Ex-Vivo Organ Repair

Established Machine perfusion connects a donor organ to a circuit that pumps oxygenated fluid through its vasculature, and in doing so converts a countdown into a window. On ice, a liver is a depleting asset with roughly eight to twelve usable hours, a heart or lung four to six, a kidney around a day. On a pump the organ is metabolically active and observable: it makes bile, clears lactate, consumes oxygen, holds a vascular resistance. Three temperature regimes are in clinical use — hypothermic oxygenated perfusion at around 8 to 12 °C, subnormothermic, and normothermic perfusion at body temperature — and they trade metabolic suppression against the ability to test the organ while it works.

Established The claim that defines this subject is that a window is also an opportunity for treatment, and that claim is at a much earlier stage than the preservation claim. An organ on a circuit can be defatted, flushed of microthrombi, dosed with drugs at concentrations no patient could tolerate, stripped of donor immune cells, enzymatically stripped of its blood-group antigens, or given genetic instructions — all without exposing a recipient to any of it. Each of those has been demonstrated at least once. None has a randomised trial showing that the treated organ works better in a person than the same organ untreated.

Frontier The measured clinical benefit so far comes overwhelmingly from assessment and better preservation, not repair. Perfusion lets a surgeon look at an organ that would otherwise have been declined on paper, which lowers discard rates and raises utilisation. That is a real gain and it is a different mechanism from repair, with different limits: selection can only recover organs that were already good enough, while repair would extend the pool. Separating the two is the central unfinished measurement in the field.

Frontier The longest reach of the field is a liver transplanted after three days on a machine, and the banking horizon beyond it is a rat result. Multi-day normothermic perfusion of a human liver and its successful transplantation is established as a demonstration; vitrification with nanowarming is established in a rat kidney at 100 days and has never been done at human organ scale. Between those two sits the honest answer: days are real, weeks are experimental, and indefinite banking is not close.

Established A note on sourcing. This brief was commissioned in September 2026 from the Institute’s research base. Reading-list entries without links are cited from the bibliographic record rather than re-fetched, and claims are dated no later than early 2026 unless carried by a linked source.

Established The problem being solved is ischaemia-reperfusion injury, and the ice box was never a good solution to it. Cold storage slows metabolism without stopping it, accumulates succinate and other substrates, and produces a burst of reactive oxygen species at reperfusion that injures endothelium and, in the liver, the biliary tree. Donation after circulatory death organs, which have already suffered warm ischaemia before recovery, are the worst affected and are also the fastest-growing part of the donor pool.

Established The founding randomised result is the European liver normothermic-perfusion trial reported in 2018, and its numbers still anchor the field. Across 220 transplants, normothermic machine perfusion roughly halved peak aspartate aminotransferase in the first week against static cold storage, cut early allograft dysfunction to about a third of the control rate, extended mean preservation time by roughly half, and approximately halved the rate at which recovered livers were discarded. It did not improve one-year graft survival. That combination — less injury, more organs used, no survival difference — is the honest shape of the whole field’s evidence, and it is mostly about supply rather than outcome.

Established The cleanest outcome benefit is hypothermic oxygenated perfusion for livers donated after circulatory death. In a randomised trial of 160 recipients published in 2021, symptomatic non-anastomotic biliary strictures — the injury that ruins DCD liver transplantation and often requires re-transplantation — occurred in about 6 per cent of the perfused arm against about 18 per cent of the cold-stored arm. This is the strongest evidence anywhere that a period on a machine changes what happens to a patient, and the mechanism is protective preconditioning, not repair.

Established The most important negative result is in kidneys, and it should be quoted whenever perfusion is described as self-evidently good. A randomised trial of 338 donation-after-circulatory-death kidney transplants, reported in 2023, applied one hour of normothermic perfusion at the end of cold storage and found no reduction in delayed graft function — rates around 60 per cent in both arms. Hypothermic machine perfusion of kidneys, by contrast, has been standard in much of Europe since a randomised trial in 2009 (Moers and colleagues) showed less delayed graft function than ice. The lesson is specific: perfusion is not a blessing conferred by contact with a pump. Duration, temperature and timing are the intervention, and a short normothermic pulse after cold storage does nothing.

