A patient twists their foot, it swells, they limp, they can no longer bear weight. The reflex is to call this an ankle sprain, and it is precisely that sentence which costs the prognosis. This article is about the midfoot: the tarsometatarsal joint, known as the Lisfranc joint. For the anterior talofibular ligament and the lateral ankle, see our dedicated article on lateral ankle sprain. The two look alike at first examination, are treated in opposite ways, and only one of them leaves three patients out of four with osteoarthritis ten years later when it is poorly reduced.

Clinical synthesis

What to have in mind before reading the rest, if this patient walks in tomorrow morning.

Why is a Lisfranc injury taken for an ankle sprain?

This question is not rhetorical: it is the error mechanism that produces most of the diagnostic delays, and it can be taken apart point by point.

The patient arrives with a swollen foot after a twisting injury. They limp, they point to “the ankle” with a broad gesture, they cannot put the foot down. Three elements of this picture are shared word for word with lateral ankle sprain, which is twenty to a hundred times more common. Clinical reasoning does what it always does: it keeps the most probable hypothesis. And it is right in the vast majority of cases, which is exactly what makes the error so hard to correct.

The problem is that the four features which separate the two injuries are all discreet, and that none of them leaps out unless it is looked for.

The first trap: the location the patient gives you

A patient almost never says “my midfoot hurts”. They say “my foot hurts” or “my ankle”, and they point to a ten-centimetre area. If the examination starts with palpation of the bundles of the lateral collateral ligament, it will almost always find something tender, because an injured, oedematous foot is diffusely painful. The test ends on a confirmation that is nothing of the sort.

The remedy is mechanical: palpate the line of the metatarsal bases systematically, from the first to the fifth ray, before even touching the malleoli. Exquisite, reproducible tenderness over the joint line between the first cuneiform and the base of the second metatarsal does not belong to the register of ankle sprain.

The second trap: the mechanism, poorly questioned

“I twisted my foot” covers two mechanisms that have nothing in common. In lateral ankle sprain, it is the foot that goes into inversion under a leg that carries on along its path. In Lisfranc injury it is the reverse: the forefoot is fixed to the ground, often in forced plantar flexion, and it is the body that topples or is loaded over the top. The registries find a clear predominance of low-energy mechanisms: in the Swedish fracture register, a simple trip and fall accounts for 31 % of the 2,084 injuries recorded between 2013 and 2022, and only one case in five arises from high-energy trauma (Juto 2025, PMID 39626601). In the Norwegian series, only 31 % of injuries were high-energy (Stødle 2020, PMID 31257042).

In other words: waiting for a spectacular injury before considering a Lisfranc means ruling out two thirds to four fifths of cases from the start.

The third trap: the normal radiograph, taken for proof

The patient leaves the emergency department with a foot film and the verdict “no fracture”. The film was taken lying down, without load, because the patient could not stand. Yet it is precisely weightbearing that separates the joint surfaces and makes the diastasis visible. A normal non-weightbearing radiograph, in a patient whose midfoot is painful, is an inconclusive test, not a negative one. We return to this in detail in the imaging chapter.

The fourth trap: the Ottawa rules, applied outside their domain

The Ottawa ankle and midfoot rules are among the best validated triage tools in trauma care: the meta-analysis of 66 studies reports a pooled sensitivity of 99.4 % (95 % CI: 97.9-99.8) for a specificity of 35.3 % (95 % CI: 28.8-42.3) (Beckenkamp 2017, PMID 27884861). That sensitivity figure is reassuring, and that is where the reasoning error slips in.

These rules were built and validated to rule out a fracture. Their midfoot criterion covers the base of the fifth metatarsal and the navicular. A purely ligamentous Lisfranc injury is not a fracture, and its site is neither the navicular nor the base of the fifth. A patient can therefore satisfy the Ottawa rules, receive no radiograph, and still have an unstable midfoot injury. This is not a flaw in the rule: it is an application outside its domain of validity.

What the error costs, in figures

The published missed-diagnosis rates vary enormously, and that variation is itself instructive: it measures the recruitment as much as the difficulty of the diagnosis.

The attentive reader will note that Renninger finds no significant difference between high and low energy in the missed-diagnosis rate: 21 % versus 18 %. This is a counter-intuitive and important result. One might think that the discreet injuries are missed and the shattered ones spotted; in reality both are missed to much the same extent, for different reasons: the former because they are faint, the latter because they are lost in a polytrauma.

The table that separates the two diagnoses

What is the midfoot, and why is the Lisfranc complex so vulnerable?

Understanding why an almost immobile joint becomes unstable after a discreet injury means looking at its geometry before its ligaments.

The midfoot is the region of the foot lying between the hindfoot and the metatarsals. The tarsometatarsal joint (known as the Lisfranc joint, after the military surgeon Jacques Lisfranc de Saint-Martin) is the line of union between the cuneiforms and the cuboid on one side, and the five metatarsal bases on the other. It crosses the foot obliquely, in a staircase rather than a straight line.

