Replantation

“Yes, if it is technically feasible and the patient is fit for it — and I would say that the traditional teaching against single-digit replantation has been overturned.

My decision rests on four things: whether the patient is stable and this is an isolated injury; the mechanism; the level; and what this patient needs the hand to do.

The evidence has moved. The contraindication to single-digit replantation comes from two papers from 1977, by Morrison and by Weiland, and Sebastin and Chung point out explicitly that neither presents data to support that objection. It was opinion. *The 2019 FRANCHISE multicentre study in JAMA Surgery, 338 patients across 19 centres, now recommends replantation for single-finger amputation excluding the thumb distal to the PIP joint*, because it achieved better patient-reported outcomes.

The honest caveat is that FRANCHISE also found revision amputation gave better two-point discrimination in that same subgroup — 6 versus 8 millimetres. So what replantation buys is length, appearance, the nail and the patient’s own digit, and I would frame the conversation that way rather than promising better function.”

“I would rather frame it as who benefits, because the classical list is not evidence-based.

The indications I would still give are: the thumb at any level; multiple digits; any digit in a child; transmetacarpal, mid-palm and wrist-level amputations; and a single digit distal to the FDS insertion. The last of those is the one with an actual evidential anchor — Urbaniak’s 1985 series of 59 single-finger replants, which found replantation distal to the FDS insertion justified and proximal to it seldom indicated.

What I would add is that the modern position is more permissive distally and more cautious proximally, and that the strongest patient-reported outcome signal in the entire literature is for the thumb — Stone’s 2021 meta-analysis found an MHQ advantage of 11.88 points for thumb replantation, which exceeds the minimal clinically important difference. For single non-thumb digits the same meta-analysis found a difference that was statistically significant but, in the authors’ own words, likely not clinically important.”

“Primary survey first — the part never takes priority over the patient. Once I know this is an isolated injury:

• When — time and mechanism of injury, and specifically the warm ischaemia time from the moment of injury, not from arrival.
• How the part has been stored — saline-moistened gauze, sealed bag, in iced water without direct ice contact.
• AMPLE — allergies, medications, past illnesses, last oral intake, events. Nil by mouth from now.
• The patient — age, hand dominance, occupation, smoking, comorbidity, and what they need the hand to do.
• Examination — the stump and the part, both under loupe magnification if possible; the level; the extent of the zone of injury; whether the vessels are ribboned or the nerves telescoped, which tells me this was avulsion whatever the history says.
• Radiographs of both the stump and the amputated part — that second film is the one candidates forget, and it is what lets me plan the osteosynthesis before the patient is asleep.

Then tetanus, antibiotics per local policy, analgesia, and a conversation with my consultant.”

“Wrapped in saline-moistened gauze, sealed in a watertight plastic bag, and that bag placed in iced water — never in direct contact with ice, never immersed, never frozen. The target tissue temperature is around 4 degrees, plus or minus 2.

That is the wording of both the BSSH/GIRFT 2022 traumatic amputations pathway and the 2025 ILCOR First Aid consensus. The experimental basis is *Partlin’s 2008 study in the Journal of Trauma***, which found a sealed container in an ice-and-water slurry held tissue in that window for 226 minutes, against 62 minutes for the worst method tested.

Direct ice contact causes freezing injury and can make an otherwise replantable part unusable. And I would be honest that the ILCOR review found 73% of the underlying literature is case reports, with no randomised comparison of cold storage techniques.”

“I would use Tamai zones for level and Yamano for mechanism, and I would be clear that they answer different questions.

Tamai zone I is from the fingertip to the base of the nail; zone II is from the base of the nail to the DIP joint. That is Tamai’s original distal classification — a five-level extension to the MCP joint and beyond is in common use but is a later addition.

Yamano classifies by mechanism, not level: type I guillotine, type II crush, type III crush-avulsion.

If more granularity is needed for the distal part, Ishikawa’s four subzones divide the region at the midpoint of the nail, the base of the nail and the midpoint between nail base and DIP joint — and that matters practically, because vein repair is almost impossible in subzone I and very difficult in subzone II.”

“Urbaniak’s 1981 classification, as modified by Kay.

Urbaniak: class I, circulation adequate; class II, circulation inadequate; class III, complete degloving or complete amputation.

Kay subdivides the inadequately perfused groups by whether the problem is arterial inflow or venous outflow, and by whether there is a skeletal injury, and adds a class IV for complete amputation. Adani further subdivides class IV into distal to FDS insertion, proximal to it, and complete degloving with intact tendons.