Frontier Utilisation is where the field’s economic case is made, and attribution is contested. Perfusion supplies decision-grade numbers — lactate clearance, bile pH and glucose, vascular resistance, oxygen uptake — that let a surgeon accept an organ a paper profile would have condemned. United States transplant volumes reached successive records through the mid-2020s while portable normothermic devices spread and the DCD share of liver and heart transplants rose sharply. Allocation policy changed over the same period, organ procurement organisations were placed under new performance metrics, and donor demographics shifted. Perfusion is plainly part of the rise and the field cannot currently say what fraction it owns.

Established The scarcity being addressed is not rhetorical. The United States waiting list runs to roughly 103,000 people, about seventeen of whom die each day, while about a quarter of recovered kidneys are never transplanted. A method that converts a declined organ into a transplanted one is competing against a discard rate, not against a technology.

Frontier Multi-day perfusion is the field’s furthest reach and rests on one programme. The Zurich group reported in 2020 a machine that kept injured human livers perfused and functioning for one week, using automated control of glucose, hormones, haematocrit and waste that a human operator cannot sustain. In 2022 the same programme perfused a declined human liver for three days, treated it during that time, transplanted it, and reported the recipient well at one year on standard immunosuppression. A small series has followed rather than a trial. This is a demonstration of duration, and duration is the precondition of every repair that takes longer than an operation.

Frontier Repair on the circuit has a growing list of single demonstrations and no randomised evidence. The most complete is enzymatic blood-group conversion: human donor lungs perfused ex vivo with bacterial glycosidases lost essentially all detectable A antigen within hours, and the treated lungs did not show hyperacute injury when challenged with incompatible plasma in an ex vivo model. Defatting cocktails reduce triglyceride content in steatotic human livers over hours; thrombolytic flushes clear the peribiliary microthrombi associated with ischaemic cholangiopathy; mesenchymal stromal cells and anti-inflammatory agents have been delivered to lungs during perfusion. Every one of these is at the demonstration stage, and the step from demonstration to a treated organ in a randomised recipient has not been taken for any of them.

Established Gene delivery during perfusion is seventeen years old and has not moved to the clinic. Adenoviral delivery of interleukin-10 to human donor lungs during ex vivo perfusion, reported in 2009, improved lung function and reduced inflammatory injury — the founding result of the field, produced in the same Toronto programme that made ex vivo lung perfusion clinical. It remains, in substance, the founding result: no ex vivo organ gene therapy is licensed anywhere. The vector landscape has since changed completely, with lipid nanoparticle delivery of messenger RNA now a mature clinical modality, and an isolated organ on a circuit is the most favourable delivery environment in medicine — closed, dosed at will, with no systemic exposure and no immune response to the vector from a recipient who has not received it yet.

Frontier The same window is what makes perfusion the enabling technology for organs that are not human. Extracorporeal cross-circulation of gene-edited pig livers with brain-dead human decedents has been used to test xenogeneic organ function outside a recipient, and decedent studies of pig-to-human kidney transplantation have become the standard human testbed for interventions too risky to try first in a patient. Perfusion circuits, decedent protocols and transplant immunology are converging into one experimental platform, and this subject supplies its plumbing.

Frontier Banking is the field’s most oversold horizon and its evidence is precise. Human livers have been supercooled to about −4 °C for 27 hours and recovered on normothermic perfusion, with no transplant attempted. Rat livers have survived partial freezing and subzero storage for days with transplant survival. The strongest result is vitrification with nanowarming: rat kidneys vitrified, stored at about −150 °C for up to 100 days, rewarmed by iron-oxide nanoparticles driven by an alternating magnetic field, then transplanted, with all five recipients surviving and renal function recovering within weeks. The prior benchmark was a rabbit kidney vitrified to about −135 °C and transplanted in 2005.

Established No human organ has been vitrified and transplanted, and the reason is scale, not principle. A human kidney is on the order of a hundred times the mass of a rat kidney and a liver a thousand times; cryoprotectant must be loaded uniformly through the vasculature without toxicity, nanoparticles must be distributed evenly enough that rewarming is uniform, and thermal gradients that a rodent organ tolerates will crack a human one. Each is a measurable engineering problem and none has been solved at human scale.

