Does Plasma Exchange Actually Reverse Biological Age?
If you have read anything about therapeutic plasma exchange in the last year, you have probably encountered one number: 2.61 years of biological age, removed. It comes from a real randomised, placebo-controlled trial, published in a real peer-reviewed journal, and it has travelled a long way.
It is also incomplete. A second randomised trial of plasma removal, published the same year with epigenetic analysis co-authored by Steve Horvath — the researcher who built the first widely used epigenetic clock — found no rejuvenation at all, and on several clocks, movement in the opposite direction. Both trials are honest. The reason they disagree is the most clinically useful finding in this literature, and almost nobody is writing about it.
What plasma exchange actually does
Aging is partly a story of accumulation. Over decades, your plasma — the liquid your blood cells travel in — collects inflammatory cytokines, autoantibodies, oxidised lipoproteins and the signalling proteins of chronic inflammation. The cells themselves age. But so does the medium they sit in.
Therapeutic plasma exchange (TPE) addresses the medium directly. A centrifuge separates plasma from blood cells, removes roughly one to one and a half plasma volumes, and returns the cells suspended in fresh replacement fluid — typically human albumin. It takes two to three hours. It is not experimental technology: TPE has been standard care in neurology, haematology and nephrology for decades.
The longevity question is newer. It comes out of heterochronic parabiosis — the mouse experiments in which joining the circulation of a young animal to an old one improved the old animal's tissues. The insight developed by Irina and Michael Conboy at Berkeley was that the benefit may owe less to something youthful being added than to something aged being diluted. In Irina Conboy's words: "Young blood or factors are not needed for the rejuvenating effect; dilution of old blood is sufficient."
That reframing matters, because dilution is something you can do to a human being with equipment that already exists.
The trial everyone cites
Fuentealba, Kiprov, Schneider et al., Aging Cell, 2025. Buck Institute for Research on Aging, with Circulate Health. Registered as NCT06534450.[1]
A single-blind, placebo-controlled trial in 42 adults aged 50 and above — about ten per arm — across four groups:
| Arm | Protocol |
|---|---|
| TPE + IVIG | Two sessions in the first week of each month for three months, with 2 g intravenous immunoglobulin immediately after each exchange |
| TPE “biweekly” | Two sessions in the first week of each month, then three weeks off, for three months |
| TPE monthly | One session per month for six months |
| Sham | Simulated procedure without fluid replacement, on matching schedules |
(The authors' "biweekly" label means twice in the first week of a month — not every two weeks.)
Rather than relying on a single epigenetic clock, the team measured across the epigenome, proteome, metabolome, glycome, immune cytokines and immune cell composition, plus physical measures including balance and strength.
- TPE with IVIG: an average reduction of 2.61 years in multi-omic biological age (p = 1.1 × 10⁻⁵) — the largest effect of any arm
- The monthly TPE regimen: an average reduction of 1.32 years (p = 5.7 × 10⁻³). Notably, the monthly schedule outperformed the more intensive non-IVIG schedule
- Recipients showed shifts in immune cell composition consistent with partial reversal of age-related immune decline, and modulation of proteins linked to chronic inflammation
- The largest benefits appeared in participants whose baseline bilirubin, glucose and liver enzymes were most elevated — that is, those in poorer metabolic health to begin with
- Effects were strongest after the first three sessions, with diminishing returns after that
Three caveats belong in the same breath as the headline. This is a small trial — about ten people per arm — and small trials produce unstable effect estimates. Randomisation was by entry date rather than concealed random allocation. And it was supported by Circulate, Inc., a company that sells plasma exchange; Dobri Kiprov, Eric Verdin and David Furman are all members of Circulate and all authors on the paper.
None of that makes the finding wrong. It does mean 2.61 years is a first result awaiting independent replication, not a settled number.
Change in biological age by treatment arm (Aging Cell, 2025; n=42)
Average reduction in multi-omic biological age versus placebo. Only these two arms had per-arm effect sizes reported. Data: Fuentealba et al., Aging Cell (2025).
The trial almost nobody cites
Borsky et al., Scientific Reports, 2025. Charles University, Hradec Králové, with epigenetic analysis co-authored by Steve Horvath.[2]
A cross-over study in 38 healthy first-time donors aged 40 to 60, stratified by age, sex and BMI, running 18 weeks with up to eight sessions at a minimum 14-day interval. Thirty-four completed. Aging was assessed with a full established panel: Horvath pan-tissue and Skin & Blood, PhenoAge, GrimAge and GrimAge2, Hannum, DunedinPACE and DNAmTL.