Three columns, three roles

Functionally, three columns are described:

• The medial column: first cuneiform and first metatarsal. It carries most of the load at push-off.
• The intermediate column: second and third cuneiforms, second and third metatarsals. This is the rigid column, the one that must not move.
• The lateral column: cuboid, fourth and fifth metatarsals. This is the mobile column, the one that lets the foot adapt to the ground.

This difference in mobility has a direct surgical consequence, which we return to: the lateral column is not fixed like the other two, on pain of stiffness.

The keystone: what the geometry of the second metatarsal imposes

The base of the second metatarsal is recessed and keyed into a mortise formed by the three cuneiforms. This mortise-and-tenon arrangement is what locks the whole line: the second ray cannot move laterally as long as the mortise is intact. It also explains why the injury concentrates there, because a force that drives the first metatarsal away from the second comes up against that keying, and it is the ligamentous link, or the bone itself, that gives way.

An anatomical fact that is decisive for reading radiographs: there is no transverse intermetatarsal ligament between the base of the first and that of the second metatarsal. Stability between these two rays is provided entirely by an oblique ligament running from the first cuneiform to the base of the second metatarsal, the Lisfranc ligament proper. That is why widening between M1 and M2 on a weightbearing film signals injury to that ligament, and not a benign anatomical variant.

Three bundles, but only one decides

The systematic review of the cadaveric literature describes the Lisfranc ligamentous complex as three distinct structures: the dorsal bundle, the interosseous bundle and the plantar bundle. Across eight cadaveric studies selected from 1,204 screened articles, it is the interosseous bundle that shows the greatest thickness, the largest bony insertion area, the highest stiffness and the best resistance to failure when loaded along the axis of its fibres (Sripanich 2021, PMID 33639336).

But thickness is not function. The sequential sectioning cadaveric study is more instructive still: across sixteen lower limbs placed in a radiolucent loading frame and scanned at each stage, section of the dorsal ligament and of the interosseous ligament produces only a slight increase in the M1-M2 and C1-M2 distances. The widening worsens on sectioning the plantar ligament between the first cuneiform and the third metatarsal, and becomes maximal after sectioning the plantar ligament between the first cuneiform and the second metatarsal. The authors' conclusion is explicit: frank instability appears only with simultaneous sectioning of the Lisfranc ligament and of the two plantar ligaments (Penev 2021, PMID 31950232).

This plantar hierarchy explains two things in the clinic. First the diagnostic value of plantar ecchymosis: it reflects bleeding from the plantar structures, the ones whose injury tips the balance towards instability. Then the weight MRI gives to the plantar bundle in the decision, which we return to in the imaging chapter.

What injury mechanisms produce a Lisfranc injury?

Reconstructing the movement is the first step of the examination, and often the one that orients you fastest. But the right question still has to be asked.

The indirect mechanism, by far the most common

The foot is in plantar flexion, the forefoot resting on the ground through the metatarsal heads. An axial load travels down the limb, or a rotation is applied to the tibia while the forefoot stays fixed. The dorsal ligaments, more slender than the plantar ones, give way first; the metatarsal bases move dorsally and laterally. Depending on the intensity, the result is a simple ligamentous sprain, a bony avulsion at the ligament's insertion, or a frank dislocation.

The typical situations are not exotic:

• A rugby or American football player tackled from behind while their toes are anchored to the ground.
• A rider whose foot stays caught in the stirrup during a fall.
• A motorcyclist whose forefoot is jammed against the gear lever.
• A walker who misses a step and loads their whole weight onto a foot in equinus.
• A dancer in demi-pointe or en pointe, in whom the mechanism is repeated rather than single: we devote a chapter to it.

The direct mechanism, rarer and more serious

A crush injury to the dorsum of the foot (a vehicle wheel, a heavy load falling) produces a different injury: more fractures, more columns involved, more soft-tissue damage, a risk of compartment syndrome. These are not the injuries that get missed, but the ones whose prognosis stays bleakest: in Myerson's original series, only one of the eight direct crush injuries achieved a good or excellent result (Myerson 1986, PMID 3710321).

Low energy versus high energy: two injuries under one name

The most useful comparison in practice comes from Renninger's series of 80 operated injuries, which sets 32 high-energy injuries against 48 low-energy ones (Renninger 2017, PMID 28693353). The differences are clear-cut:

The clinical reading fits in one sentence: the low-energy injury is isolated, ligamentous, limited to two or three rays, that is to say exactly the injury that shows almost nothing on a non-weightbearing radiograph. And it is the one that dominates in numbers.

One point deserves a pause. Stødle reports that unstable injuries are more common in women (p = 0.016), and identifies two further predictors of instability: the presence of intra-articular fractures of the two lateral tarsometatarsal joints (p = 0.007), and a reduced height of the second tarsometatarsal joint (p = 0.036), that is, a shallower mortise (Stødle 2020, PMID 31257042). This last point links the anatomy described above directly to clinical risk: the less deeply the keystone is set, the less it holds.

