Well, whilst volumes have been written on this topic, here are some nuts and bolts that might help you have a sensible discussion with your patient about what to do.
- Not all scaphoid fractures are created equal. About 20% of them are distal fractures or only involve the tubercle and seem to heal up about 100% of the time. It is highly questionable whether anything needs to be done about these (i.e. immobilisation) at all. On the other side of the bone, proximal pole fractures occur in <5% but have a much higher non union rate (30-50%) - so these deserve special attention.
- Undisplaced fractures visible on plain film in the waist of the scaphoid heal up well in about 90% of cases with immobilisation.
- Fractures you can't see on plain films are, by definition, undisplaced.
- Plain films miss about 5-10% of scaphoid fractures. The upper end of the scale is for MRI diagnosis, the lower end is plain film diagnosis. Much higher rates than these are commonly reported, but are usually from studies biased towards high fracture rates, not average ED patients.
- About 30-50% of patients reviewed at 5-7 days following injury have no pain or tenderness, and need no treatment or investigation. Delaying MRI until this time will increase your pick up rate for fractures.
- MRI picks up at least as many minor fractures of other bones and ligamentous injuries as it does scaphoid fractures!
- If a fracture is only visible on MRI, then how significant is it? It is clearly non displaced, so should have a good prognosis anyway (if it does not involve the proximal pole). MRI fractures appear to have the same distribution as plain film fractures, but unfortunately I haven't yet found any reports on the natural history of them.
- A delay in immobilisation of suspected fractures that turn out to be true fractures of up to 4 weeks does not appear to have a significant effect on outcome.
- Immobilisation of confirmed fractures in a standard forearm cast (without thumb immobilisation) appears to have the same outcomes, but less inconvenience for the patient. If this is the case, then I can't see any reason why a simple wrist splint wouldn't be equally good.
- Immobilisation may prevent an undisplaced fracture becoming displaced in certain patients (occupation, intoxication) - but there doesn't really appear to be much information on this. I very much doubt that lack of immobilisation is likely to lead to displacement of a fracture that cant even be seen on plain films.
- Immobilisation may help the healing of other occult injuries, although I suspect that nearly all of them will heal up anyway.
- Studies of clinical tests for scaphoid fracture have not reported the ability of clinical signs to detect the different fracture types. My experience is that the axial compression test is more likely to be positive in patients with fractures that have a greater need for immobilisation (proximal pole or waist) than tubercle only fractures - so I get more worried if it is positive and less so if it is negative.
- MRIs for suspected scaphoid fractures can now be ordered by GPs in Australia - but they cost about A$450 compared to $50 for plain films.
So the maths goes something like this:
If the missed fracture rate on plain films is 5% but only 80% of fractures are at risk of complications then only 4% of patients are at risk. At this point you will need to immobilise about 96% of patients unnecessarily and do 25 MRIs to find one injury worth treating.
If you wait a week only 2% of patients will be at risk (the others will be better!), but 8% of persistently symptomatic patients will be (number to MRI now 12.5).
So I suggest for suspected scaphoid fracture with normal plain films for most patients
- no immobilisation of patients at low risk for re-injury
- review by GP/OPD in 1 week
- suggest investigation of symptomatic patients at 1 week with MRI
- simple cast or wrist splint for confirmed, undisplaced fractures.
- referral for surgery for any displaced fractures (and possibly proximal pole ones)
Free our innocent scaphoids from unnecessary incarceration!