Frontier The economics are set by reimbursement rules as much as by biology. In the United States the cost of recovering and preparing an organ flows through the transplant centre’s organ acquisition cost centre and is reimbursed, which means a perfusion run costing tens of thousands of dollars per case does not sit on a surgeon’s departmental budget. That is the single clearest reason adoption has been faster in the United States than in systems that must find the money inside a fixed hospital budget, and the leading device vendor’s revenue has grown several-fold since 2022 on its own reporting. Against that cost stands the alternative: a dialysis patient costs a health system roughly six figures a year indefinitely, so a kidney recovered from the discard pile repays a great many perfusion runs.

Frontier Does perfusion repair organs, or only identify the good ones? Every utilisation gain published so far is consistent with pure selection. The question is answerable with a trial that holds the organ pool fixed and varies only the intervention, and it has not been run at power.

Frontier How long does a repair need? Defatting and thrombolysis work in hours; regeneration of biliary epithelium, clearance of donor passenger leukocytes or expression of a delivered transgene plausibly need days. If the useful repairs are multi-day, the field’s clinical infrastructure — built around transport times — is the wrong shape for its own biology.

Frontier What viability biomarker actually predicts graft function? Perfusion-derived criteria vary between centres, and a false negative discards a usable organ while a false positive transplants a bad one. No biomarker has been prospectively validated across centres against one-year outcomes, and this is the gap most responsible for variation in who accepts what.

Frontier Can immunological tolerance be started on the pump? Depleting donor antigen-presenting cells, delivering immunomodulators, or altering the endothelium before the organ meets a recipient is the most consequential possible use of the window, because it attacks lifelong immunosuppression rather than early graft injury. Evidence is preclinical and thin.

Established Multi-day perfusion is a sterility and homeostasis problem before it is a biology problem. A circuit running for days must control glucose, electrolytes, hormones, pH, haematocrit and waste without an operator at the bedside, avoid haemolysis in the pump, and stay sterile through repeated interventions. The one programme that has done it built automated control to get there, which is why the capability has not spread by imitation.

Frontier Perfusate supply is a constraint nobody plans for. Normothermic circuits are commonly primed with packed red cells, which places elective organ repair in competition with a blood supply that is already strained. Acellular oxygen carriers would remove that dependency and have a long history of clinical disappointment.

Established The biliary tree is the liver’s weak point and it is poorly instrumented. Cholangiocytes are supplied by a fragile peribiliary plexus, injury there produces strictures months later, and no ex vivo measurement predicts that outcome reliably. The best current proxies — bile pH, bicarbonate and glucose — are indirect.

Frontier Cryoprotectant loading and nanoparticle distribution do not scale by making the machine bigger. Uniform perfusion of a cryoprotectant through a human-sized vascular tree at subzero temperatures, with viscosity rising as the solution cools, is an unsolved fluid-mechanics and toxicity problem, and rewarming uniformity depends on nanoparticles reaching every capillary bed.

Frontier A treated organ has no settled regulatory category. An organ dosed with a gene therapy during perfusion is arguably a biologic product, arguably a transplanted organ regulated through transplant oversight, and arguably both. The uncertainty is itself a bottleneck: sponsors will not fund a trial whose product definition is unknown.

Frontier This subject waits on delivery vehicles it does not build. Lipid nanoparticle formulation, targeting ligands and messenger RNA chemistry are advancing for systemic medicine, and an isolated organ is the easiest target any of them will ever have. The transfer is not automatic: an organ perfused at 10 °C will not translate messenger RNA efficiently, so the delivery window and the preservation window are in tension.

Established It waits on cryobiology, which is a small field. Ice-blocking molecules, low-toxicity cryoprotectant cocktails and magnetic nanowarming are the work of a handful of laboratories, and the rate-limiting inputs to organ banking are being produced by fewer people than a single large drug programme employs.

Frontier It waits on transplant immunology for the endpoint that would justify the machine. If ex vivo conditioning could reduce immunosuppression, the value of perfusion would stop being measured in discard rates and start being measured in decades of recipient health. That is an immunology result, not a device result.

Established It does not wait on organ manufacturing, and organ manufacturing waits on it. As Lab-Grown Organs sets out, no engineered construct has an endothelialised vascular tree that survives whole blood at physiological pressure. Any such construct would be tested, conditioned and kept alive on exactly this equipment, which makes perfusion infrastructure a precondition of that field rather than a competitor to it.