Which clinical signs should alert you?

No single sign makes the diagnosis. It is their accumulation that must trigger the request for the right imaging, and you have to know which ones to look for.

Inspection, before touching

Three things are visible, provided the patient is completely unshod and the two feet are compared side by side, soles included.

Swelling of the dorsum of the midfoot. It is central, in front of the instep, and not retromalleolar. Compared with the other foot, it visibly widens the region of the metatarsal bases.

Plantar ecchymosis. This is the sign that carries a name. Described in 1996 in a series of patients whose radiographs did not show the true extent of the injuries, it consists of an area of ecchymosis over the plantar aspect of the midfoot, and it reflects potentially major injury to the plantar tarsometatarsal ligaments. The authors conclude that patients carrying this sign must be assessed aggressively, with early recourse to stress films (Ross 1996, PMID 8932671).

Collapse of the medial arch. It can only be judged under load, and therefore only when the patient can put the foot down. It is a late sign, which already signals an advanced-stage injury.

Palpation, systematic and topographic

The rule is to palpate the whole tarsometatarsal line, from the first to the fifth ray, before anything else. You are looking for:

• Exquisite tenderness over the first cuneiform – second metatarsal base joint line, which is the elective site.
• A palpable gap, a “hollow”, between the bases of the first and second metatarsals.
• Tenderness on palpation of the base of the second metatarsal itself, which should also raise the possibility of a stress fracture in an athlete or a dancer.

The stress manoeuvres

They reproduce the pain by loading the ligamentous complex. None has published diagnostic values in the sources we verified; they serve to reinforce a suspicion, not to settle it.

• Passive abduction-pronation of the forefoot, with the hindfoot held fixed. The movement separates the medial column from the intermediate column and puts the Lisfranc ligament under tension.
• Vertical mobilisation of a single ray (the “piano key test”), stabilising the neighbouring ray. Abnormal mobility or reproduced pain points to instability of that ray.
• Transverse compression of the forefoot at the level of the metatarsal bases, which reproduces the pain if the line is injured.

The functional tests, which speak loudest

These are the ones that sort best, once the hyperalgesic phase of the first three or four days has passed.

• Single-leg stance. Being unable to stand on the injured foot, or a markedly shortened time compared with the healthy side, is a strong signal. The single-leg stance deficit in fact persists in the long term after an operated injury, and correlates with isokinetic strength (Mehlhorn 2017, PMID 28043054).
• Heel raise onto the toes. It loads the tarsometatarsal line directly through push-off. A patient who cannot perform it, or who performs it while carefully avoiding the first ray, should be imaged.
• A small hop on the spot, then running on the toes. In Wadsworth and Eadie's case report, it is precisely the ability to run on tiptoe shortly after the injury, together with the absence of bony displacement, that led to conservative treatment in a professional basketball player (Wadsworth 2005, PMID 15839309). The test therefore works in both directions.

Which imaging to request, and in what order?

This is the chapter that decides. A well-chosen investigation rescues a hesitant clinical examination; a badly chosen one turns hesitation into false certainty.

The non-weightbearing radiograph: an investigation that does not conclude

This is the investigation the patient has almost always already had. Three views (anteroposterior, lateral, 30-degree oblique) taken lying down, because they could not stand. What is classically looked for is the fleck sign, a small bony avulsion in the space between M1 and M2, which signals that the Lisfranc ligament has been torn from its insertion.

The problem is not that this investigation is useless, because it detects frank dislocations and fractures. The problem is logical: without load, the joint surfaces are held in place by bony congruence alone, and a torn ligament shows no widening at all. The systematic imaging review concludes that conventional radiography can show a frank diastasis, but that weightbearing improves the ability to detect subtle injuries (Sripanich 2020, PMID 31368007).

A normal non-weightbearing radiograph in a patient whose midfoot is painful is therefore an inconclusive investigation. Putting it that way, in the report as in the conversation with the patient, stops them leaving convinced that “all is well”.

The bilateral weightbearing radiograph: the first-line investigation

Three views of the foot, weightbearing, and above all the contralateral foot under the same conditions. It is the comparison that gives the information, because the normal distance between C1 and M2 varies widely between individuals.

The most useful Australian study on this point included 117 patients who had both bilateral weightbearing radiographs and a CT scan. The diastasis ratio, that is the measured gap on the injured side divided by the gap on the healthy side, was 1.53 (95 % CI: 1.41-1.65) in the operated patients, versus 1.11 (95 % CI: 1.07-1.16) in those treated without surgery, a highly significant difference (p < 0.001). The correlation between the diastasis measurement and the decision to operate was R = 0.576 (p < 0.001) (Kennelly 2019, PMID 30780193).