- Total cholesterol, non-HDL cholesterol and triglycerides all fell significantly
- Apolipoprotein A also fell — which is not an improvement, since Apo A-I is the cardioprotective, HDL-associated apolipoprotein
- "No significant epigenetic rejuvenation was observed based on epigenetic clock measurements. Instead, plasmapheresis was associated with increases in DNAmGrimAge, the Hannum clock, and the Dunedin Pace of Aging."
- Calcium and albumin fell; homocysteine and red cell distribution width rose
- No participant stopped the study because of severe long-term effects or illness caused by the procedure, though one discontinued during a first session for low blood pressure
The authors' own conclusion is worth quoting rather than paraphrasing: the protocol "has not provided conclusive data supporting benefits," and "based on epigenetic clock parameters, it may accelerate epigenetic aging."
Why the two disagree — and why it matters clinically
The instinct is to treat this as one trial contradicting another. It isn't, and the distinction is the practical heart of this entire topic.
The Buck trial performed therapeutic plasma exchange. The Czech trial performed plasmapheresis for plasma donation. The procedures look similar and are not the same thing.
In therapeutic plasma exchange, removed plasma is replaced — volume for volume — with human albumin. In the Czech protocol, donors received physiological saline only: 250 ml when 750 ml or more of plasma was collected. No albumin went back in.
The authors themselves name this. Their stated recommendation for future work is to "consider prolonging the intervals between each plasmapheresis and to substitute for vital parts of plasma that are being diminished by the procedure, such as albumin," and they call for "further clinical studies incorporating albumin replacement."
So there are two candidate explanations, and the authors put both on the table: the missing albumin, and the session frequency. Their protocol ran eight sessions at fortnightly intervals; the best-performing non-IVIG arm in the Buck trial ran six sessions across six months. Falling calcium and albumin in the Czech cohort is exactly what you would predict from repeatedly removing a protein-rich fluid without putting protein back.
The honest reading is therefore not "plasma exchange doesn't work." It is narrower and more useful:
"What appears to matter is not only what you take out, but what you put back — and how often you do it."
If that is right, the replacement fluid isn't an incidental detail of the procedure. It may be part of the active ingredient. And it means results from a plasma donation centre tell you very little about a clinical exchange protocol, in either direction.
The biggest trial in this field isn't about biological age at all
Lost in the biological-age conversation is much the biggest controlled trial of plasma exchange with albumin ever run.
The AMBAR trial — Alzheimer's Management By Albumin Replacement — screened 496 patients aged 55 to 85 with mild-to-moderate Alzheimer's disease and randomised 347 to plasma exchange with albumin replacement, at varying doses with and without immunoglobulin, versus sham. Treatment ran 14 months: six weeks of weekly conventional exchange, then twelve months of monthly low-volume exchange.[3]
Two findings worth separating carefully, because they are routinely conflated:
- In the moderate subgroup, treated patients showed 61% less decline on both ADCS-ADL (daily function) and ADAS-Cog (cognition) versus sham.
- Across all treated patients, decline on the CDR-Sb scale was 65–71% lower than sham.
AMBAR matters here for two reasons. It is an order of magnitude larger than the biological-age trials. And its endpoints were cognition and daily function — things patients actually experience — rather than a molecular estimate of age.
It is not a longevity trial. It was run in a disease population, and its results do not transfer automatically to healthy aging. But it is the most substantial evidence that removing and replacing plasma changes a clinical outcome in humans.
AMBAR: reduction in decline versus sham (n=347 randomised)
The first two bars are the moderate-severity subgroup; the third is the full treated population, where the range across active arms was 65–71%. Data: Boada et al., Alzheimer's & Dementia (2020).
What "biological age" actually means here
This deserves saying directly, because it is where most coverage of this topic quietly overstates the case.
Epigenetic and multi-omic clocks are surrogate endpoints. They are statistical models trained to predict chronological age, mortality risk or pace of aging from molecular patterns. They are genuinely useful research instruments. They are not a measurement of how long you will live.
- Different clocks disagree. The Czech trial is a clean illustration: several established clocks moved in a direction that contradicted the multi-omic estimate from the other trial.
