Best Shoes for Freiberg’s Disease: Offloading, Not Treatment
Freiberg's hits the second metatarsal head. No shoe treats it, but drop and a stiff forefoot offload it. Eight shoes with measured geometry and sources.
On this page
- What Freiberg’s disease actually is
- What the research actually supports — and where it stops
- What non-operative treatment actually is, and how long it is supposed to take
- The one thing a shoe can change: how far the joint has to bend
- The shoes we hold measured geometry for, steepest drop first
- The metatarsal pad goes behind the sore head, not under it
- What to check, in order
- What to stop wearing
- What we cannot tell you
- When footwear stops being the answer
- Questions patients ask
- What happens next on this page
- Sources
- Related guides & shoe pages
Spec analysis — not yet tested · Published September 2026
Nothing on this page has been measured by us. Every stack height and drop below is either the brand’s own published figure or a laboratory measurement published by RunRepeat, which cuts retail shoes in half and measures them. Each row says which. No shoe here carries a score or the Podiatrist Tested seal, because none has been on our bench. How we test →
Freiberg’s disease is a collapse of the bone at the end of a lesser metatarsal — nearly always the second. No shoe repairs it. What a shoe can do is take load off that joint and stop it bending so far, and the geometry that does the offloading is measurable, which is what this page is for. It is also the page that tells you the research on footwear for Freiberg’s barely exists.
What Freiberg’s disease actually is
The metatarsal head — the knuckle at the ball of your foot — loses its blood supply, the bone underneath the cartilage softens, and the top of the joint surface flattens or collapses. It happens to the second metatarsal far more often than any other, and the pain is characteristically at the top of the step: fine standing still, worse the moment the toe bends and you push off.
Two things about it are widely repeated and only half true. The first is that it is a teenage condition. In the largest review of the literature ever assembled — 163 publications, 1,121 patients — 516 patients were 18 or older against 194 under 18. Adults are the majority of the published cases. The second is that it is common. The same review produced the first population-based estimate of how often it happens: 1 in 2,833 livebirths. It is rare, which is exactly why the footwear evidence is so thin.
What the research actually supports — and where it stops
There is no trial comparing shoes for Freiberg’s disease. Not a small one, not a poor one — none. We looked, and what the literature contains instead is surgery. In that same review of every retrievable publication going back to 1914, of 1,169 affected feet, 939 had an operation and 116 did not; for the remaining 114 it was not recorded. Dorsal closing-wedge osteotomies, osteochondral transplants and modified Weil osteotomies dominate the outcome data, and the authors’ own conclusion is that those procedures do well “where nonoperative management failed”.
So be clear about what this page is. It is mechanical reasoning applied to measured shoe geometry. It is not evidence that a shoe changes the course of Freiberg’s disease, because that evidence does not exist. Anyone who tells you a particular shoe treats this condition is telling you something the literature cannot support.
What non-operative treatment actually is, and how long it is supposed to take
The paragraph above is about footwear trials, which do not exist. It is not a statement that the condition has no non-operative literature — it does, and the most useful piece of it is a 2024 treatment algorithm published in Cartilage by a group of foot and ankle surgeons working from the published articles plus specialists’ consensus. It is worth reading it precisely, because it says two things people on this page will want to know.
First, what the aim is. The authors state it plainly: the aim of non-operative treatment is “to relieve pain and minimize metatarsal head deformation.” Not to regrow the head — to stop it flattening further while it settles. And the primary non-operative management they list is: pain medications, activity modification, immobilization, shoe wear modifications, and orthotics. That is the literature naming footwear as part of first-line care. It is not the literature naming a shoe, and the distinction is the whole reason this page prints geometry instead of a winner.