The most counter-intuitive result of this study concerns CT. When the initial weightbearing radiograph was positive, 54 % of the CT scans that followed were reported as equivocal or negative. And when the weightbearing radiograph was negative or equivocal, only 12 % of the CT scans were positive. The authors conclude that “CT offers limited benefit to the diagnosis and initial management of subtle Lisfranc injuries in the emergency setting” and argue for the bilateral weightbearing radiograph as the first-line investigation.

CT: bony geometry, not stability

Conventional CT, performed without load, provides what radiography does not: undisplaced fractures, minimal bony subluxations, the state of the lateral columns, the exact map before surgery. The systematic review confirms it: “CT is more useful than radiography for detecting undisplaced fractures and minimal bony subluxations” (Sripanich 2020, PMID 31368007).

What it does not give, performed without load, is stability, that is, the information that decides the treatment. That is the whole import of Kennelly's result. CT remains essential as soon as surgery is considered, and for injuries with a bony component; it does not replace the weightbearing film for the question “is this injury unstable?”.

Weightbearing CT: the most promising avenue, and what is really known about it

Weightbearing CT (weightbearing CT) combines the three-dimensionality of CT with physiological loading. The systematic review of nine studies concludes that it “appears better able to detect subtle Lisfranc instability than conventional CT, presumably because it allows the injury to be assessed under load” (Talaski 2023, PMID 38027457).

The most accomplished work to date is the Oslo prospective cohort. Thirty-eight patients with undisplaced fractures (less than 2 mm) or avulsions in tarsometatarsal zone 1 to 3 had a bilateral single-leg weightbearing CT. The C1-M2 measurements, assessing the integrity of the dorsal bundle and of the interosseous bundle, were combined and compared with the healthy side to produce a difference score. Results:

• Eight patients (21 %) had a difference score greater than 3 mm under load. Instability was confirmed in all eight by a dynamic test under fluoroscopy.
• The other thirty, below the 3 mm threshold, were treated without surgery. Re-examined by a further weightbearing CT at twelve weeks, none showed any sign of instability, with a median difference score of 0.6 mm.
• The reproducibility of the measurement method was excellent: an inter-observer intraclass correlation coefficient of 0.95 and an intra-observer one of 0.97 (Poulsen 2025, PMID 41170783).

MRI: the only investigation that sees the ligaments, with one precise predictor

MRI is, according to the systematic review, “clearly the best investigation for detecting ligamentous abnormalities”, but “its usefulness for detecting subtle Lisfranc instability requires further work” (Sripanich 2020, PMID 31368007). That nuance is the heart of the matter: seeing a torn ligament and predicting instability are not the same thing.

The study that narrowed this gap compares MRI with intraoperative findings in twenty patients (twenty-one feet), with the manual stress test under anaesthesia and surgery as the reference. Seventeen joints proved unstable, four stable. The best predictor of instability was neither the dorsal bundle nor the overall appearance of the complex, but rupture of the plantar ligament between the first cuneiform and the bases of the second and third metatarsals (pC1-M2M3), with a sensitivity of 94 %, a specificity of 75 % and a positive predictive value of 94 %. Nineteen of the twenty-one complexes were correctly classified by MRI (Raikin 2009, PMID 19339574).

Two practical lessons. First, MRI is not read by looking for “the Lisfranc ligament”: it is read by looking for the plantar bundle, consistent with what the sequential sectioning cadaveric study had established. Second, a plantar bundle of normal appearance points to a stable midfoot, and may spare a stress test under anaesthesia in a patient whose clinical picture and radiograph are equivocal, which is the authors' explicit conclusion. Conversely, rupture of the second plantar tarsometatarsal ligament, present in eighteen of the twenty-one feet, had “little clinical correlation with instability”: one more torn ligament is not one more argument.

The decision tree

How should the injury be classified, and what does that change?

Three systems coexist. They do not answer the same question, and only one answers the question that concerns the physiotherapist.

Hardcastle and Myerson: describing the displacement

The anatomical classification, from Myerson's work on 76 fracture-dislocations, describes the direction and the extent of the displacement: involvement of all rays in the same direction (type A, homolateral), partial involvement with one or more rays spared (type B), divergent involvement where the columns separate in opposite directions (type C).

This classification has considerable historical and descriptive value, since it is the one that established that the quality of the initial reduction is the major determinant of unacceptable results, and that displacement greater than 2 mm after closed reduction mandates open reduction (Myerson 1986, PMID 3710321). It also retains prognostic value: type C is one of the three long-term risk factors for osteoarthritis identified, alongside non-anatomical reduction and smoking (Dubois-Ferrière 2016, PMID 27147683).

Its limitation is structural: it presupposes a visible displacement. It says nothing about the injury that appears only under load, that is, the one that gets missed.

Nunley and Vertullo: classifying the athlete's sprain

Published in 2002, this classification was built for midfoot sprains in athletes, on the basis of the clinical picture, weightbearing radiographs and bone scintigraphy:

• Stage I: pain over the Lisfranc complex, normal radiographs, increased uptake on scintigraphy. No displacement. This is the only stage treated without surgery in the authors' protocol.
• Stage II: diastasis of 1 to 5 mm between M1 and M2 compared with the opposite side, without loss of arch height.
• Stage III: diastasis greater than 5 mm with loss of arch height.

