What Works
Strict non-weight-bearing immobilisation, minimum 6 weeks HIGH
Usually 6 to 8 weeks in a below-knee cast or boot, with crutches. This is the single highest-value decision in the whole condition, and it is the strongest evidence in the literature.
Non-weight-bearing cast: 86% returned to sport (19 of 22). Rest with continued walking: 26% (9 of 34). p<0.001 (Khan 1992, 86 CT-proven fractures). Independently, pooled analysis found the walking version significantly worse than both the cast and surgery (Torg 2010).
Early imaging, with a CT scan to stage it HIGH
A bone scan or MRI to find it, then a CT to see how far the crack has travelled. The CT is what decides the treatment.
Average delay before diagnosis is 8.8 to 9.9 months (Attia 2021; Saxena 2017). In a 320-athlete series, tarsal bones took the longest of any site both to diagnose and to recover (Matheson 1987).
Surgery (screw fixation), for defined reasons MODERATE
Indicated for displaced or complete fractures, fractures reaching a second surface of the bone, hardened or dead bone at the fracture edges, failed proper casting, and repeat fractures in young athletes.
Lowers the chance of re-breaking it: 1.28% vs 23.53% (Attia 2021), supported independently by zero re-fractures among previously operated patients (Saxena 2017). It does not reliably get people back faster LOW — see The Nuance.
Staged rehabilitation after the cast MODERATE
Four blocks of two weeks, with a check-up at the end of each one deciding whether you move on or repeat the block.
15 of 17 athletes returned to their previous competitive level; 24 weeks on average to full sport, range 17 to 32 (Bojanic 1997). The only staged protocol in the literature.
Bone health and fuelling check MODERATE
For runners, teenagers, and anyone in a sport where leanness matters.
43% of male athletes with lower-limb bone stress injuries had low bone density; runners carried 6.1 times the risk of non-runners (Tenforde 2018).
Tier 3 — Emerging (open)
Vascularised bone grafting for the hardest surgical cases: 100% healing, against 80% for screws alone and 75% for screws with a standard graft. No difference in return to sport between the groups (Nunley 2022, 43 patients, single centre).
Ultrasound pain-threshold monitoring as a way of tracking recovery objectively in elite athletes (Malliaropoulos 2017, 10 patients). Interesting because it attempts to solve the real problem here, which is that symptoms mislead. Not established.
Shock-absorbing insoles reduced stress fracture rates across four trials (Cochrane). That is prevention of lower-limb stress fractures generally, not treatment of this one, and the review's search closed in 1997.
Exercise Prescription
| Exercise | What it's for | Dose | Pain guide |
|---|---|---|---|
| Anything that loads the injured foot | Not prescribed, and that is deliberate. There is no published, tested exercise or return-to-running programme for this injury, and the safe timeline depends on your fracture type and whether you have had surgery. That progression has to come from the clinician who has seen your CT. | ||
| Upper body and core, seated or lying | Keeping your strength while the foot heals | As normal, 2–4× per week | No foot pain at all, during or after |
| Good leg and both hips, non-weight-bearing positions | Losing as little as possible on the other side | As normal, 2–3× per week | No foot pain at all, during or after |
| Arm-bike conditioning | Keeping your heart and lungs going | 20–40 min, 3–5× per week | No foot pain at all, during or after |
What Doesn't Work
- "Rest from running but keep walking." The most commonly given version of rest and the worst-performing treatment on record: 26% back to sport against 86%. This is not a mild under-treatment, it is the difference between most people getting back and most people not.
- Being reassured by a normal X-ray. It catches as few as 12% of these.
- Using the scan alone to decide you are ready. Scan healing lags behind real recovery, and the original researchers said so explicitly.
- Using tenderness on the top of the foot as your sign you are better. It persists in people who have competed successfully for two or more years (Potter 2006).
- Borrowing the return-to-running rules used for shin and forefoot stress injuries. Those are built for low-risk bones where pain is a fair guide. Here the pain settles first, so those rules give permission to run on a crack that has not healed.