Frontier The decisive experiment is the paired-kidney repair trial, and the donor supplies the randomisation for free. A deceased donor gives two kidneys of near-identical biology and history. Assign one to perfusion plus the candidate intervention and the other to perfusion alone, transplant both, and report delayed graft function and one-year function. That design removes donor variation, the largest confounder in the field, and it isolates repair from selection — the one thing utilisation statistics can never do. It requires no new device and no new drug for the first candidates, and nobody has run it at power.

Frontier Second: randomise organs that every centre has already declined. The comparator for an extended perfusion-and-repair protocol on an organ bound for the incinerator is not standard care; it is discard. A registry-controlled trial of multi-day perfusion on universally declined organs, with one-year graft survival as the endpoint, would establish whether the pool can be extended rather than merely sorted, and the ethical review is simpler than it looks because the alternative for the organ is destruction.

Frontier Third: prospective cross-centre validation of one viability biomarker. Fix a threshold in advance, transplant a defined set of organs that fail it, and report outcomes. Until someone transplants organs their own criteria reject, the criteria are unfalsified, and the field will keep discarding organs on numbers whose predictive value is assumed.

Speculative Fourth: the first-in-human ex vivo gene delivery trial, seventeen years late. The 2009 lung result defined a protocol; a modern vector, a transient anti-inflammatory payload and a randomised design would finally test whether the closed-circuit delivery advantage translates into a recipient benefit.

Speculative Fifth: the human-scale vitrification attempt, staged honestly. Vitrify and rewarm a human kidney rejected for transplant, assess it on normothermic perfusion rather than in a recipient, and publish the damage. A negative result with a measured injury profile would be more useful to the field than another rodent success.

Established The device requirements split cleanly by mission. A transport machine must be portable, battery-powered, robust to aircraft handling and simple enough for a recovery team; a repair machine must run for days, control composition automatically, allow sterile intervention, and instrument the organ continuously. These are different products, and most of the installed base is the first kind.

Frontier Perfusate and gas exchange set the ceiling on duration. Red-cell-based perfusates haemolyse over days in roller and centrifugal pumps; oxygenators foul; waste accumulates unless dialysis is integrated. The one-week human liver result came from solving these as a systems problem rather than by improving any single component.

Frontier Instrumentation is the cheapest available improvement. Continuous in-line measurement of oxygen consumption, lactate, bile chemistry and perfusate cytokines would convert perfusion from an interval into a dataset, and would make the cross-centre biomarker validation described above straightforward instead of laborious.

Frontier Nanowarming hardware is the least mature piece of the banking chain. Driving iron-oxide nanoparticles with an alternating magnetic field at rat-kidney scale is a benchtop apparatus; doing it uniformly across a human liver requires coil geometry, field homogeneity and thermal monitoring that has not been demonstrated.

Frontier Xenotransplantation is the nearest neighbour and the most demanding customer. A gene-edited pig organ still has to be recovered, preserved, assessed and possibly conditioned, and decedent studies in brain-dead recipients have become the standard human testbed for both fields at once. The equipment, the immunological questions and the ethical review are shared.

Established Organ manufacturing needs this infrastructure before it needs anything else. Lab-Grown Organs and Whole Organ Regeneration both stop at a construct that has never been challenged with whole blood at pressure; the challenge rig is a perfusion circuit, and the maturation bioreactor for any recellularised scaffold is the same machine described here.

Frontier A perfused human organ is also a research platform, and that use is growing for regulatory reasons. Declined human organs on pumps are among the few human systems where a drug can be dosed at organ scale with real vasculature, which connects this subject to Microphysiological Systems and to the regulatory push for human-relevant models that the FDA’s animal-testing roadmap represents.

Speculative Cold biology connects it to hibernation and preservation research. Emergency preservation and resuscitation — cooling a trauma patient to profound hypothermia to buy surgical time — and the metabolic suppression work in Human Hibernation are the whole-body version of the same bet, and the banking question links directly to the cryobiology discussed in Cryonics, where the same vitrification results are read for a different purpose.