The reason I use it is prognostic. Bamba’s 2018 systematic review of 572 patients found completion amputation in 5.9% of class II but 30.6% of class III, with microsurgical failure rates of 9% and 15.9% respectively.”

“The Chen classification — four domains: return to work, range of motion, sensibility and muscle power.

Grade I, excellent: back to original work, collective range of motion over 60% of normal, high-grade sensibility without excessive cold intolerance, power M4 to M5.

Grade II, good: some gainful work but not the original job, ROM over 40%, near-normal median and ulnar sensibility, M3 to M4.

Grade III, fair: independent in activities of daily living, ROM over 30%, poor but useful sensibility, M3.

Grade IV, poor: tissue survival with no useful function.

It is a major-limb instrument and it is not validated for digits. For digits I would use the MHQ or the DASH, following the FRANCHISE validation work — and specifically not the brief MHQ, which was shown not to be reliable in digital amputation, and not SF-36, which discriminated poorly.”

“Two teams, and the single biggest saving in ischaemia time is that one team prepares the part while the other prepares the stump.

Under tourniquet, both ends are debrided radically back out of the zone of injury, and every structure is identified and tagged before I start repairing anything. Then:

• Bone — shortened enough to allow tension-free repair of everything else, and stably fixed. This is the platform for the whole operation.
• Flexor tendons — repaired before the artery, while the field is dry.
• Artery — moved earlier in the sequence if ischaemia has been long.
• Nerves — primary repair wherever shortening allows it.
• Extensor tendons.
• Veins — at least two per repaired artery where the calibre allows.

Then skin closed without tension, dressings that will not constrict, and the digit visible for observation.

The one thing I would emphasise is that the sequence is a default, not a rule. If ischaemia is long I restore inflow first, and I would say so to the examiner rather than reciting a fixed order.”

“I would still replant it, artery-only, and plan the venous outflow strategy before I start.

The evidence supports this. *Crowe’s 2022 systematic review in PRS, 55 studies and 1,498 digits, reports 78.5% survival with artery-only replantation and concludes explicitly that the absence of a repairable vein is not a contraindication. Kobayashi’s 2026 series of 207 digits found no significant difference between artery-only and artery-plus-vein replants* in any Tamai zone.

For outflow I would remove the nail plate and apply heparin-soaked gauze, or make a paraungual stab incision and maintain a controlled bleed with heparinised saline. Leeches if that fails. The physiological endpoint is neovascularisation, which takes about five to seven days — so I would plan for that duration and consent for transfusion.

The honest caveat is the cost: on Buntic’s standardised artery-only protocol, all 19 replants survived, but 11 of 17 patients were transfused, mean 1.8 units, with a mean hospital stay of nine days.”

“I would replant it. The thumb is the one digit where I would attempt replantation at almost any level and in almost any mechanism, because it contributes around 40% of hand function, there is no adjacent digit to compensate, and secondary reconstruction of a thumb is far harder than of a finger.

The evidence is the strongest in the whole field for the thumb — Stone’s 2021 meta-analysis found an MHQ advantage of 11.88 points, 95% confidence interval 7.78 to 15.99, which exceeds the minimal clinically important difference. By contrast the same meta-analysis found the advantage for single non-thumb digits was statistically significant but, in the authors’ words, likely not clinically important.

Technically, the thumb digital arteries are awkward to anastomose directly because of their orientation and the pronated position, so I would have an interpositional vein graft from the ulnar or radial digital artery to the dorsal branch of the radial artery in the snuffbox as my plan B, and I would prepare the snuffbox at the start rather than discovering I need it at hour four.”

“Multiple digits is a clear indication, and my priority order is functional importance and probability of survival — thumb first, then the digit contributing most to grip.

Practically: two teams, and my registrar explores and prepares all the amputated parts while I prepare the recipient stumps. Under the first tourniquet I would identify and tag all recipient vessels, nerves and tendons. Under the second, bony shortening, osteosynthesis and tendon repairs. Then down with the tourniquet — thumb revascularised first, then the digits in order, then veins, then nerves.

I would not feel obliged to do everything tonight. Woo’s 2015 series of delayed and suspended replantation replanted the important digits immediately and suspended the rest to the following morning; 28 of 32 parts survived, 88%, against 84% in 711 immediate replants at the same institution over the same period. With a mean second-stage cold ischaemia of nearly 16 hours. That is the evidence that legitimises operating in daylight with a fresh team rather than at three in the morning.”

“This is a muscle-containing part — there are intrinsics in the hand — so I am working to major-replant ischaemia limits, not digital ones, and the BSSH standard sets a UK target of reattachment within 4 hours for any part containing muscle. That governs everything that follows.