Return to Training
- Full immobilisation period completed (minimum 6 weeks non-weight-bearing, usually 6 to 8)
- Healing confirmed by the treating clinician, understanding that the scan lags behind and should not be the only gate
- No tenderness on the top of the foot — necessary, but explicitly not enough on its own
- Pain-free full weight-bearing walking for at least 2 straight weeks
- Each 2-week progression block passed at a formal check-up, with no block repeated for symptoms
- Pain-free single-leg heel raise and single-leg hop, cleared by the treating clinician
- No pain during, straight after, or 24 hours after each new loading stage
- Bone health and fuelling check completed where indicated
- Return to running approved by the clinician holding the scan, not self-directed on how it feels
Red Flags — Get Seen Urgently
- The diagnosis itself is the red flag. This is one of a small number of "high-risk" stress fractures, because the bone has strong pulling forces across it and a poor blood supply. It needs a specialist opinion and imaging, not physical therapy alone.
- Load-related pain across the top of your midfoot with a normal X-ray, in someone who runs or jumps. A normal X-ray catches as few as 1 in 8 of these. It rules nothing out, and being reassured by one is a major reason these get missed for the better part of a year.
- Symptoms that have not improved after weeks of "rest" that involved walking normally. That is a reason to change the plan and get imaged, not a reason to rest for longer.
- Pain at night, pain at rest, fever, unexplained weight loss, or a lump. This is not a stress fracture pattern and needs looking at promptly.
Refer to: an orthopaedic foot and ankle surgeon or a sports medicine physician, for imaging and staging. Onward to a bone health service where fuelling or bone density is a concern.
If a scan has confirmed this, get completely off the foot as instructed. Including around the house.
In the published research, athletes who stayed non-weight-bearing got back to their sport 86% of the time. Athletes who rested from sport but kept walking on it got back 26% of the time. Same injury, same bone, one decision. If you have not been scanned yet and this pain is load-related and has been building for weeks, the action is to book that scan and stop running today.
Conviction
MODERATE
Split by claim, because the evidence here is not one block. Non-weight-bearing beats walking rest HIGH. Slowest-healing stress fracture site in the body HIGH. Diagnostic delay is the dominant problem HIGH. Surgery lowers re-fracture MODERATE. Surgery is better first-line LOW. Surgery gets you back faster LOW. Any specific return-to-running programme: no conviction at all, because no data exists.
What would change my mind on "non-weight-bearing beats walking rest"
A prospective study in which patients were randomly assigned to a non-weight-bearing cast or to activity restriction with continued walking, with the walking group's adherence actually monitored, showing no meaningful difference in return to sport. The current evidence is two independent analyses agreeing on a large effect in the same direction, but neither randomised anyone, and it is possible that the people put in casts differed systematically from those told to take it easy.
What would change my mind on "surgery is not clearly better first-line"
A randomised trial of at least 120 athletes with confirmed type I or II fractures, comparing protocolised, adherence-monitored non-weight-bearing casting against primary screw fixation, with return to pre-injury sport at 12 months and re-fracture at 5 years. That single trial would settle whether the surgical advantage survives when the comparison group is done properly, which is the question the entire literature currently begs.
Don't want to guess which injuries are the serious ones? The Verdict breaks down one condition a week, with the evidence and the honest gaps.
The Full Picture — Anatomy, Diagnosis & Evidence
What's Actually Going On
Bone is living tissue in constant turnover. Repeated loading creates tiny areas of damage, and the bone clears them out and lays down new bone in their place. A bone stress injury is that balance failing: damage builds up faster than it can be cleared.
The navicular sits at the top of the arch, the keystone of the middle of the foot. Two things make it dangerous rather than merely annoying, and they are the same two things shared by every high-risk stress fracture in the body: strong pulling forces across it, and poor blood flow. The middle third of the bone is a relative dead zone for blood supply, which is why this one heals slowly, unreliably, and sometimes not at all.