Frontier The regulatory gap is the treated organ. Devices are regulated as devices, organs are allocated under transplant oversight, and an organ that has been dosed with a biologic during perfusion falls between the two. Until a regulator states which pathway applies, ex vivo repair has no product definition, and without a product definition there is no sponsor.

Frontier Allocation rules were written for organs that could not wait. If perfusion reliably buys a day, geography stops constraining matching, elective scheduling becomes possible, and better immunological matching becomes affordable in time. None of that is in current allocation policy, which optimises for speed because speed used to be the binding constraint.

Frontier Normothermic regional perfusion has opened a genuine institutional dispute. Restoring circulation in situ in a donor after circulatory death, with the cerebral vessels clamped, recovers better organs and has been objected to by professional bodies on the grounds that it restarts circulation in a person declared dead by circulatory criteria. The practice is in routine use in some jurisdictions and prohibited or contested in others. That disagreement is about the definition of death, and it sits inside the logistics of this field.

Frontier Consent for a repaired organ has no established form. A recipient can consent to receive a treated organ; the donor family consented to donation, not to an experimental intervention on their relative’s liver. Where the treatment is a gene therapy, the resulting graft carries the modification into the recipient, and neither consent document currently anticipates that.

Established Decedent research is now a standard route and deserves its own scrutiny. Studies in brain-dead human decedents, used for xenograft physiology and for extracorporeal circuits, sit outside ordinary clinical trial protections because the subject has died. They are defensible and they depend entirely on family consent, institutional review and public trust that is easier to lose than to build.

Frontier Equity cuts in an unexpected direction. Recovered marginal organs are disproportionately offered to patients who are sickest or least able to wait, which means the first beneficiaries of perfusion are often those with the least negotiating power. Whether a recovered organ is a gift or a second-best offer depends on outcome data that do not yet exist for many of these grafts.

Speculative Banking would change the ethics rather than removing them. A bank implies ownership, pricing, prioritisation and expiry, and organ allocation systems have no vocabulary for an asset that can be stored. The ethical work here is best done before the technology arrives, and it is barely begun.

Established The near-term prize is measured in organs, not in decades of life extension. Roughly a quarter of recovered kidneys are discarded and thousands of people die waiting each year in the United States alone; halving discard would be one of the larger single gains available in transplant medicine, and it requires no new biology.

Speculative Banking would restructure transplantation the way refrigeration restructured food. Stored organs allow elective scheduling, immunological matching across a national pool rather than a regional one, pre-transplant conditioning of the recipient, and quality control before use. The gain is logistical and it is enormous; the evidence for it at human scale is a rat kidney and a rabbit kidney.

Speculative Ex vivo conditioning is the plausible route to less immunosuppression. Lifelong immunosuppression costs recipients kidney function, infection risk and cancer risk; an intervention applied to the organ rather than the patient is the only approach that does not trade one systemic drug for another. This is the largest claim in the brief and it rests on preclinical work.

Handwave The claim that ex vivo repair ends the organ shortage belongs here as a claim. The shortage is set by the number of people who die in circumstances allowing donation, and no amount of repair changes that denominator; recovering discarded organs narrows the gap and does not close it.

These horizons track results that would change what can honestly be written, not forecasts of clinical availability.

• 10 yr: Frontier A randomised trial separating repair from selection — most cheaply the paired-kidney design — is the defining near-term result, and its absence would say that the field prefers utilisation statistics to evidence. Frontier Multi-day perfusion either spreads beyond the one programme that has done it or remains a Zurich capability. Speculative A first-in-human ex vivo gene delivery trial is technically ready and needs a regulatory category. Established Machine perfusion continues to displace ice for marginal organs regardless, because reimbursement already rewards it.
• 25 yr: Speculative A human organ is vitrified, rewarmed and transplanted, or the banking programme concedes that supercooling and partial freezing at days-to-weeks is the achievable endpoint. Speculative Ex vivo immunological conditioning enters trials with immunosuppression reduction as the endpoint. Speculative Allocation policy is rewritten around organs that can wait, which is a bigger institutional change than the technology that caused it.
• 50 yr: Speculative Organ banks operate as routine infrastructure, making transplantation elective and matched, on the assumption that human-scale vitrification is solved. Handwave Whole-body or near-whole-body preservation for medical purposes is asserted on this horizon by parties whose interest depends on it, and rests on solving every organ simultaneously rather than one at a time.
• 100 / 250+ yr: Handwave Indefinite storage and revival of complex organs and organisms is a statement about cryobiology nobody can currently evaluate, and the honest position is that the rat kidney is the state of the art and the gap to a human is measured in unsolved engineering, not in years.