Second, how long. The same algorithm reports that stage I–II disease “usually undergo nonoperative management for at least 6 months,” and puts non-operative success in the early stages at about 60 per cent. Six months is a long time to be inside a shoe decision, which is the practical argument for making that decision on numbers rather than on a ten-minute fitting. The authors are also candid about the other end of it: they write that no clear algorithm for surgical decision-making has been established, and describe the shortening osteotomy — the common operation — as achieving “offloading of the metatarsal head.” Even the surgery is described in the language of offloading. That is the logic this page is borrowing, at a far smaller scale.
And the surgical proportion above needs a counterweight, because read alone it frightens people. A 2026 series in the Journal of Pediatric Orthopaedics followed 19 adolescent patients (21 metatarsals) — 68 per cent female, median age 12.7 years, the second metatarsal involved in 67 per cent, and 90 per cent at Smillie stage I–III. Sixteen of the 19 — 84 per cent — were cleared by their provider to return to sport without an operation, most often after a period in a CAM boot or a cast. Dance accounted for 32 per cent of the group, soccer 27 per cent and track 18 per cent.
Our reading, stated as a reading and not as a finding: a review that gathers every retrievable publication since 1914 is going to be weighted towards the cases somebody wrote up, and operations get written up. A consecutive clinic series does not have that shape. Nineteen patients from one centre is small and we are not going to pretend otherwise — but it is the better answer to “does this always end in surgery” than a century of case reports is.
The one thing a shoe can change: how far the joint has to bend
Walking loads the metatarsal heads at push-off, and it does it while the toe joint is dorsiflexed — bent upwards. A damaged second metatarsal head is being compressed at precisely the angle the joint is least able to tolerate. That gives footwear two levers, and only two:
- Stop the joint bending so far. A forefoot that resists bending, usually combined with a rocker, carries you over the front of the shoe instead of making the toe joint do it.
- Move load off the head. A shoe with more material under the heel than the forefoot shifts pressure rearward, and a metatarsal pad behind the head takes load onto the shaft of the bone instead of its end.
Both are geometry, and geometry is measurable. Neither is a cure. The point of both is to let you walk a mile without provoking the joint while the decision about what to do next gets made.
The shoes we hold measured geometry for, steepest drop first
Heel-to-toe drop is the figure that tracks forefoot offloading most directly: the more material under the heel relative to the forefoot, the further back the load sits. These are the shoes in our library with the steepest drops, and the forefoot stacks are genuinely thin — which is the part that surprises people, because a “cushioned” shoe with a 23 mm forefoot is not cushioning your metatarsal heads at all.
| Shoe | Heel | Forefoot | Drop | Who published these figures |
|---|---|---|---|---|
| Dansko XP 2.0 | 43 mm | 25 mm | ~18 mm | Dansko, converted from its published 1.7 in and 0.98 in; the drop is derived by us, because Dansko publishes none |
| Orthofeet Coral | — | — | 15.7 mm | Orthofeet publishes a drop but no stack heights at all |
| New Balance Made in USA 990v6 | 39.1 mm | 23.8 mm | 15.3 mm | RunRepeat laboratory; New Balance publishes no stack height |
| KURU ATOM 2 | 39.9 mm | 26.0 mm | 13.9 mm measured; KURU states 10 | RunRepeat laboratory; KURU’s own 28 mm figure is cushion thickness, not stack |
| Brooks Addiction Walker 2 | 35.7 mm | 23.5 mm | 12.2 mm | RunRepeat laboratory; Brooks publishes no offset for this shoe |
| Brooks Beast GTS 26 | 38 mm | 28 mm | 12 mm stated; its 38/28 stacks imply 10 | Brooks for the drop, Running Warehouse for the stacks — and the two do not agree |
| Brooks Adrenaline GTS 24 | 39 mm | 27 mm | 12 mm stated; 13.5 measured | Brooks; the measured figure is RunRepeat’s |
| HOKA Transport 2 | 38 mm | 26.8 mm | 5 mm stated; 11.2 measured | RunRepeat laboratory stacks against HOKA’s stated drop — the largest disagreement in our library |
None of these figures is ours, and no shoe on this site has been worn and measured by us yet. Where a brand publishes nothing, the cell says so rather than guessing. The full dataset behind this table, with every source named, is in the measured spec database, and the stated-versus-measured gaps are collected on the drop chart.