- Moving a clock is not the same as moving an outcome. No trial has yet shown that lowering a biological age estimate with plasma exchange lengthens life or delays disease.
- The effect looks front-loaded and may not persist. The Buck trial's own finding — most of the benefit in the first three sessions, diminishing returns after — is not what an indefinitely repeatable intervention looks like. Nobody has published durable follow-up.
The intellectually honest position in August 2026 is this: TPE has a plausible mechanism, one supportive small randomised trial with an industry sponsor, one contradictory randomised trial run under a materially different protocol, and one large trial with real clinical endpoints in a disease population. That is a genuinely promising evidence base. It is not a proven longevity intervention, and anyone telling you otherwise is selling something.
If you are considering it
TPE is a medical procedure, not a wellness service. Practical points worth knowing:
- A session runs two to three hours, with IV access and continuous monitoring.
- The best-supported protocols are courses, not single sessions — and the Buck data suggest the first three sessions carry most of the effect.
- Citrate anticoagulation lowers calcium. This is the most common adverse effect of any apheresis procedure and the usual cause of the tingling around the mouth and fingers, or light-headedness, some people feel mid-session. It is monitored and managed, not ignored.
- Repeated removal without adequate replacement lowers albumin. This is precisely the difference between a clinical protocol and a donation protocol.
- It is not right for everyone. Active infection, coagulopathy, haemodynamic instability and some cardiac conditions are contraindications or require careful risk–benefit assessment before proceeding.
- Pre and post biomarker panels are the point. Without them you have no way of knowing whether anything happened.
We deliver TPE at Peak Human in a clinical, biomarker-tracked setting — you can read how the protocol is structured on our therapeutic plasma exchange service page, see the intervention summarised in the Peak Human intelligence database, or browse the underlying TPE research.
- The headline figure — 2.61 years, from the TPE + IVIG arm — comes from one small (n=42), industry-supported randomised trial. Real, peer-reviewed, and awaiting independent replication.
- A second 2025 randomised trial found no epigenetic rejuvenation, and possible acceleration — but it was plasma donation without albumin replacement, at fortnightly intervals.
- The replacement fluid may be part of the active ingredient. Take-out-and-put-back is a different intervention from take-out.
- The largest controlled dataset, AMBAR (347 randomised), used albumin replacement and moved real clinical endpoints — in Alzheimer's disease, not healthy aging.
- Biological age clocks are surrogates. No trial has yet linked TPE to longer life or delayed disease.
- Benefit appears front-loaded in the first three sessions, with no published evidence of durability.
Dr. Sanjeev Goel will review your goals, your bloodwork, and whether TPE is appropriate for you.
Disclosure: Peak Human offers therapeutic plasma exchange as a clinical service. I have tried to write this the way I would want it written if I did not — which is why the contradictory trial is in here, and why the funding source and author affiliations of the favourable trial are named. If TPE is right for you, the evidence above should be why. If it isn't, that should be clear from the same evidence.
This article is for general education and is not medical advice. The findings described come from studies in specific populations and do not predict individual results; several are small or preliminary. Therapeutic plasma exchange is a medical procedure with real risks and is not appropriate for everyone. Please discuss it with a qualified physician who can review your history and bloodwork.
- Fuentealba M, Kiprov D, Schneider K, et al. Multi-Omics Analysis Reveals Biomarkers That Contribute to Biological Age Rejuvenation in Response to Single-Blinded Randomized Placebo-Controlled Therapeutic Plasma Exchange. Aging Cell. 2025;24(8):e70103. PubMed ↗ · DOI ↗
- Borsky P, Brooke RT, Milciute M, Gordevicius J, Horvath S, et al. Human clinical trial of plasmapheresis effects on biomarkers of aging (efficacy and safety trial). Scientific Reports. 2025. DOI ↗
- Boada M, López OL, Olazarán J, et al. A randomized, controlled clinical trial of plasma exchange with albumin replacement for Alzheimer's disease: Primary results of the AMBAR Study. Alzheimer's & Dementia. 2020;16(10):1412–1425. PubMed ↗ · DOI ↗
- Kim D, Kiprov DD, Luellen C, et al. Old plasma dilution reduces human biological age: a clinical study. GeroScience. 2022;44(6):2701–2720. PubMed ↗ · DOI ↗
Literature retrieved from PubMed and journal sources, August 2026.