A scaphoid fracture is easy to miss, and the reasons are specific enough to be worth understanding in their own right. This is not a story about a rare injury slipping through the net. It is a story about a common injury that closely resembles a much more minor one, examined with tests that are individually imperfect, and investigated with an X-ray that is genuinely unreliable for this particular bone. Each of these problems is well documented in the published literature, and together they explain why a wrist injury dismissed as a sprain can, months or years later, turn out to have been a fracture all along.
Scaphoid fractures happen overwhelmingly by one mechanism: a fall onto an outstretched hand. The wrist extends, and the scaphoid, which sits at the base of the thumb, is very often the first structure to absorb the impact. This is precisely the same mechanism that causes the far more common and far more minor injury of a simple wrist sprain, and in the first hours after injury, the two can look almost identical. Both cause pain, swelling, and reduced movement around the same part of the wrist. Without a specific examination aimed at the scaphoid itself, there is little in a patient's initial presentation that reliably separates one from the other.
Doctors use a few simple hands-on checks to work out how likely a scaphoid fracture is, and each of these checks has been tested in real patients to see how well it actually works. When researchers test a sign like this, they measure two separate things: how often it correctly picks up a real fracture, and how often it correctly tells someone without a fracture that they do not have one. A test can be excellent at the first of these and still be quite poor at the second, which turns out to matter a great deal for the scaphoid.
Tenderness in the anatomical snuffbox, the small hollow visible on the back of the wrist at the base of the thumb, is the best-known of these checks. A study of 221 patients with a suspected scaphoid injury found that this sign correctly picked up around 90 percent of real fractures. However, the same study found that it also flagged up a fracture in a lot of people who did not actually have one, correctly clearing only around 40 percent of people without a fracture. In other words, a positive result here is a strong reason to investigate further, but on its own it does not tell a doctor very much either way.
Tenderness felt over the scaphoid tubercle, a spot on the palm side of the wrist rather than the back, picked up a very similar proportion of real fractures, around 87 percent, but did meaningfully better at correctly clearing people who did not have one, around 57 percent. The same study found that a third check, gently pushing along the length of the thumb toward the wrist, performed the best of the three in that particular group of patients, though later studies have found this test to be less consistent.
None of these checks is perfect on its own. What the research consistently shows is that no single sign is enough, by itself, to either confirm or rule out a fracture. This is why a proper scaphoid examination checks several of these signs together rather than relying on just one: a patient with none of them present is genuinely unlikely to have a fracture, while a patient with several of them present needs an X-ray regardless of how minor the injury looks.
What each check found, out of 100 people tested:
Source: Grover R. "Clinical Assessment of Scaphoid Injuries and the Detection of Fractures." Journal of Hand Surgery (British and European Volume). 1996;21(3):341–343. Figures reflect a single prospective cohort of 221 patients and are not a guarantee for any individual case.
Even where clinical suspicion correctly leads to an X-ray, the X-ray is a genuinely difficult test to interpret for this specific bone. The scaphoid has an irregular, curved shape, unlike the straight profile of many other bones, and a fracture line is only clearly visible when the X-ray beam happens to pass almost exactly parallel to the break. This is why a dedicated scaphoid series uses four separate angled views rather than the usual two, each attempting to catch the fracture line from a slightly different direction. Even with this dedicated protocol, a pooled analysis drawing on 2,507 patients across 42 published studies found that 21.8 percent of scaphoid fractures were missed on the initial X-ray series and only identified afterward, using more advanced imaging.
Some fractures also become genuinely easier to see over the following one to two weeks, as the broken bone edges resorb slightly and the gap becomes more visible. This is the basis for a well-established practice: a patient with a normal X-ray but a clinical presentation that still raises suspicion is often placed in a splint and asked to return for a repeat X-ray after this interval, rather than being discharged outright on the strength of a single normal film.
Where suspicion remains after a normal X-ray, further imaging is the appropriate next step, and the same pooled analysis directly compared how well the main options perform against each other, using the same two measures explained above: how many real fractures each test catches, and how many people without a fracture it correctly clears.
How the imaging options compare:
Source: Bäcker HC, Wu CH, Strauch RJ. "Systematic Review of Diagnosis of Clinically Suspected Scaphoid Fractures." Journal of Wrist Surgery. 2020;9(1):81–89. Figures reflect pooled published data across 2,507 patients and are not a guarantee for any individual case.
MRI's advantage lies in what it is actually detecting. Rather than relying on a visible break in the bone's outer surface, MRI identifies bone oedema, a build-up of fluid within the bone that behaves like an early bruise and is present even before a fracture line becomes visible on any other form of imaging. This is why NICE guideline NG38 specifically recommends MRI as the preferred next investigation where clinical suspicion of a scaphoid fracture persists despite a normal X-ray.
Given how unreliable X-ray is for this bone, it would be reasonable to assume the solution is simply to image every patient with a suspected wrist injury using MRI from the outset. A detailed review of 52 medical negligence cases involving missed scaphoid fractures tested this assumption directly, and the finding is more specific than "imaging needs to improve."
