Previously in Parts 1 and 2
In AMBAR Part 1, we explored how the AMBAR trial achieved effect sizes of 52–71% in slowing Alzheimer's progression — strongest in moderate-stage patients, with some mild-stage patients showing actual improvement on certain measures. Benefits emerged gradually over 6–9 months and continued diverging through month 14.
In AMBAR Part 2, we examined the mechanistic evidence and independent validation. CSF biomarker patterns didn't fit a simple amyloid-removal hypothesis, suggesting multi-factorial mechanisms at work. A 2025 Argentina real-world study confirmed 45–88% reduction in decline across different measures. And when compared head-to-head with FDA-approved monoclonal antibodies, AMBAR showed larger effect sizes (52–71% vs. 26–40%) with a superior safety profile (0.6% vs. 12–40% ARIA risk).
Now in Part 3, we explore the most surprising discovery from AMBAR and what it means for the future of Alzheimer's treatment.
The Amyloid-Negative Finding: A Game-Changer
Perhaps the most important finding from AMBAR came not from the original 2020 publication, but from a follow-up analysis published as a Letter to the Editor in 2022.
The Discovery
When the AMBAR investigators analyzed CSF samples to determine amyloid status, they made a surprising discovery. Of the 298 patients with available biomarker data, approximately 70% were amyloid-positive (CSF A-beta-42/A-beta-40 ratio <0.05) and approximately 30% were amyloid-negative. This happened because AMBAR enrolled patients based on clinical diagnosis of Alzheimer's disease using NINCDS-ADRDA criteria, not based on biomarker confirmation — a phenotypic diagnosis based on progressive memory loss and cognitive decline.
When they analyzed treatment response by amyloid status, both amyloid-positive and amyloid-negative patients showed similar clinical benefit — with amyloid-negative patients showing statistically significant improvements in cognitive function (ADAS-Cog), global function (CDR-sb), and clinical impression of change (ADCS-CGIC).
Why This Is a Major Finding
This discovery has profound implications for how we understand and treat Alzheimer's disease.
1. It's not just about amyloid. If plasma exchange worked solely by removing amyloid from the brain, it shouldn't work in amyloid-negative patients. The fact that it does means other mechanisms are at play.
2. It addresses a critical unmet need. Current FDA-approved treatments (lecanemab, donanemab) require amyloid-positive biomarker status. If you have an Alzheimer's phenotype but are amyloid-negative, you have no treatment options. The CTAD Task Force (Clinical Trials on Alzheimer's Disease) specifically highlighted this gap: patients with AD-like symptoms but negative amyloid biomarkers represent an "unmet need" that current therapies don't address. Plasma exchange appears to work for these patients.
3. It suggests broader applicability. If the treatment works through multiple mechanisms — reducing inflammation, oxidative stress, vascular dysfunction, removing various toxic proteins — it might benefit other neurodegenerative conditions beyond just amyloid-driven AD.
4. It changes how we think about treatment targets. The pharmaceutical industry has spent 20+ years and tens of billions of dollars targeting amyloid. The implicit assumption has been: remove amyloid plaques, cure Alzheimer's. The AMBAR results, particularly the amyloid-negative finding, suggest we should be targeting the multiple downstream consequences of aging and neurodegeneration rather than focusing solely on upstream amyloid accumulation.
"About 30% of AMBAR patients were amyloid-negative — yet they benefited just as much as amyloid-positive patients. This means the treatment works through mechanisms beyond just removing amyloid plaques."
Why Haven't You Heard About This?
If plasma exchange showed effect sizes of 52–71% with minimal safety concerns, and these findings have been independently replicated, why isn't this the standard of care for Alzheimer's disease? That's a reasonable question with several answers.
The monoclonal antibodies are backed by major pharmaceutical companies — Biogen, Eisai, Eli Lilly — with massive marketing budgets and extensive relationships with prescribing physicians. These companies spent billions on drug development and clinical trials. They have dedicated sales forces, direct-to-consumer advertising campaigns, and fund continuing medical education programs and patient advocacy organizations. AMBAR was sponsored by Grifols, a substantial company in the plasma protein therapeutics industry but one without a large neurology presence, a neurology sales force, or a comparable marketing budget. Grifols is promoting an off-label use of existing products rather than a patented new drug. The economic incentive structure heavily favors the monoclonal antibodies despite potentially inferior efficacy and safety.
The regulatory landscape creates additional barriers. The monoclonal antibodies went through the traditional FDA drug approval pathway — physicians can prescribe with confidence, insurance companies are required to cover them in many cases, there's a clear standard of care, and legal liability is reduced. Plasma exchange with albumin replacement remains an off-label use of existing approved products. Both plasma exchange and albumin are individually FDA-approved, but not specifically for Alzheimer's disease. This means some physicians are hesitant to prescribe, insurance coverage is less certain, and implementation requires more institutional support.
There are practical and logistical challenges as well. Prescribing lecanemab or donanemab is straightforward: write a prescription, the patient goes to an infusion center every 2–4 weeks for a 30–60 minute infusion. Plasma exchange is more complex — it requires specialized apheresis equipment, trained nurses and technicians, 2–3 hours per session, appropriate venous access, and coordination between neurology and apheresis services. Many medical centers don't have apheresis capabilities on-site. It's simply easier to prescribe a drug than to implement a procedural intervention.
Medical conservatism plays a role, too. Neurologists are trained to think in terms of pharmaceutical interventions. Plasma exchange represents a different paradigm — a procedural, extracorporeal therapy more commonly associated with hematology or nephrology. There's natural resistance in any specialty to adopting treatment modalities that fall outside the typical scope of practice. A neurologist prescribing an antibody infusion feels familiar. Referring a patient for repeated plasma exchange sessions feels foreign.