The authors report 93 % excellent results at a mean follow-up of 27 months, in fifteen athletes. This figure circulates widely; it should be quoted with its sample size (Nunley & Vertullo 2002, PMID 12435655).

The lasting contribution of this work is not the success rate, it is the demonstration that diastasis is measured by comparison with the healthy side and under load, and that loss of arch height is the severity marker that separates stage II from stage III.

The stability-based classification: the one that governs treatment

The Bone & Joint Journal review of December 2024 states the problem head-on: “the original classification systems were anatomical, and limited as tools to guide treatment decisions”. It proposes instead a classification based on stability, with the weightbearing radiograph and CT as the reference diagnostic tools (Poutoglidou 2024, PMID 39615511).

The shift fits in one sentence: the question is not “what shape is the displacement?” but “does this midfoot hold under load?”. And the answer governs management directly:

• The stable injuries generally do well with non-operative treatment, “most reliably by immobilisation and non-weightbearing for six weeks”.
• The displaced or comminuted injuries require surgery.

Which criteria tip the balance towards surgery?

The physiotherapist does not set the indication, but sees the patient before, between, and after. Understanding what was decided, and why, conditions everything else.

The criterion that commands consensus: stability, and the 2 mm threshold

The principle has been stable for forty years and has not been contradicted: a displaced injury must be reduced anatomically and held there. Myerson already stated that “displacement of more than 2 mm or a talometatarsal angle of more than 15 degrees on radiographs after closed reduction mandates open reduction” (Myerson 1986, PMID 3710321).

Fourteen years later, the Seattle series confirms this in 48 patients followed for a mean of 52 months: “the major determinant of a good result was anatomical reduction” (p = 0.05). The mean AOFAS midfoot score reached 77 points out of 100, with 25 % post-traumatic osteoarthritis, six of those patients requiring an arthrodesis (Kuo 2000, PMID 11097452). The same work notes that purely ligamentous injuries tend towards poorer results despite anatomical reduction and screw fixation, an observation that still feeds today's debate on primary arthrodesis.

The contemporary literature takes up the threshold: “residual displacement of more than 2 mm is associated with inferior outcomes and significantly reduced return-to-play rates” (Hammad 2026, PMID 41522288).

What becomes of an unstable injury left without surgery

The systematic review of non-operative treatment is the most honest document on this question, because it measures failure as much as success. Across eight studies and 220 patients, mean age 39.8 years, mean follow-up 4.3 years:

• Four papers report good results, with adjusted functional scores of 82.6 to 100 out of 100.
• One study reports late displacement in 54 % of patients.
• The rate of secondary osteoarthritis ranges from 5 %–38 %.
• The rate of surgical conversion rises to as much as 56 %.
• Patients whose CT showed no displacement on the C1-M2 measurement had the best outcomes (Guerreiro 2023, PMID 36841140).

The authors' conclusion deserves to be taken up as it stands: the heterogeneity is too great, a clear definition of a stable injury is lacking, and prospective work is needed. In other words, non-operative treatment is not a bad treatment: it is a treatment whose right candidates cannot yet be identified with certainty, which is what Poulsen's work on weightbearing CT is beginning to correct.

Internal fixation or primary arthrodesis: a debate still open

Two meta-analyses published in 2024 reach conclusions that must be read together, because they do not draw on the same material.

The first pools eighteen studies comparing primary arthrodesis and internal fixation in acute injuries. It favours arthrodesis on almost every criterion: AOFAS of 84.4 versus 75.7, VAS pain of 1.4 versus 2.0, return to previous activity in 79.2 % versus 65.7 %, midfoot osteoarthritis in 2.8 % versus 17.3 %, and above all a rate of unplanned reoperations of 14.7 % versus 38.3 % (p < 0.001) once planned metalwork removals are excluded. The pooled effect sizes confirm this: AOFAS 0.41 (95 % CI: 0.13-0.68), osteoarthritis OR 0.29 (95 % CI: 0.11-0.77), reoperations OR 0.16 (95 % CI: 0.06-0.44) (O'Connor 2024, PMID 39680239).

The second retains only the randomised trials: five trials, 241 patients, 121 arthrodeses against 120 internal fixations. It finds an advantage for arthrodesis on pain at two years (mean difference 0.89; 95 % CI: 0.18-1.59), on satisfaction (OR 10.04; 95 % CI: 1.78-56.76) and on all-cause reoperations (OR 27.31; 95 % CI: 12.72-58.63), but no difference in the AOFAS midfoot score at two years, nor in the SF-36, nor in unplanned reoperations. And the authors conclude that “these findings are not powered to confer an advantage to either approach, because of substantial heterogeneity” (Mactier 2024, PMID 39444525).

The constructs, and why they change your timetable

Three families coexist, and the construct chosen determines the timelines the surgeon will pass on to you.