Two teams. One prepares the hand: radical debridement, washout, identification and tagging of the radial and ulnar arteries, the dorsal and cephalic veins, the median and ulnar nerves and the superficial branch of the radial nerve.

Skeletally, my aim is stable fixation with enough shortening for tension-free repair of vessels, nerves and skin. At this level the options are proximal row carpectomy, radiocarpal arthrodesis, or transarticular K-wire stabilisation — and I would choose PRC because I need the shortening or because the proximal row is unreconstructable, not simply because of the level.

Then the standard sequence, with the artery brought forward if ischaemia is long — and if it is, a temporary shunt while I do the osteosynthesis. I would repair both radial and ulnar arteries and at least two veins, using a coupler where calibre allows.

I would finish with forearm and hand fasciotomies and a carpal tunnel release, and leave the skin open for delayed grafting.”

“This is a decision I would make with a consultant colleague, and I would want to be clear that the honest goal here may not be a functioning hand.

The factors are the patient’s physiology, the mechanism, the warm ischaemia time and the state of the nerves.

Mechanism first, because at this level the relevant framework is Chuang’s classification of traction-avulsion amputation. Type I is at the musculotendinous junction with muscles essentially intact. Type II is within the muscle bellies with proximal muscles still innervated. Type III involves the motor nerve or the neuromuscular junction, so the muscle is denervated at the moment of injury — and replanting it restores a limb with no motor unit. Type IV is through a joint. If this is a Chuang type III, vascular success is irrelevant to function and I would say so.

And I would be explicit about the nerve problem. Axons regenerate at about a millimetre a day, so from a transhumeral coaptation to the hand intrinsics is three to four years — and functional neuromuscular junctions fail to re-form after about 18 months of denervation. The window closes long before the axons arrive. I would plan on the assumption that intrinsic hand function will not return.

What is realistically achievable is elbow flexion, protective sensation, and a limb the patient uses for gripping, supporting and balance — and, if the replant partially fails, the conversion of an above-elbow amputation into a functional below-elbow level. That last is a legitimate stated goal and it is exactly what Wood and Cooney described in their transhumeral series.”

“The governing principle is that this is a systemic operation, not a limb operation. The limb can kill the patient at reperfusion.

• Resuscitate as polytrauma. Life before limb. Trauma CT, group and save, cross-match, coagulation, baseline potassium, creatine kinase and lactate.
• Assess the part for salvageability, dissect out the vessels I will use, and mark a saphenous vein donor site before I start.
• Restore perfusion early with a temporary shunt — a Javid, Pruitt-Inahara or Argyle shunt, or an improvised paediatric feeding tube — while osteosynthesis is done.
  Rasmussen’s wartime series reported 86% patency for proximal shunts without systemic heparin.
• Radical debridement out of the zone of trauma, and bone shortening — typically 2 to 5 centimetres — with plate fixation.
• Flush the limb and discard the first venous effluent before completing venous outflow, so the ischaemic load does not go straight into the circulation.
• Arterial repair out of the zone of injury, with a reversed saphenous interposition graft if needed. Veins repaired with a coupler — the bigger the better.
• Primary nerve repair wherever shortening allows it, and honesty with the patient that in an avulsion the result may be poor regardless.
• Mass closure of muscle, then fasciotomies of forearm and hand, and carpal tunnel release. Skin left open for delayed grafting, with a planned second look at 48 to 72 hours.
• Post-operatively: anticipate hyperkalaemia and myoglobinuric renal injury — calcium, insulin-dextrose and bicarbonate available at declamping, aggressive fluids, serial CK, potassium, lactate and gas.”

“Warm, wet, pain-free and observed — and observed by people who do this regularly.

• Nursed in a warm side room or HDU, with a warming blanket for at least the first 24 hours, and the patient confined to the ward.
• Hourly observations for the first 24 to 48 hours, comparing the replanted digit against an adjacent uninjured one, in consistent lighting: colour, capillary refill, turgor and temperature.
• Well filled — copious oral fluids with intravenous crystalloid, targeting a urine output above 50 millilitres an hour and a systolic above 100. Diuretics avoided.
• Pain controlled, ideally with a regional catheter.
• Caffeine, nicotine and vasoconstrictors prohibited.
• Elevation just above the level of the heart — high enough for oedema, not so high that arterial inflow suffers.
• Hand therapy involved from day one, with mobilisation dictated by the fixation and the tendon repairs.

And I would be honest that almost none of this is evidenced. A 2023 survey of the largest Nordic replantation units concluded there is no robust evidence to support or reject most strategies in post-operative care. It is physiologically rational and universally practised, which is not the same thing.”

“Clinically, and the single most discriminating test is the pinprick.