Read that last column before you read the numbers. Three of these eight shoes are sold with a drop a laboratory could not reproduce, and in every case the real figure is steeper than the claim — the HOKA Transport 2 by more than 6 mm. A fourth, the Beast GTS 26, has Brooks stating 12 mm while the retailer stack heights in the same row imply 10. For most buyers those discrepancies are trivia. For a forefoot you are trying to protect, they are the whole specification.
The metatarsal pad goes behind the sore head, not under it
This is the single most useful thing on this page, and it is the thing people get wrong. A metatarsal pad works by loading the shaft of the metatarsal so the head carries less. Put it under the head instead and you are pressing directly on the part that hurts.
There is a measurement for this. In twenty people fitted with total-contact inserts, pads placed 6.1 to 10.6 mm proximal to the second metatarsal head — that is, just behind it — reduced peak pressure under the forefoot by an average of 32%. Move the pad outside that window and the reduction fell to 16% and became unpredictable. Move it more than 1.8 mm past the head, towards the toes, and peak pressure went up.
The limitation matters and we are not going to bury it. Those twenty people had diabetes, peripheral neuropathy and previous forefoot ulcers. Nobody has run the same study in Freiberg’s disease. What transfers is the mechanical principle and the placement window; what does not transfer is any claim about outcomes in your condition.
Practically: find the sore knuckle with your thumb, and the pad’s front edge should sit a centimetre behind it. If a pad makes the pain worse, it is almost always too far forward, not too thin.
A second study closes the population gap — and finds the same direction
The honest limitation above — twenty people with diabetes and neuropathy — has since been partly answered by a study in a completely different population. In 2020 a group at La Trobe took 36 community-dwelling adults aged 65 or older with a history of forefoot pain, no diabetes requirement, put them in standardised shoes, and measured in-shoe plantar pressure across seven conditions: no pad at all, and two different prefabricated metatarsal domes each tested in three positions — 5 mm proximal to the metatarsal heads, in line with them, and 5 mm distal. They also built a three-region mask — proximal to the heads, beneath them, distal to them — so that for the first time they could see where the pressure went rather than only that it fell.
Every pad condition significantly reduced peak pressure distal to the metatarsal heads, where the highest readings sit, against no pad at all — by roughly 45 to 60 kPa. But dropping pressure in one place is trivial if you are willing to raise it in another, and that is the trap a pad sets. Only the two proximally positioned conditions reduced pressure distally without significantly raising it proximally, under the bulk of the pad itself. The in-line and distal placements did not manage both. Two studies, two unrelated populations, fourteen years apart, pointing the same way: behind the head.
The part nobody expects: the softer pad won. The two domes differed in hardness, and the paper prints the numbers — the Emsold dome measured Shore A 11, the Langer PPT pad Shore A 20. In the proximal position the softer of the two was the more effective, and the authors’ explanation is mechanical: a softer dome moulds to the shafts behind the heads, so it offloads distally without building a new pressure point under itself. A firmer pad in the right place is still a lump.
Which leaves a hole you cannot shop your way out of. That result is only actionable if you know the hardness of the pad you are buying, and no retail metatarsal pad we can find publishes a durometer figure. The study identifies the property that decides the outcome, and the market declines to print it. We own a durometer; that measurement is on our bench list, and when it exists this section will name products instead of a principle.
One caveat travels with this one too, and the authors raise it themselves. They measured pressure, not pain, and write that “randomised trials using patient-reported outcome measures are needed” before anyone concludes that forefoot pads reduce forefoot pain. The participants were also 65 and older — not the adolescents and young adults Freiberg disease most often lands on. What transfers is where a pad belongs, not a promise about how you will feel.
What to check, in order
- Bend the shoe. Hold the heel and push the toe up. If it hinges at the ball of the foot, it will make your toe joint do the work. You want it to resist, or to roll.