The crack starts at the top surface and works its way down and across. That is exactly what the CT scan measures when it "stages" the injury: type I is a break in the top surface, type II runs into the body of the bone, type III reaches a second surface. It matters. Type I returned to activity at 3.0 months on average, type II at 3.6, and type III at 6.8.
Here is why the biology matters more than it sounds. In a well-supplied bone, pain settling is decent evidence that healing is going well. In this bone it is not. The pain can settle while the crack is still open, which is precisely why the usual pain-guided approach is unsafe here.
How to Identify It
The presentation is the problem. Patients describe "diffuse, poorly defined symptoms" and there is a striking lack of physical findings, so the first assessment frequently misses it (Lee 2004). It is vague, hard to point to, and walking usually stays comfortable, which is the single most misleading feature of the whole condition.
Ask how long it has been going on. Expect months, and do not treat that as a reason to doubt the diagnosis.
| Test | How good is it at catching this? | How good is it at ruling it out? |
|---|---|---|
| Tenderness on top of the navicular ("N spot") | Never measured | Never measured |
| Single-leg hop | Never measured | Never measured |
| Any other physical test | Never measured | Never measured |
What it is not. Müller-Weiss disease can produce a fragmented navicular that mimics this directly, usually in an older patient. A stress avulsion of the navicular is a separate injury with a small triangular fragment on the top, best seen on a standing side-on X-ray. A Lisfranc injury sits further forward with bruising underneath. And finding an extra small bone on the scan does not rule out a fracture sitting right next to it.
Imaging is where the diagnosis is actually made, and the numbers are humbling. Plain X-ray catches somewhere between 12% and 56% of these. MRI catches 68% to 99%, but how well it rules other things out varies enormously between studies, with a floor of 4%, so a hot midfoot on MRI is not automatically this. A bone scan is a sensitive screen. CT is required to stage it. CT accuracy figures were not available in the evidence reviewed, and have deliberately not been estimated here rather than filled in from a study that was not designed to measure them.
The Debate
There is no clinical practice guideline for this condition as of July 2026. In its place sit two meta-analyses that reach opposite conclusions, and understanding why they disagree is the most useful thing on this page.
Cast or operate?
Honest Limitations
1. The evidence describes athletes, and mostly elite ones
The research: 97% of the fractures in the largest meta-analysis were in athletes, and a third of the largest single group were elite or professional.
The gap: the whole framework, including the acceptability of six to eight weeks off your feet and a four-to-six-month timeline, is built on people whose career or identity justifies that cost. A recreational runner, a self-employed tradesperson, or a parent of young children faces a materially different decision, and no study describes that group.
The adjustment: separate the medical recommendation from the life logistics, and have the hard conversation openly. If staying off the foot genuinely cannot be sustained, that is a real argument to bring to the surgical team, not something to quietly compromise on by walking anyway.
2. Nobody randomised anyone, and the bias runs both ways
The research: every comparison in this literature is retrospective. Who got which treatment was decided by how bad the fracture looked.
The gap: in the adolescent study, the operated group were older, heavier, and had a visible fracture line 88% of the time against 38% in the non-operated group. The surgical group was selected for worse fractures. In older surgical series the bias flips, because surgery was often reserved for people who had already failed casting.
The adjustment: treat published success rates as descriptions of what happened to particular groups of people, not as the effect of the treatment. Stage the fracture and decide on the one in front of you.
3. The rehabilitation half of the pathway is nearly empty
The research: exactly one prospective staged protocol exists in the whole literature, in 17 athletes, and it specifies the stages and the check-ups but not the exercises or the running distances.
The gap: there is an excellent, well-reasoned pain-guided return-to-running framework available, and it is exactly the wrong tool here, because it is explicitly written for low-risk shin and forefoot injuries. At this site the pain settles before the bone heals, so a pain-guided rule gives permission to load a fracture that has not healed. That scope boundary is the original authors' own, not one imposed on them.
The adjustment: progression here is gated on time, imaging and clinical checks, not on symptoms. Where a structure is needed, use the two-week block-and-reassess pattern rather than importing a low-risk protocol.