In 41 of the 52 cases reviewed, the recorded mechanism of injury, a fall onto an outstretched hand, was entirely consistent with a possible scaphoid fracture. Despite this, there was no record of an examination for scaphoid tenderness at all in 37 of the 52 cases, and in a further 10, tenderness was checked for and specifically not found. In total, a scaphoid fracture was never considered, or was actively excluded on clinical examination, in 49 of the 52 cases. Applying a policy of MRI for every patient with a normal X-ray and any suspicion of scaphoid injury would only have identified 3 of these 52 missed fractures, because in the other 49, no imaging beyond a possibly normal X-ray was ever requested in the first place. The diagnostic failure, in the great majority of cases actually reviewed, happened earlier than any imaging decision: in recognising the mechanism of injury as significant, and in performing and properly recording a scaphoid-specific examination.
The consequences of this pattern are measurable in more than clinical terms. A review of UK National Health Service Litigation Authority data covering scaphoid fracture mismanagement claims between 1995 and 2010 identified 85 closed cases, of which 57 related specifically to missed fractures, the single largest category of complaint. The mean cost per case was £41,680, with individual cases ranging up to £206,789, and a cumulative cost across the full 15-year period of £3,542,855. This scale of cost reflects the same underlying pattern identified in the negligence case review above: the majority of these claims relate to a fracture that was never diagnosed at the point it should have been, not to a technically difficult case that genuinely defeated available imaging.
A wrist injury that has not settled within the timeframe expected for a simple sprain, particularly one following a fall onto an outstretched hand, warrants specific reassessment for a possible scaphoid fracture, even where an initial X-ray was reported as normal, and even where the original injury happened some time ago and was labelled as a sprain at the time. Persistent pain on the thumb side of the wrist, reduced grip strength, or ongoing discomfort with gripping or pushing movements are all reasons to seek this reassessment rather than assume the diagnosis given at the time of injury was necessarily correct.
"If the X-ray was normal, there is no fracture."
Published data puts the initial X-ray miss rate at close to 22 percent when compared against more advanced imaging. A normal X-ray reduces the likelihood of a fracture considerably but does not rule one out on its own, particularly where clinical examination findings point the other way.
"A doctor pressing on my wrist and finding no tenderness means it is definitely not broken."
Individual hands-on checks are imperfect. Tenderness over the scaphoid tubercle, for example, only picks up around 87 in 100 genuine fractures, meaning roughly one in eight real fractures will not produce this particular sign. This is precisely why several checks are used together rather than relied upon individually.
"Scanning every wrist injury with MRI would solve the missed diagnosis problem."
The most detailed available data on missed scaphoid fractures found that in 49 of 52 reviewed cases, the fracture was never suspected or examined for in the first place, meaning no imaging was ever requested. Blanket MRI cannot compensate for a mechanism of injury that was never recognised as significant at the point of initial assessment.
"A wrist injury from years ago that was called a sprain could not still be an old undiagnosed fracture."
This is one of the more common ways an old scaphoid fracture eventually comes to light. Because the initial presentation of a fracture and a sprain can look very similar, and because a proper scaphoid-specific examination is not always performed, a wrist injury that never fully settled after being dismissed as a sprain is a reasonable prompt for reassessment regardless of how long ago it happened.
Why does a scaphoid X-ray need four views instead of the usual two?
Because of the scaphoid's curved, irregular shape, a fracture line is only clearly visible on X-ray when the beam passes nearly parallel to the break. A dedicated scaphoid series uses four differently angled views specifically to improve the chance of catching the fracture line, though even this protocol still misses a significant proportion of fractures.
If my wrist X-ray was normal but I still have pain, what should happen next?
Where clinical suspicion remains, the appropriate next step is either a period of splinting with a repeat X-ray after one to two weeks, or further imaging, most often MRI. NICE guidance specifically supports MRI as the preferred investigation in this scenario rather than discharge on the strength of one normal X-ray alone.
Which single physical examination test is most reliable?
None is reliable enough to be used alone. Tenderness in the anatomical snuffbox is the best of the individual checks at catching real fractures, but it also wrongly flags up a lot of people who turn out not to have one. The most accurate approach combines several checks together, since a patient with none of the recognised signs present is genuinely unlikely to have a fracture.
Why are so many scaphoid fractures missed if X-ray and examination tests are both well established?
The most detailed available data suggests the core problem is rarely a technical limitation of a specific test. In the majority of reviewed cases, the fracture was never specifically examined for at all, because the injury was assumed to be a simple sprain. The tests themselves perform reasonably well when they are actually used.
Is a missed scaphoid fracture a common reason for medical negligence claims?
Yes. A review of 15 years of UK litigation data found missed scaphoid fractures to be the single largest category of claim relating to scaphoid mismanagement, with a mean cost of £41,680 per case, reflecting how frequently and how significantly this particular diagnosis is missed in practice.
A wrist injury that has not settled as expected, whether recent or long-standing, warrants proper assessment by a hand and wrist specialist rather than an assumption that the original diagnosis was necessarily correct. What happens when a scaphoid fracture is confirmed late, and how a fracture that has already failed to heal is assessed and treated, is covered in full in the companion article on scaphoid nonunion. iiS Health, The Cheshire Hand Clinic, provides scaphoid assessment, including on-site X-ray, with MRI or CT arranged where required.
.png)