- Look at the drop, not the cushioning. A soft 4 mm-drop shoe puts more load through the forefoot than a firm 12 mm one.
- Check the insole comes out. You will probably want a pad in there, and it needs somewhere to go.
- Check the toe box height, not just its width. A collapsing metatarsal head sits dorsally; pressure from above is a real source of pain in this condition and almost nobody checks for it.
What to stop wearing
- Anything that folds in half. Minimal shoes, most flats, and the flexible-forefoot trainers sold as “natural” all force the joint through its full range at every step.
- Heels above roughly two inches. A raised heel sounds like offloading and is the opposite: it tips load forward onto the metatarsal heads and holds the toe joints extended.
- A metatarsal pad you positioned by feel, under the sore spot. See above — that placement measurably raises pressure.
What we cannot tell you
We do not publish a forefoot stiffness figure for any shoe, and stiffness is half of the mechanism described above. Stack heights and drop we can source; how hard a given sole resists bending we cannot, because no brand publishes it and we have not yet measured it ourselves. So the table above ranks shoes on the lever we can quantify and is silent on the one we cannot. Treat the hand test as the stiffness measurement until we have a better one — and if you see a page that ranks shoes for Freiberg’s on “stiffness” without saying where the number came from, there is no number.
When footwear stops being the answer
See a foot and ankle specialist rather than adjusting your shoes if the joint is swelling repeatedly, if it is stiffening or grinding, if night pain has started, or if you have been managing it with footwear for three months without the pain settling. Freiberg’s is graded on imaging, and the grade changes what is worth trying — a flattened head with preserved cartilage is a different problem from a collapsed one. The literature above is overwhelmingly surgical for a reason: in the published cases, footwear and inserts were what came before the operation, not instead of it.
The general-education version of the forefoot argument, if you would rather hear it than read it. It is about metatarsalgia rather than Freiberg’s disease specifically, because that is the complaint most people arrive with before anybody has looked at an x-ray — but the footwear habits it describes are the ones that keep loading the joint this page is about.
Questions patients ask
Will a stiff-soled shoe heal Freiberg’s disease?
No. It reduces how much the damaged joint is asked to bend and how much load passes through it. Whether that changes the natural history of the disease has never been tested.
Is a rocker sole or a stiff sole better?
They are usually the same intervention. A stiff forefoot stops the joint bending; a rocker gives you somewhere for that motion to go, so you do not simply trip over your own shoe. Most shoes that work for this have both.
Where exactly does the metatarsal pad go?
Just behind the sore knuckle — the measured window is 6 to 11 mm proximal to the metatarsal head. In front of it, pressure rises.
Does it only affect teenagers?
No. In the largest published series, adults outnumbered under-18s by more than two to one. The onset may well be in adolescence; the presentation often is not.
Can I keep running?
That is a question for the person who has seen your imaging, not for a shoe page. Mechanically, running multiplies the forefoot load that this condition cannot tolerate.
Is a carbon plate the same thing as a stiff sole?
Functionally it is the strongest version of it, which is why carbon-plated inserts are used for stiff-toe conditions. We have a page on carbon fibre inserts that covers what they do and do not do.
What happens next on this page
The gap here is forefoot stiffness. When we start measuring shoes on our own bench, a bend-resistance figure for the shoes in the table above is the number this page needs most, and it is the one that would let us rank rather than list. We will also add a measured toe-box height if we can find a repeatable way to take it, because dorsal pressure on a collapsing metatarsal head is the complaint we hear most and the one no spec sheet addresses.
Sources
- Rehm A, Seah M, Chase HE, Gompels B, Newton AC, Fawi HMT, Ahmed M, Hatzantoni K, Butt J, Ashby E. Freiberg’s disease: variation of surgeries, outcomes, and first population-based incidence. Journal of Pediatric Orthopaedics B. 2025;34(5):488–497. doi:10.1097/BPB.0000000000001243 (PMID 40014303). Source of the incidence figure of 1 in 2,833 livebirths, the 939-of-1,169 surgical proportion and the adult-majority age split.
