Plasma Exchange and Biological Age: What Two Trials Actually Found

What the procedure is, precisely

Therapeutic plasma exchange is not new, not experimental, and not glamorous in its original setting. Blood is drawn, a centrifuge or membrane filter separates the straw-coloured plasma from the red cells, platelets and white cells, the plasma is discarded, and the cellular fraction is returned to the body alongside a replacement fluid — usually human serum albumin, sometimes with donor immunoglobulin. Citrate (C₆H₅Na₃O₇) keeps the circuit from clotting. It takes two to three hours. Hospitals have used it for decades to treat myasthenia gravis, Guillain–Barré syndrome and thrombotic thrombocytopenic purpura, where the aim is to strip a specific pathogenic antibody out of circulation.

The longevity proposition borrows the machine and changes the target. If ageing is partly driven not by what the body lacks but by what accumulates in circulation — inflammatory cytokines, damaged proteins, senescence-associated signals — then removing a few litres of that milieu and replacing it with clean albumin should, in principle, reset the signalling environment every cell reads.

That is a coherent hypothesis with real experimental roots. It grew out of parabiosis work, in which the circulations of a young and an old mouse are joined, and out of the finding that the benefit does not appear to require anything youthful being added. In 2020, dilution of old mouse plasma with saline and albumin alone — no young blood whatsoever — improved cognition and reduced neuroinflammation in two-year-old mice.^[4] The active ingredient, on this reading, is subtraction. A 2026 review in Experimental & Molecular Medicine frames blood as both the mirror and the modulator of ageing, and places plasma dilution and TPE among the few interventions that act on the systemic milieu rather than on a single tissue.^[7]

The trial the clinics quote

In 2025, a team from the Buck Institute for Research on Aging published the study that anchors nearly every marketing claim now made for the procedure. Forty-two healthy adults over 50 were randomised to biweekly TPE, biweekly TPE with intravenous immunoglobulin (TPE-IVIG), monthly TPE, or placebo, and followed with multi-omics: epigenome, proteome, metabolome, glycome, cytokines, immune cell composition.^[1]

The headline result is genuinely striking. Fifteen epigenetic clocks showed rejuvenation relative to placebo at a false discovery rate below 0.05. The TPE-IVIG arm was the strongest, producing coordinated cellular and molecular responses, reversing aspects of age-related immune decline and shifting proteins associated with chronic inflammation — the same low-grade inflammatory drift measured by the inflammatory age clock. Long-term treatment was reported as safe, with two adverse events requiring discontinuation, one attributable to the immunoglobulin.^[1]

An earlier study from Irina Conboy's laboratory at Berkeley pointed the same way, reporting that rounds of TPE shifted the human proteome toward a younger profile, restored pro-regenerative and apoptotic regulators, reduced markers of cellular senescence and DNA damage in circulating cells, and rebalanced JAK–STAT, MAPK, TGF-β, NF-κB and Toll-like receptor signalling.^[3]

Read on their own, these are the most encouraging human data any systemic longevity intervention has produced. Which is precisely why the second trial matters.

The trial nobody quotes

Two months after the Buck paper, Scientific Reports published a randomised cross-over trial from Charles University in Hradec Králové, with Steve Horvath — the man who built the first widely used epigenetic clock — among the authors. Healthy blood donors underwent either eight or four plasmaphereses over 18 weeks. The critical design difference: this was plasmapheresis without volume replacement by albumin or young plasma.^[2]

The result was not neutral. It was negative. No significant epigenetic rejuvenation was observed on any clock. Instead, the authors documented increases in DNAmGrimAge, the Hannum clock and the Dunedin pace of ageing — three of the better-validated measures in the field. Total cholesterol, non-HDL cholesterol, triglycerides, apolipoprotein A, total protein and albumin all fell; red cell distribution width rose. The paper's own conclusion is unusually blunt for the genre: the protocol "may accelerate epigenetic aging", and more research into its long-term safety is needed.^[2]

The variable is what goes back in

Two randomised trials of what a brochure would call the same procedure, pointing in opposite directions. This is the part worth understanding, because it is not a contradiction so much as a clarification.