The Nuance
The honest position on surgery refuses both cheap endings. Not "surgery is superior, the newest analysis says so". Not "surgery is unproven, keep casting".
There are zero randomised trials in this condition. The claim that surgery gets athletes back faster is the weakest part of the case for it, and it is the part patients are most often told: the largest meta-analysis reports a p-value of 0.60 for time to return, which is about as null as a result gets, and two later groups found surgery slower rather than faster. One measured 4.56 months against 3.97 for casting. The other, in 110 teenagers, found the operated group took significantly longer to get back to walking, to running, and to full sport.
The claim that surgery lowers the chance of re-breaking it is the strongest part, it holds up across independent groups, and it deserves real airtime with a young athlete: 23.5% against 1.3%, with zero re-fractures among previously operated patients in a separate cohort. A summary that only found reasons to doubt surgery would be just as biased as the reviews that only find reasons to recommend it.
For a non-displaced type I or II fracture in someone who can genuinely stay off the foot, properly-run non-weight-bearing immobilisation remains a legitimate first choice, and roughly 9 in 10 athletes get back to their previous level by one route or the other. What nobody can honestly tell you, because the trial has never been run, is how much of the surgical advantage survives when the comparison is done properly.
Sources
- Khan KM et al., 1992, Am J Sports Med — 82 athletes, 86 CT-proven fractures. Non-weight-bearing cast 86% return to sport against 26% for activity limitation with continued walking (p<0.001). The largest treatment contrast in the condition.
- Torg JS et al., 2010, Am J Sports Med — meta-analysis. Non-weight-bearing 96% against surgery 82%, not statistically significant; walking rest significantly worse than both. The only analysis that separates the conservative group.
- Attia AK et al., 2021, Int Orthop — meta-analysis, 315 fractures. Surgery 97.9% against conservative 78.1%; return-to-play outcomes null; re-fracture 1.28% against 23.53%; average delay before diagnosis 9.86 months.
- Mehta S et al., 2023, Am J Sports Med — 110 adolescents across 8 centres. Non-surgical treatment succeeded in 85%; the operated group returned significantly later.
- Saxena A et al., 2017, J Foot Ankle Surg — 62 fractures, prospective. Surgery 4.56 months against 3.97 for casting; 91.9% returned to their previous level; symptoms present 8.8 months before treatment.
- Hoenig T et al., 2023, Br J Sports Med — 76 studies, 2974 bone stress injuries. Navicular 127 days to return (95% CI 102 to 151), the longest of any site.
- Bojanic I, Pecina MM, 1997 — 17 athletes, prospective. Six to eight weeks non-weight-bearing then four two-week staged blocks; 24 weeks to full sport, range 17 to 32.
- Potter NJ et al., 2006, Br J Sports Med — 32 fractures at two or more years. No difference between surgery and casting in pain, function or scan appearance; surgical patients more often still tender.
- Wright AA et al., 2016, Am J Sports Med — systematic review of imaging accuracy. Plain X-ray catches 12% to 56%.
- Nunley JA et al., 2022, Foot Ankle Int — 43 operated patients. 100% healing with vascularised graft against 80% for screws alone.
- Tenforde AS et al., 2018, Am J Sports Med — 28 male athletes. 43% had low bone density; runners 6.1 times the risk of non-runners.
- Saxena A et al., 2000, J Foot Ankle Surg — CT staging system. Return to activity 3.0, 3.6 and 6.8 months for types I, II and III.
- Warden SJ et al., 2021, JOSPT — cited here for its scope boundary. Pain-guided return to running is written for low-risk shin and forefoot injuries, not for this one.
- Lee S, Anderson RB, 2004, Foot Ankle Clin — "diffuse, poorly defined symptoms" and a lack of physical findings.
- Matheson GO et al., 1987, Am J Sports Med — 320 stress fractures. Tarsal bones took the longest of any site both to diagnose and to recover.
Full evidence base: 33 sources, every citation verified against the source database. This page is educational and is not a substitute for assessment by a clinician who can see your imaging.