• Transarticular screws. The historical construct. It crosses the cartilage, which has fed the argument in favour of arthrodesis, and it frequently requires metalwork removal, in 82 % of the internally fixed patients in the military cohort (Koehler 2022, PMID 32517507).
• Bridge plates. The Bone & Joint Journal review notes “an emerging consensus in favour of plate bridging” for displaced or comminuted injuries (Poutoglidou 2024). In the Swedish register, plate fixation is the most used construct, in 51 % of operated patients (Juto 2025).
• Flexible suture-button fixation. A systematic review of fourteen articles and 243 feet reports a mean postoperative AOFAS score of 90.1, a mean VAS of 1.5, 100 % return to activity and 100 % maintenance of radiographic alignment, with no complications and no metalwork removals reported. But these figures come from eleven clinical series without a comparison group, with sample sizes from 35 to 216 depending on the criterion, and the three biomechanical studies included are discordant, with a trend in favour of screws for control of the diastasis (Chona 2023, PMID 37538534).

What the lateral column imposes

The lateral column, that is the fourth and fifth tarsometatarsal joints, is mobile by nature, and fixing it rigidly produces stiffness. The recent literature recommends treating it with K-wires rather than with screws or a plate, with early removal (Hammad 2026, PMID 41522288). In practice this means that a patient whose lateral column has been wired will have an extra removal step in their timetable, often at around six to eight weeks, and that your mobility work on the lateral rays will begin after that.

What phased rehabilitation, and at what timelines should weightbearing resume?

This is the chapter where the physiotherapist decides. First we must say honestly what it rests on: a level V expert consensus, a systematic review of postoperative protocols, and a handful of biomechanical series. There is no randomised trial of rehabilitation after Lisfranc injury.

The reference timelines, and where they come from

The most structured source is the international consensus on foot and ankle trauma in sport, conducted in 2024 by a modified Delphi method over four rounds, backed by a systematic review. Its figures:

• Patients with a ligamentous injury were allowed to start weightbearing at 3.5 weeks on average in the literature reviewed.
• The bony injuries were allowed partial weightbearing at 4.5 weeks on average.
• The mean time to return to sport for ligamentous injuries was 8.9 months in the studies surveyed.
• Bony injuries treated by internal fixation allowed a return to sport at a median of 8 weeks (range 3 to 12 weeks) in some series.
• Unanimous agreement of the panel: the athlete can expect to return to their sport 4 to 6 months after surgery, and unstable ligamentous injuries allow a return to full weightbearing 8 to 12 weeks after surgery.
• More than 90 % of athletes do return to competition (Webber 2026, PMID 41582692).

For a stable injury treated without surgery

The reference protocol is the one adopted by the Bone & Joint Journal review: “stable injuries generally do well with non-operative treatment, most reliably by immobilisation and non-weightbearing for six weeks” (Poutoglidou 2024, PMID 39615511).

Two points that are not in the sentence but make it applicable. First, this approach presupposes that stability has been established, not presumed, that is, by a comparative weightbearing film, ideally completed by a CT. Second, it presupposes a weightbearing radiographic review later on, because late displacement is the dominant failure mode: up to 54 % in one of the series surveyed, and a surgical conversion rate of up to 56 % (Guerreiro 2023, PMID 36841140).

The five phases

What must be rehabilitated, as identified by measurement

The only study to have quantified the late functional sequelae is a biomechanical analysis of seventeen patients operated on for a Lisfranc fracture-dislocation, reviewed on average 50.5 months after surgery. Its results guide the content of sessions better than any intuition:

• Significant reduction in peak plantar flexor torque on the injured side compared with the healthy side. Expected.
• Significant reduction in dorsiflexor torque, and it is the dorsiflexor torque, not the flexor torque, that correlated with the clinical outcome. This result is the least intuitive of the study, and the most actionable.
• Significant reduction in single-leg stance time, with a strong correlation between that time and the isokinetic measurements.
• Significant reduction in peak pressure under the midfoot and in the force-time integral under the second metatarsal, the signature of a foot that has stopped transmitting load through its intermediate column.

The authors conclude that “sufficient rehabilitation is decisive for the clinical outcome” and support “a rehabilitation approach centred on restoring proprioception and triceps surae strength, including isometric exercises for the dorsiflexors” (Mehlhorn 2017, PMID 28043054).

The progression table, phase by phase

The modality by level-of-evidence table

What does the context change: athlete, dancer, adolescent?

Three populations in which the injury takes a particular form, and in which it is missed for three different reasons.

The athlete: the midfoot sprain, or the injury that looks like nothing

In athletes the injury is most often purely ligamentous, low-energy, and without displacement visible off load. This is the Nunley-Vertullo stage I or II injury. The athlete walks, limps moderately, and says “it's just a sprain”. Nunley and Vertullo already noted that half the athletes in their series had normal non-weightbearing imaging.