The dilution hypothesis was never that removing plasma is therapeutic. It was that replacing a portion of an old, signal-saturated milieu with clean fluid resets the concentration of accumulated factors. Remove plasma and give back albumin, and you have diluted the milieu while restoring oncotic pressure and the body's principal carrier protein. Remove plasma and give back nothing, and you have subjected the body to repeated depletion of albumin, immunoglobulins and lipoproteins without the reset — which looks less like rejuvenation and more like a mild, recurring physiological insult. The clocks appear to have read it as exactly that.

There is a second, less comfortable reading available. Epigenetic clocks are exquisitely sensitive to the composition of the blood cells being sampled, and TPE demonstrably changes immune cell composition. When a procedure alters the cellular mixture a clock is trained on, some of the movement in that clock may be a change in what is being measured rather than in the age of the person. The Buck authors themselves report immune cell composition as one of the domains that shifted.^[1] This is the standing methodological caveat across the whole field of biological age reversal, and it does not disappear because a result is favourable.

The strongest evidence is not about ageing

If you want to see plasma exchange tested at scale with a hard clinical endpoint, you have to leave the longevity literature entirely.

The AMBAR trial randomised 347 patients with mild-to-moderate Alzheimer's disease across four arms: three plasma-exchange regimens with differing doses of albumin and immunoglobulin replacement, and a sham procedure. Over 14 months, treated patients declined 52% less on the activities-of-daily-living scale (P=0.03), with a 66% less decline on the cognitive subscale that fell just short of significance (P=0.06). In the moderate subgroup, both co-primary endpoints were met, with 61% less decline on each.^[5] A later analysis of 142 AMBAR participants found the procedure significantly reduced IFN-γ, eotaxin, MIP-1α and ICAM-1 in serum, and eotaxin-3 and MIP-1β in cerebrospinal fluid — evidence that a purely peripheral intervention reaches the central compartment.^[6]

Note the shape of that evidence. It is a properly powered, sham-controlled trial in a diseased population, using plasma exchange with albumin replacement, and it slowed decline rather than reversing anything. That is a real finding. It is also not the finding being sold to a healthy 52-year-old in a private suite.

What it costs, and what is actually purchased

Reported United States pricing in 2026 runs from roughly $5,000 to $15,000 per session, with typical longevity protocols specifying four to six sessions over two to three months. Business press coverage of the trend has described full courses in the tens of thousands of dollars.^[8] Nothing is reimbursed, because ageing is not an approved indication anywhere.

For that sum, the honest inventory of what is known is short. A small randomised trial shows movement in surrogate markers, strongest when immunoglobulin is added. Another randomised trial with a different replacement protocol shows movement in the wrong direction. A large trial in Alzheimer's disease shows slowed functional decline. No trial in a healthy population has yet measured what any of this does to illness, function or survival, and none has run long enough to try. Between the surrogate and the outcome sits the entire gap that has swallowed most of the longevity interventions of the last two decades — a gap we have traced through senolytics and through the biological-age-reversal claims made for alpha-ketoglutarate, and one that arrives faster than the marketing does.

What a discerning reader should take from this

Plasma exchange is the rare longevity intervention that is not pseudoscience, is not a supplement, and has a mechanistic rationale supported by decades of parabiosis and dilution work.^[9] It is also a procedure with an intravenous line, a cost of a small car, an unapproved indication, and two randomised trials that disagree about its central claim.

The intelligent position is neither dismissal nor enthusiasm. It is precision. If a clinic offers this, the questions that separate a serious protocol from an expensive one are specific: what replacement fluid, at what dose; is immunoglobulin included, given that it drove the strongest result and the only treatment-related discontinuation; who supervises the apheresis; and what outcome, beyond a clock reading sold back to you as an age, is being promised. Any programme that cannot distinguish its own protocol from the one that made GrimAge rise has not read its own evidence.

The broader lesson is the one this field keeps re-teaching. The distance between a plausible mechanism and a proven benefit is measured in properly powered trials with real endpoints, and no amount of marble, discretion or price closes it. There is a version of this procedure that may turn out to matter. Establishing which version, in whom, and to what end, is work that has not been done — and until it is, the person in the chair is paying to be part of the experiment rather than to benefit from its conclusion.

Common questions