• Depends on results other briefs are producing for their own reasons: the vascular and endothelial work tracked in Lab-Grown Organs, whose named binding barrier — an organ-scale vascular tree that survives whole blood at pressure — is the same barrier a perfusion circuit tests, and the cold-biology results read differently in Cryonics and Human Hibernation. Nothing on the map blocks the clinical use of perfusion itself, which is already routine.
• Requires (not on this map) first, a biomarker measured on the circuit that has been prospectively validated against one-year graft function across centres, without which every acceptance decision is a local custom. Second, a regulator’s statement of what a treated organ is, since no sponsor will fund a trial of a product with no category. Third, uniform delivery of cryoprotectant and warming nanoparticles through a human-sized vascular tree, the specific unsolved step between the rat result and an organ bank. Fourth, an oxygen carrier that does not consume the blood supply, because elective multi-day perfusion at scale competes with transfusion medicine. Fifth, a payment rule that rewards recovering an organ that would have been discarded rather than penalising the outcome statistics of the centre that tries. Sixth, a circuit that holds sterility and homeostasis for days under repeated intervention, which one programme has built and nobody has productised.
• Enables lower discard rates and higher utilisation of donation-after-circulatory-death organs; organ-level drug and gene delivery without systemic exposure; conditioning aimed at reducing lifelong immunosuppression; the assessment and maturation rig that engineered and xenogeneic organs will require; and, if banking is solved, elective and better-matched transplantation.
• Adjacent to Whole Organ Regeneration, which asks whether an organ can repair itself in the body while this brief asks what can be done to it outside one, and to Microphysiological Systems, which shares the argument that perfused human tissue is the most human-relevant test system available.

Frontier “Machine perfusion repairs damaged organs.” Almost everything it has been shown to do clinically is preservation and assessment. The randomised evidence supports less reperfusion injury, fewer biliary strictures in donation-after-circulatory-death livers, and fewer discards. Defatting, thrombolysis, cell therapy and blood-group conversion are demonstrations, not treatments, and the word repair is doing promotional work in most coverage of the field.

Established “Perfusion always helps.” A randomised trial of one hour of normothermic perfusion in kidneys found no benefit at all. Delayed graft function was around 60 per cent in both arms of a 338-transplant trial. Temperature, duration and timing are the intervention; contact with a pump is not.

Frontier “The three-day liver means organs can now be stored for days.” It means one organ was kept metabolically alive for three days by a machine that few centres can run. Normothermic multi-day perfusion consumes oxygen, blood products, staff attention and sterile handling continuously. It is life support, not storage, and the distinction is the entire difference between a transport window and a bank.

Established “Vitrification is solved — they transplanted a frozen kidney.” They transplanted rat kidneys, five of five, after up to 100 days. The rabbit kidney result that preceded it dates from 2005. The human problem is mass, viscosity, thermal gradients and nanoparticle distribution, and no human organ has been vitrified and transplanted.

Frontier “Supercooling already gives weeks of preservation.” In human livers it has given 27 hours, with no transplant. The multi-day subzero results are in rats. Scaling subzero storage is hard for the same reason vitrification is: ice nucleation probability rises with volume and time.

Established “Ex vivo gene therapy for organs is around the corner.” It has been around the corner since 2009. Interleukin-10 delivery to human donor lungs during perfusion worked in that study, and no ex vivo organ gene therapy has been licensed since. The obstacle has not been the vector or the target; it has been that no sponsor knows what kind of product a treated organ is.

Frontier “Restarting circulation in a declared-dead donor is a technical detail.” Professional bodies disagree, in public, and jurisdictions differ. Normothermic regional perfusion improves organ quality and is opposed by some clinicians precisely because it restores circulation in a donor declared dead on circulatory criteria. Any honest account of the field’s utilisation gains has to state that some of them come from a contested practice.