Two NFL series shed light on what becomes of this injury when it is managed properly, and they do not say quite the same thing:

The convergence on timing, medians of 11.1 and 11.0 months over two different periods, is the most solid thing in these two papers. The divergence on performance after return is partly explained by the indicators used, which do not measure the same thing. What can be said without forcing it: return is the rule, it is slow, and the question of a return to the previous level remains open.

In the high-demand non-professional athlete, the American military cohort gives a useful landmark: at a mean follow-up of 3.5 years, 69 % had returned to the daily running required by their role, and 80 % had remained on active duty or completed their term of service. Two predictors stand out, and the first is directly actionable: a BMI below 30 multiplied the probability of returning to running by 13 (OR 13.14; 95 % CI: 2.50-69.19), whereas a BMI of 30 or more multiplied the risk of medical discharge by nearly 18 (OR 17.67; 95 % CI: 3.69-84.53) (Koehler 2022, PMID 32517507).

The dancer: when the mechanism is repeated, not single

The classical dancer spends part of their time in demi-pointe and en pointe, that is, in the exact position that loads the tarsometatarsal line. The injury then takes two forms that resemble one another and are treated differently: the stress fracture of the base of the second metatarsal and the synovitis or sprain of the Lisfranc joint.

Harrington's work on eight ballet dancers sets out the problem and its solution: “diagnosis of this syndrome usually requires distinguishing synovitis of the Lisfranc joint from a stress reaction of the base of the second metatarsal. Prompt diagnosis is important because the treatment of these two conditions differs significantly and, in the case of a bone stress reaction, delay can worsen the injury”. The authors report good clinical results in this group thanks to early diagnosis and treatment, in contrast with the poor results described in the literature when management is late (Harrington 1993, PMID 8368422).

Kriz's case report shows that the two can coexist in the same patient: a 14-year-old pre-professional dancer presenting with both a stress fracture of the base of the second metatarsal and a sprain of the Lisfranc complex, two midfoot conditions “often missed or diagnosed late in these young athletes” (Kriz 2015, PMID 26045400).

The adolescent: a growing midfoot

In children and adolescents, the presence of growth plates and ligamentous laxity change the behaviour of the midfoot under load. Adult radiographic landmarks do not apply directly, and incomplete ossification of the cuneiforms makes the interpretation of gaps more delicate. The sources verified for this article do not allow quantified paediatric thresholds to be proposed, and we shall therefore propose none: the reasonable course is the opinion of a paediatric orthopaedic surgeon for any persistent midfoot pain in a child after trauma, with systematic comparison against the healthy side.

Two populations in which the picture is not traumatic

Two situations deserve to be named because they present in the clinic under the same mask and do not belong to the same register:

• The neuropathic patient, particularly the diabetic one. A warm, swollen midfoot with pain out of all proportion to the clinical appearance should raise the possibility of early Charcot neuroarthropathy, whose site of election is precisely the tarsometatarsal region. Time to management matters as much here as in the traumatic injury, but the approach is not the same, and weightbearing is the main enemy. This situation does not fall within the present article: it calls for prompt referral.
• The older patient after a minor injury. The registries are a reminder that Lisfranc injury is not a matter for the young: the mean age is 43 years in men and 49 years in women in the Swedish register, with cases up to 92 and 96 years, and the most frequent mechanism is a simple trip and fall (Juto 2025, PMID 39626601). Midfoot pain after a same-level fall in an older person is not a contusion by default.

What do concrete clinical cases teach us?

Four published cases, all verifiable by their identifier. They prove nothing, which is the nature of the case report, but they show four trajectories that statistics do not convey.

Case 1. The professional basketball player who was not operated on

Source: Wadsworth DJ, Eadie NT. Journal of Orthopaedic & Sports Physical Therapy, 2005 (PMID 15839309). Published by Australian physiotherapists.

The patient: a 21-year-old professional basketball player with a recurrent ligamentous injury of the Lisfranc joint.

What was decided: conservative treatment, on two grounds: the absence of bony displacement, and the player's ability to run on tiptoe shortly after the injury. Management was explicitly global: assessment and treatment of the whole lower limb and pelvis, not of the midfoot alone.

The outcome: return to the previous level of competition at 12 weeks. At two years of follow-up, the player was still playing professionally, asymptomatic.

What the authors conclude: in patients without bony displacement and able to run on tiptoe shortly after the injury, conservative treatment may be appropriate. Its key elements are recognition of a prolonged recovery time, enough rest for soft-tissue healing, restoration of a normal gait pattern to avoid chronic overload of the injured tissues, an appropriate orthosis prescription, and proprioceptive retraining.

Case 2. The injury missed in the emergency department on the day

Source: van Rijn J et al. Journal of Foot and Ankle Surgery, 2012 (PMID 22168954).

The patient: a 60-year-old sports instructor, the victim of an indirect injury, with no crush and no high energy.

What happened: she attended a hospital the same day, where the Lisfranc fracture-dislocation was not recognised. The diagnosis was made subsequently at another institution, which performed an immediate reduction. Internal fixation had to be deferred by two weeks because of swelling. The course: a non-weightbearing cast for six weeks, then progressive weightbearing in a walking brace, screw removal at three months.

What the authors conclude: these injuries “are missed initially in up to one third of cases”. A careful clinical examination and radiographs in three planes are necessary; CT helps when these remain inconclusive.

The lesson for the physiotherapist: age and the innocuous nature of the mechanism protect against nothing. A 60-year-old sports instructor after an indirect injury is exactly the dominant profile in the Swedish register, not an exception.

Case 3. Six weeks of delay, and a good result all the same

Source: Amirthalingam S et al. Journal of Orthopaedic Case Reports, 2023 (PMID 37255642, PMCID PMC10226627).

The patient: a 53-year-old man presenting with chronic pain and swelling of the foot.

The diagnosis: a homolateral Lisfranc fracture-dislocation (Hardcastle and Myerson type A), neglected for six weeks. The initial injury had been taken for a simple foot sprain.

The treatment and the outcome: open reduction and internal fixation with plates and screws, with an excellent functional outcome at six months of follow-up.

What the authors conclude: Lisfranc injuries are frequently undetected, “with estimates ranging from 20 % to 80 %”. The radiological workup must include anteroposterior, lateral and medial oblique views after a complete physical examination. And a delayed presentation can still be treated by open reduction with a good functional outcome.

Case 4. The 14-year-old dancer, and both diagnoses at once

Source: Kriz P et al. Journal of Dance Medicine & Science, 2015 (PMID 26045400).

The patient: a 14-year-old pre-professional ballet dancer.

The diagnoses: two coexisting midfoot conditions: a stress fracture of the base of the second metatarsal and a sprain of the Lisfranc joint complex, which the authors describe as “often missed or diagnosed late in these young athletes”.

What the authors conclude: these are injuries that may be decisive for a professional dancing career, and a high index of clinical suspicion, a careful physical examination, appropriate imaging and prompt treatment are essential to the best possible outcome.

The lesson: finding one diagnosis does not entitle you to stop looking. Finding the stress fracture and being satisfied with it would have let the sprain slip through, and the other way round.

How is this applied in practice?

What this chapter changes, concretely, in a practice where most of what is seen is ankle sprains.

The three-second rule

It fits in one sentence, to be applied to oneself in front of any injured foot: before concluding “ankle sprain”, palpate the line of the metatarsal bases and look at the sole. It costs fifteen seconds and covers most of the risk.

The five questions to ask the patient who has already been labelled

The most frequent scenario in practice is not the acute patient: it is the patient referred “for rehabilitation of an ankle sprain” two or three weeks later. Five questions are enough to reopen the case when it must be reopened:

1. “Where exactly does it hurt, if you have to show me with one finger?” A finger on the dorsum of the midfoot changes everything.
2. “At the moment of the injury, was your foot flat or up on the toes?”
3. “Did you have a bruise under the foot?” This question is almost never asked, and the patient has often noticed it.
4. “Your radiograph, did you have it standing up or lying down?” The patient always remembers, and the answer “lying down” restores that film to its true value.
5. “Can you rise onto the toes of that foot, here, right now?”

What to ask the referrer, and how

When the arguments accumulate, the request to be made is a precise one. “The patient should perhaps have further tests” gets nowhere. The wording that gets something is the one that names the investigation and its reason:

What to extract from a postoperative file before the first session

• Which columns were fixed? A wired lateral column imposes an extra removal step before working on the mobility of the lateral rays.
• Which construct? Transarticular screws, bridge plate, suture button or arthrodesis: the loading timetable and the existence of a second operation depend on it.
• Is the reduction judged anatomical? It is the leading prognostic determinant, and it tells you what to say to the patient about the long term.
• What date is weightbearing allowed, and with what progression? To be obtained in writing, not by deduction.
• What date is metalwork removal planned? It structures the whole of phase 3.

What to tell the patient

Three messages, to be given early, because they determine adherence over the first six weeks:

• “This is not an ankle sprain, and it does not heal on the same timescale.” A patient who believes they have a sprain goes back too early.
• “Offloading is not a precaution, it is the treatment.” In a stable injury not operated on, offloading is the intervention.
• “If your foot collapses or widens compared with the other one, another radiograph is needed.” Late displacement is the main failure mode, and the patient is best placed to spot it between two appointments.

Useful links within the corpus

For neighbouring diagnoses and the conditions with which a painful midfoot is confused:

• Lateral ankle sprain is the diagnosis to be distinguished from, and by far the most common.
• Plantar fasciopathy covers plantar pain of gradual onset, without trauma.
• Tibialis posterior dysfunction and adult acquired flatfoot is the other cause of arch collapse, and it is not traumatic.
• Post-traumatic ankle osteoarthritis shows what becomes of a poorly reduced joint of the foot.
• Medial tibial stress syndrome in runners covers repeated-loading pain in the endurance athlete.

Frequently asked questions