Midfoot Osteoarthritis in Runners: What Does the Evidence Actually Say About Orthoses, Stiff Inserts, and Injections? | Find Your Stride | Edinburgh Podiatrist
Introduction - Midfoot Osteoarthritis
If you've ever treated or run through a nagging, dorsal midfoot ache that flares on hard surfaces and hill work, there's a decent chance midfoot osteoarthritis (OA) is on the list of differential diagnoses. It's a common source of foot pain and disability, affecting roughly 1 in 8 adults over 50, and yet clinicians managing it are working with almost no solid evidence to guide treatment decisions.
A 2023 systematic review published in Rheumatology International by Lim and colleagues set out to fix that gap pulling together every trial evaluating non-surgical treatment of midfoot OA. The result is a useful, well-conducted review of a genuinely weak evidence base. That combination is worth unpacking, because the headline conclusions ("orthoses may help," "injections work short-term") are easy to skim past without appreciating just how shaky the underlying data really is.
What the authors actually did
The authors searched five databases and a clinical trials registry without date or language restrictions, screening 803 records down to six eligible trials covering 231 participants in total. That's the entire published evidence base for non-surgical midfoot OA management — six trials, worldwide, ever. Of those six: one was a randomised feasibility trial with a control group (37 participants), and five were uncontrolled case series (ranging from 6 to 63 participants). Every single trial was rated poor quality using the NIH Quality Assessment Tools. Three broad interventions were covered:
Arch contouring foot orthoses (2 trials)
Shoe stiffening inserts (2 trials)
Image-guided intra-articular corticosteroid injections (2 trials)
No trials examined NSAIDs, physical therapy, footwear modification alone, or bracing, despite these being commonly used first-line strategies in clinical practice.
What the authors found
Arch contouring orthoses showed effects ranging from negligible to large on pain, and medium to very large on function, depending on the trial. The catch: the trial reporting the biggest effects (Ibuki et al., a 57-person case series with no control group) had 36 of its 57 participants also fitted with rigid carbon fibre footplates in their shoes alongside the orthoses — meaning the orthotic effect and the stiffening effect are essentially tangled together in that dataset.
Shoe stiffening inserts produced a similarly messy picture. In one small case series, stiff carbon fibre inserts did nothing for people with mild midfoot OA but produced a huge immediate reduction in pain for those with moderate OA — within a single walking session. A separate trial found only a medium effect on pain and no meaningful effect on function at four weeks. Two trials, two different insert designs, two different magnitudes of effect, and no way to reconcile them.
Corticosteroid injections followed the pattern seen in other foot and ankle conditions: meaningful short-term pain relief (up to 12 weeks) that faded substantially by the medium and long term. In one retrospective series, just over half of participants reported improved pain at 1–3 months, but that dropped to 10% by 6–12 months and 2.5% beyond a year. Adverse events were minor (flare reactions, transient post-injection swelling) and uncommon.
Across all three interventions, nobody measured health-related quality of life, and only three of the six trials used a validated pain or function outcome measure at all.
The critical read: why "may be effective" is doing a lot of work
This is where the review earns real credit — the authors are unusually candid about the limitations of their own conclusions, and clinicians should read the discussion section as closely as the abstract. A few things stand out:
There's no agreed definition of midfoot OA. Diagnostic criteria varied across nearly every trial, some relied on radiographs, some on MRI, some on bone scan uptake, some on clinical presentation alone. That means the "midfoot OA" in one trial may not be directly comparable to the "midfoot OA" in another, which undermines any attempt to generalise findings across the pooled evidence.
Nothing was blinded. Assessor blinding was absent in five of six trials. Only the feasibility RCT attempted blinding, and even there, the clinician fitting the orthoses obviously knew which device was which — an unavoidable limitation of any orthotic trial, but one that still matters for a condition with a well-documented placebo response to foot orthoses.
Sample sizes were tiny and mostly unpowered. Only the feasibility trial reported a prospective sample size calculation. Everything else, including the trial with the most dramatic effect sizes was a small, uncontrolled case series vulnerable to regression to the mean, natural symptom fluctuation, and selection bias.
Outcome measures were inconsistent. Different pain scales (NRS, VAS, FFI-R subscales), different function measures (MFPDI, SEFAS, FFI-R disability subscale), different follow-up windows. This is precisely why the authors couldn't run a meta-analysis, the data simply weren't homogeneous enough to pool.
Confounded interventions. As above, the strongest orthotic result is contaminated by a co-intervention (rigid footplates) in the majority of participants, which the original authors themselves flagged but which is easy to lose in a summary abstract.
What this means for clinical practice — and for runners
For clinicians, the honest takeaway is this: arch contouring orthoses, shoe stiffening inserts, and corticosteroid injections are all reasonable, low-risk options to trial in midfoot OA, and with what we know works for other lower-limb OA presentations. But none of them are supported by evidence that would meet the bar clinicians would expect for, say, an ACL rehab protocol or a bone stress injury return-to-run plan. There is no trial evidence at all for NSAIDs, manual therapy, taping, gait retraining, or graded loading in this population, despite these being staples of conservative sports podiatry management.
For runners and active patients with midfoot OA, particularly those trying to maintain training volume, the practical implication is that a trial of a stiffer-soled shoe or a carbon-plated insert is a sensible, evidence-informed first step. Not because the RCT evidence is strong, but because the risk is low, the mechanistic logic is sound, and it's genuinely the best data we have. Corticosteroid injections may be useful for getting through a short block of racing or a specific event, but shouldn't be sold as a durable fix. The evidence suggests the benefit largely evaporates within a year.
The real headline: we need better trials
The most important output of this review isn't really "what works" it's the roadmap for what needs to happen next. We need a consensus clinical and radiographic definition of midfoot OA (work already underway via the International Foot and Ankle Osteoarthritis Consortium), adequately powered randomised controlled trials with blinded assessors, standardised validated outcome measures, and critically some data on quality of life and long-term function beyond 12 months.
Lim and colleagues have done the field a service by mapping exactly how much (or how little) we know. It's a call to action for better-funded, better-designed trials in a condition that affects a huge number of active, weight-bearing adults and currently has no clinical guidelines at all.
Find Your Stride!
Citation:
Lim, P. Q. X., Lithgow, M. J., Kaminski, M. R., Landorf, K. B., Menz, H. B., & Munteanu, S. E. (2023). Efficacy of non-surgical interventions for midfoot osteoarthritis: a systematic review. Rheumatology International, 43(8), 1409–1422. https://doi.org/10.1007/s00296-023-05324-3
Comments