- Hastings MK, Mueller MJ, Pilgram TK, Lott DJ, Commean PK, Johnson JE. Effect of metatarsal pad placement on plantar pressure in people with diabetes mellitus and peripheral neuropathy. Foot & Ankle International. 2007;28(1):84–88. (PMID 17257544). Source of the 6.1–10.6 mm proximal placement window, the 32% average pressure reduction and the finding that pressure rises when the pad sits distal to the head. Population: 20 adults with diabetes, peripheral neuropathy and prior forefoot ulceration — not Freiberg’s disease.
- RunRepeat laboratory measurements of stack height and drop, per shoe, read September 2026. Individual figures are linked to each shoe’s page in the table above.
- Brand-published stack heights, drops and prices, taken from each manufacturer’s current product listing, September 2026.
- Yoshimura I, Takao M, Wagner E, Stufkens S, Dahmen J, Kerkhoffs GMMJ, Glazebrook M. Evidence-Based Treatment Algorithm for Freiberg Disease. Cartilage. 2024;15(1):58–64. doi:10.1177/19476035231205676 (PMID 37815268). Open access. Source of the stated aim of non-operative treatment, the list naming shoe-wear modification and orthotics in primary management, the “at least 6 months” window for stages I–II, the approximately 60 per cent early-stage success figure, the description of shortening osteotomy as offloading the metatarsal head, and the authors’ own statement that no clear surgical algorithm has been established. Built from published articles plus specialists’ consensus, not from a trial.
- Touban BM, Amaral JZ, Chhabra BN, Miggins J, Kushare I. Provider-Cleared Return-to-Sport After Freiberg Infraction in Adolescent Athletes. Journal of Pediatric Orthopaedics. Published online 20 July 2026. doi:10.1097/BPO.0000000000003420 (PMID 42474446). Source of the 19-patient, 21-metatarsal series, the 67 per cent second-metatarsal involvement, the 90 per cent Smillie I–III distribution, the sport breakdown and the 16-of-19 (84 per cent) return to sport without surgery. Single centre, nineteen patients — small.
- Landorf KB, Ackland CA, Bonanno DR, Menz HB, Forghany S. Effects of metatarsal domes on plantar pressures in older people with a history of forefoot pain. Journal of Foot and Ankle Research. 2020;13(1):18. doi:10.1186/s13047-020-00388-x (PMID 32375847). Open access. Source of the seven-condition placement comparison, the 45–60 kPa reductions, the finding that only the 5 mm-proximal positions offloaded distally without raising pressure proximally, and the Shore A 11 against Shore A 20 hardness result. 36 adults aged 65 or older; pressure measured, pain not measured, as the authors state.
Nothing on this page is medical advice or creates a doctor–patient relationship. Freiberg’s disease is diagnosed on imaging and graded, and the grade changes what is worth trying — persistent forefoot pain, a swelling or stiffening joint, or night pain should be assessed in person.
Related guides & shoe pages
Recently reviewed
Individual shoe pages state their evidence tier, published measurements and who should skip them.
HOKA Clifton 11
Spec AnalysisHOKA’s product page carries a podiatric seal claim the APMA database does not support for this version, and a women’s stack height that contradicts HOKA’s own drop. Not tested by us.
Skechers Hands Free Slip-ins
Spec AnalysisUntested. A heel panel engineered to stay open cannot also be a rigid heel counter, and Skechers publishes no stiffness figure for it — nor a drop, a weight or a width.
New Balance Made in USA 990v6
Spec AnalysisNew Balance publishes no weight, drop or stack height for the 990v6, but it does publish B through 6E widths — and that range is why it still matters. Not tested by…
Measured, worn, scored — never sponsored
No brand has ever paid for a review, a score, a ranking or the seal. Where we earn a commission the link says so, and the score does not change.