Study size and replication lag also matter. While AMBAR was well-designed and adequately powered, it enrolled 322 patients — compared to 1,795 for lecanemab and 1,182 for donanemab. In medicine, there's often a desire to see large-scale replication before widely adopting a new treatment. The Argentina real-world study provides important validation, but it was smaller (32 patients) and used historical controls rather than concurrent randomization. A large, multicenter, randomized controlled trial comparing plasma exchange to standard of care would probably be needed to shift widespread practice.
The complexity of mechanism works against adoption as well. Monoclonal antibodies have a simple, compelling narrative: "Remove the toxic plaques that cause Alzheimer's." Plasma exchange has a more complex story — it reduces damaged proteins, lowers inflammation, restores antioxidant capacity, improves vascular function, and clears various toxic factors through multiple mechanisms that we don't fully understand yet. The first story is easier to communicate, even if the second might be more accurate and effective.
Finally, there's an uncomfortable economic reality. The monoclonal antibodies cost around $26,000–28,000 per year for the drug alone, plus PET scans, MRI monitoring, and infusion services — generating significant revenue for healthcare systems. Plasma exchange with albumin replacement is also expensive, but the reimbursement structure is different and potentially less favorable to hospitals in some settings.
"AMBAR's effect sizes of 52–71% exceed those of FDA-approved monoclonal antibodies at 26–40%, with a fraction of the safety risk. The question isn't whether the evidence supports plasma exchange — it's why the medical establishment hasn't caught up."
Limitations and Uncertainties
In the interest of balanced reporting, it's important to acknowledge the limitations of the AMBAR study and the uncertainties that remain. Science requires honest evaluation of both strengths and weaknesses.
Study Limitations
Industry sponsorship. The study was funded by Grifols, which manufactures therapeutic albumin. While the trial was independently conducted with oversight from academic investigators and a Data Safety Monitoring Committee, the potential for bias exists.
Single trial. Until the Argentina study, AMBAR was a single trial. While the results were internally consistent and showed effects across multiple outcomes, medicine generally requires replication before establishing new treatments.
Sample size. At 322 patients, AMBAR was adequately powered for its primary endpoints but smaller than the pivotal trials for competing therapies.
Treatment arm ambiguity. None of the three treatment arms (low-albumin, low-albumin+IVIG, high-albumin+IVIG) showed clear superiority when analyzed separately. The significant effects were seen when all three were pooled, making it unclear which specific protocol is optimal.
Dropout rates. While similar to other AD trials, the dropout rates of 34–35% in the IVIG-containing arms and 20–22% in the non-IVIG arms are substantial, likely reflecting procedural burden and potential IVIG side effects.
Biomarker inconsistency. The CSF biomarker results don't provide a clear mechanistic explanation for the clinical effects, particularly in mild AD patients.
Blinding challenges. While the study used sham plasma exchange to maintain blinding, procedural studies are inherently challenging to blind perfectly. However, the high completion rate in the placebo group (80%) suggests blinding was reasonably effective.
Remaining Uncertainties
The optimal treatment protocol remains undefined — including the number of intensive-phase sessions needed, the ideal maintenance schedule, the necessary plasma exchange volumes, and whether IVIG is beneficial, neutral, or harmful.
Treatment duration is another open question — benefits were still emerging at 14 months, and we don't know how long treatment should continue or whether benefits persist after stopping.
Patient selection criteria need refinement — the evidence suggests moderate AD patients benefit most, but questions remain about the role of APOE4 status, the ideal point in the disease course to begin treatment, and whether mild cognitive impairment would also respond.
The mechanism of action is incompletely understood — we don't know whether the primary benefit comes from amyloid clearance, anti-inflammatory effects, antioxidant restoration, vascular improvement, or some combination, and the amyloid-negative finding only deepens the mystery.
Cost-effectiveness, long-term outcomes (nursing home placement, survival, effects beyond 14 months), and the potential for combination therapies with monoclonal antibodies or other approaches all require further investigation.
Conclusion
The AMBAR study demonstrated that plasma exchange with albumin replacement can slow cognitive and functional decline in Alzheimer's disease patients, with effect sizes of 52–71% across multiple outcome measures. These findings have been independently validated in a real-world cohort from Argentina.
Looking Ahead
The AMBAR findings — particularly the amyloid-negative response and multi-factorial mechanism — suggest plasma exchange may be addressing fundamental aging processes rather than just Alzheimer's-specific pathology. This has profound implications for longevity and healthspan optimization beyond neurodegenerative disease.
The mechanism remains incompletely understood. The biomarker data suggest plasma exchange may work through multiple pathways — reducing inflammation, oxidative stress, and vascular dysfunction while potentially preventing new amyloid deposition rather than just removing existing plaques. The fact that it benefits amyloid-negative patients supports a broader mechanism of action.
The practical reality is complex. Despite promising results, plasma exchange hasn't become standard of care due to regulatory pathways, logistical challenges, limited awareness, and the practical difficulties of implementing a procedural intervention in a field accustomed to pharmaceutical treatments.
The AMBAR study may have stumbled onto something bigger than anyone realized.
Continue the Discussion
This concludes our three-part analysis of the AMBAR study. If you found this series valuable, I'll be publishing additional content exploring practical treatment considerations for patients and families, the anti-aging hypothesis behind AMBAR's benefits, and emerging research on plasma exchange for longevity and cognitive optimization.
Read the Complete Series: