, logistical strain, or changes in their clinical condition.
Medicare’s coverage framework reinforces these barriers, though by design rather than accident. When the Centers for Medicare and Medicaid Services finalized its national coverage policy for anti-amyloid monoclonal antibodies, it built in a two-track approach reflecting institutional caution about clinical benefit. CMS distinguished between scenarios requiring randomized trial evidence and those where registry-based data collection would suffice, and it emphasized the need for ongoing real-world information on both safety and effectiveness.
After Leqembi received traditional FDA approval, CMS announced broader Medicare access, but with strings attached. Patients and providers must participate in an evidence-collection registry. Treatment sites must demonstrate the clinical infrastructure to handle ARIA monitoring, infusion delivery, and emergency response. In practice, this limits access to well-resourced medical centers, putting the therapy out of reach for patients in rural areas or those without strong caregiver networks.
Plaque clearance helps, but not equally
For donanemab, a secondary analysis of the TRAILBLAZER-ALZ 2 randomized clinical trial examined whether clearing amyloid plaques from the brain actually translates into better cognition and daily functioning. The analysis of posttreatment amyloid levels found that greater plaque clearance did track with slower cognitive and functional decline, but the relationship was far from uniform.
Patients who achieved very low residual amyloid tended to fare better on cognitive tests and daily activity scales. Yet the size of that benefit varied substantially depending on how much brain pathology a patient carried at the start of treatment. In other words, donanemab’s impact appears to be shaped by disease stage and the underlying burden of neurodegeneration, not just amyloid removal alone. For patients with more advanced damage, clearing plaques may not be enough to meaningfully change the trajectory of decline.
The questions that still lack answers
Long-term cognitive outcomes for Leqembi users beyond roughly 18 months of real-world treatment have not been established in published, peer-reviewed data as of spring 2026. The CMS evidence-collection registry is only beginning to accumulate the multi-year information needed to determine whether slowing amyloid buildup preserves memory, daily functioning, and independence over years rather than months.
There is also a troubling selection problem. The cohort study’s documentation of high attrition between screening and infusion raises the possibility that patients who complete the gauntlet of MRI scans, eligibility checks, and infusion scheduling are systematically healthier, better supported, and closer to specialized centers than the typical person with Alzheimer’s. If only this relatively advantaged subset receives and maintains treatment, real-world outcomes may look better than what the broader dementia population would experience. No published study has isolated the effect of access barriers, socioeconomic factors, or geography on who ultimately gets anti-amyloid therapy and how they fare.
Cost-effectiveness remains another blind spot. The list price of these drugs and the expense of repeated MRI monitoring are widely discussed, but no formal economic model from a federal agency has been published weighing those costs against documented clinical gains. Independent organizations such as the Institute for Clinical and Economic Review (ICER) have conducted their own analyses, but federal cost-effectiveness data that could guide Medicare policy decisions remain absent from the public record. Whether these therapies reduce nursing home placement, delay full-time caregiving needs, or lower other medical costs is still unknown.
Questions also persist about how anti-amyloid drugs fit into the broader landscape of Alzheimer’s care. Standard treatments like cholinesterase inhibitors and memantine continue to be prescribed alongside the newer antibodies, but optimal sequencing or combination strategies have not been established. Whether patients with mixed vascular and Alzheimer’s pathology derive enough added benefit to justify the risks and burdens is unclear. As more patients receive these therapies, patterns of off-label use, discontinuation, and switching between agents may emerge without robust evidence to guide those choices.
What the evidence hierarchy tells families right now
Not all evidence in this space carries equal weight, and understanding the hierarchy matters for anyone making treatment decisions. The FDA’s Drug Safety Communication is a primary regulatory document based on the agency’s own review of adverse event reports. It represents the strongest type of safety signal short of a label change or market withdrawal. The CMS coverage decisions are similarly authoritative: they set the actual terms under which patients access these drugs and codify the government’s assessment of how much remains unknown.
The peer-reviewed cohort study provides direct observational evidence from patients treated outside trial conditions. Its value lies in documenting what trials cannot: the friction of biweekly infusions, the burden of frequent imaging, and the reasons people stop treatment. Observational data are subject to confounding, but they reveal how age, comorbidities, and social support shape who actually receives therapy and how tolerable it proves to be. The TRAILBLAZER-ALZ 2 secondary analysis, while rigorous, draws on a trial population screened for specific characteristics, making its findings less directly applicable to the broader patient pool seen in community clinics.
What is largely missing is aggregated outcome data from the CMS registry itself. The registry exists, and providers must participate as a condition of Medicare coverage, but anonymized summaries of cognitive trajectories, functional status, and adverse events have not been released publicly. Until that data becomes available, the strongest evidence about real-world performance comes from independent academic studies rather than the government’s own tracking system.
For families weighing these treatments now, the practical starting point is a detailed conversation with a neurologist or memory specialist about eligibility, the MRI monitoring schedule, and whether the treating facility participates in the CMS evidence-collection program. That discussion should cover not only the potential for slowed decline but also the likelihood of ARIA, the time and travel demands of infusions and imaging, and the real possibility of having to stop treatment if side effects or logistical barriers emerge. In the absence of comprehensive long-term data, decisions about Leqembi and donanemab remain deeply personal, balancing uncertain benefits against concrete risks and substantial demands on patients and the people who care for them.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