Arterial insufficiency: pale or mottled, sluggish or absent capillary refill, reduced turgor — a soft, empty, prune-like pulp — cold, and on pinprick nothing, then scant bright blood. It comes on suddenly.

Venous congestion: blue or dusky, brisk capillary refill that looks too fast, increased turgor — tense and swollen — and on pinprick immediate brisk dark blood. It comes on insidiously.

So: dark blood immediately is venous; nothing then bright red is arterial.”

“Intensively for the first 24 to 48 hours, and then I would not relax as much as the free flap literature would suggest.

For free flaps the timing data are clear — Chen’s series of 1,142 flaps found 82% of compromised flaps presented within 24 hours and 95.6% within 72, and 85% of those presenting within 72 hours were salvaged.

But that does not transfer cleanly to digits. Florczynski’s 2022 cohort of 434 digits found 71% success, 18% early failure in hospital, and 11% late failure — digits that were viable at discharge and failed afterwards. Complete amputations and leeching were strongly associated with both. The authors’ own recommendation is that congested digits should stay in hospital longer.

So: roughly one in ten replanted digits fails after discharge, and I would say that at consent.”

“I would escalate in order, and I would start with the things that cost nothing.

• Elevate.
• Take the dressing down completely and release any tight sutures. This is the commonest reversible cause and it takes two minutes.
• Remove the nail plate and apply heparin-soaked gauze, or make a paraungual stab incision and maintain a controlled bleed with heparinised saline.
• Medicinal leeches if that fails, with antibiotic prophylaxis started before the first leech.
• Return to theatre if a venous anastomosis was performed and conservative measures fail.

And I would go back early rather than late. Bui’s series of 1,193 free flaps found salvaged flaps were re-explored at a mean of 4 hours against 9 hours for those that failed — and that venous thrombosis is both more common than arterial, 74% versus 26%, and more salvageable, 71% versus 40%.”

“Aspirin, heparinised irrigation intra-operatively, and low molecular weight heparin for venous thromboembolism prophylaxis — and I would be honest that almost none of this is evidenced.

The only Cochrane review in the field is negative — Lin 2020, four randomised trials and 273 digits, no evidence of benefit for low molecular weight heparin over unfractionated heparin or over no treatment, at low to very low certainty.

The largest cohort argues against therapeutic heparin — Zhu 2023, 1,155 digits: 79.9% success with post-operative therapeutic intravenous heparin versus 92.8% without, with more bleeding. I would immediately add that this is retrospective and heavily confounded by indication — heparin went to the worse injuries — and the authors’ own conclusion is the weaker “no benefit and more complications”, not that heparin causes failure.

Dextran I would not use. Disa’s randomised trial found no survival benefit and systemic complications in 51% at 120 hours and 29% at 48 hours, against 7% for aspirin.

So my position is: aspirin because it is cheap, safe and biologically plausible; heparinised irrigation because it is local; therapeutic systemic anticoagulation reserved for a specific intra-operative concern such as a difficult anastomosis or a vein graft, not as routine.”

“Failure of capillary perfusion to be restored despite a patent, technically satisfactory arterial anastomosis and adequate inflow, after a critical period of ischaemia. It is a microcirculatory failure, not a macrovascular one.

It was described in the brain by Ames in 1968, and translated to flaps by *May, O’Brien and Hurley in PRS in 1978, who showed in a rabbit free flap model that the obstruction is progressive and reaches a point of irreversibility after 12 hours of ischaemia.*

The mechanism has three components: endothelial and parenchymal cell swelling from ATP depletion and sodium-potassium ATPase failure; intravascular aggregation of red cells, platelets and leucocytes; and interstitial oedema compressing the capillaries from outside.

Clinically it is why ischaemia time matters even when the anastomosis is perfect, why cooling the part is critical, and why a technically flawless replant of a digit ischaemic for 24 warm hours can still fail.”

“Possibly — but I would be clear that the evidence for thrombolysis is weak and largely negative, and I would not present it as a rescue.

Panchapakesan’s series found salvage in only 6 of 20 flaps given streptokinase or urokinase, 30%, against 54% in those judged not to need thrombolytics — and no flap with a vein graft was salvaged. Chang’s series: anastomotic revision alone salvaged 92.7%, revision plus thrombolytics 84.8% — no significant benefit. Brouwers’s systematic review of 27 studies concluded no thrombolytic agent had superior salvage outcomes and that there is no consensus on indication, dose or route.

The conceptual case for it is narrower and better: Namgoong’s series identified six flaps with intra-flap microthrombosis and no thrombus at the anastomosis — all six were salvaged with a urokinase protocol. That is the situation you have described, and it is the one situation where thrombolysis makes mechanistic sense, because anastomotic revision cannot address microthrombus.

I would not quote a dose. There is no validated replantation-specific protocol, and delivery is local into the arterial inflow with the venous outflow open to drain externally, precisely to avoid systemic bleeding.”

“Four things, and I would say them before the operation rather than after.

Survival is mechanism-dependent — around 85% for a sharp injury and 45 to 55% for a crush or an avulsion, and I would give this patient their number, not an average.

A surviving replant is not automatically a successful replant. Roughly half of surviving digital replants need at least one further operation, and after a major limb replant the average is three. Around one in eight replants in the pooled literature ends in re-amputation.

Cold intolerance is very likely and it does not improve. Povlsen’s 12-year prospective study is explicit that cold-induced vasospasm in replanted digits does not improve with time — patients with moderate symptoms may perceive improvement, but that is changed habits, not physiology.

Return to work takes months, not weeks. Treger’s systematic review of 31 studies found 82.9% of patients returned to work at a mean of 4.7 months, with 90.9% resuming their previous occupation — but only 66% for finger-only replantation, against 82.8% for the thumb.”

“Early surgery is soft tissue; late surgery is tendon. Yu’s series divided at two months and found early procedures were 92% soft tissue coverage and late procedures 67% tendon. Overall, tendon procedures account for 47% of all secondary surgery after digital replantation, joint procedures 19%, skeletal stabilisation 12%, skin coverage 11%, nerve reconstruction 9%, and late amputation 1.6% — Wang’s review.

The sequence I would work to is: supple skin coverage, then skeletal stability, then protective sensation, then joint reconstruction, then tendon reconstruction. That order matters because a tenolysis under unhealed skin over an ununited fracture achieves nothing.

And flexor tenolysis is the one secondary procedure with a demonstrated functional benefit — Yu found it significantly improved digital function.

The honest range for how often: Wang reports published rates from 2.9% to 93.2%, driven by level, mechanism and how aggressively the unit offers reconstruction.”

“That is a fair question and the evidence says volume matters.

Brown and Chung’s analysis of 631 replantation attempts found that hospital annual replantation volume increased the odds of success, and that an annual volume of three replantations was needed to achieve a 70% success rate. Mahmoudi found risk-adjusted thumb replantation success 12% higher in high-volume hospitals.

For major limb replantation the signal is even stronger: Khelfallah’s 2026 multicentre analysis found surgeon expertise level was the dominant predictor of survival, adjusted odds ratio 15.8.

And the practice is disappearing. Reavey’s analysis of US national data found replantations fell from 930 in 2001 to 445 in 2011, a more than 50% fall, while amputation incidence was unchanged — and around half to two-thirds of the hospitals still doing them performed only one a year.

So my answer is: this should be done in a designated centre by a team that does it regularly, and the UK has no formal designation of replantation centres — which I think is a genuine gap. In practice I would discuss it with the regional hand trauma service tonight.”

“Yes — I would replant more readily, but not indiscriminately.

The case for replanting in a child is growth potential, superior nerve regeneration and the psychological consequence of a lifelong deficit. Özdemir’s comparative series found survival in children no different from adults, but static two-point discrimination significantly better — 4.0 against 6.2 millimetres — and time to regain sensation of 1.3 months against 4.1.

Technically: vessels often under half a millimetre, so 11/0 nylon and sometimes 12/0; a marked vasospastic tendency, so a warm theatre, topical papaverine or lidocaine and a good block; smooth K-wires avoiding the physis where possible; and a higher transfusion requirement relative to circulating volume if I use an external bleeding technique — which was the one significant difference Özdemir found.

In a very young child, composite grafting is a genuine alternative for a fingertip.
Murphy’s series of 96 children, median age 2.4 years, found 68% had some degree of take and only 2% needed secondary revision — and neither time to surgery nor level of amputation significantly influenced the outcome.

The caveat I would give the parents is that around one in five paediatric replants goes on to revision amputation, from the US national data.”

“Replantation is the restoration of a completely amputated part. Revascularisation is the restoration of circulation to an incompletely severed part that has lost its circulation, irrespective of the nature or the amount of tissue still connecting it. That is Biemer’s 1980 definition in the British Journal of Plastic Surgery.

Macroreplantation is conventionally an amputation proximal to the wrist or ankle — that is, one containing significant muscle mass — and microreplantation is distal to that. I would add that the functional boundary is muscle mass rather than the anatomical landmark itself, because muscle is what sets the ischaemia clock and generates the systemic risk. O’Brien in 1976 drew the line more distally, at the MCP joint.”