to the editor: I read with great interest the article by Phillips and colleagues in the September 1, 2004 issue of American Family Physician,¹ and would like to share our experience with a different management technique for patients with clinically suspected scaphoid fracture and negative initial radiographs. As stated in the article,1 there is great variability in the accuracy of initial radiographs to detect scaphoid fractures immediately after injury, and 15 to 65 percent may remain radiographically occult.² The clinical examination also has a great deal of variability in detecting scaphoid fractures, usually showing acceptable sensitivity (around 90 percent), low specificity (rarely above 50 percent), and poor positive predictive value (weighted average of approximately 20 percent).³ Thus, the standard of care suggested for using only positive clinical examination and negative radiographs as a screening tool to justify wrist immobilization may result in unnecessary casting in about one half of patients.⁴ At our institution, we recommend additional imaging in this situation to avoid overtreatment, to decrease the rate of complications, and to increase the quality of life for the patient. Magnetic resonance imaging (MRI) has been advocated as the modality of choice because of its excellent sensitivity and specificity, especially when T1-weighted images are coupled with coronal short tau inversion recovery images. Ultrasonography may be an alternative because it is more readily accessible, less expensive, and less time consuming. We usually suggest limited wrist MRI protocol at presentation of patients with a clinically suspected scaphoid fracture and negative initial radiographs. Practicality naturally precludes the use of MRI as a first-line diagnostic study in some cases. If an experienced radiologist is readily available, ultrasonography may be performed after negative radiographs, and only negative studies referred to early MRI. This approach adds little in terms of financial costs compared with standard immobilization and radiographic follow-up.^3,5,6 When losses of productivity and quality of life caused by unnecessary immobilization are considered, early additional imaging is favored, especially MRI. Opinions may vary because data are subject to interpretation and the availability of imaging equipment is not uniform. The literature on this subject demonstrates the level of continuing debate about this seemingly well-understood clinical problem. Personally, I think that there is sufficient support to encourage more use of early MRI and ultrasonography in patients with clinically suspected scaphoid fracture and negative initial radiographs.

in reply: Dr. Arend makes an excellent point concerning the poor sensitivity of the clinical examination and plain radiographs to diagnose scaphoid fractures. We support his opinion that an early magnetic resonance imaging (MRI) approach can be a preferred method for evaluating patients with a suspected scaphoid fracture and normal radiograph. Additionally, wearing a thumb spica cast for two weeks is problematic for many patients. A number of studies^1–3 have shown that early MRI is a cost-effective approach, and early MRI may soon become the preferred approach.

Nevertheless, it may be premature to completely abandon the traditional approach of casting patients for two weeks and then re-evaluating them. There are two reasons for this. First, in our review of the literature it was suggested, but not specifically determined, that MRI is 100 percent sensitive in detecting occult fractures. Because many of the studies failed to identify a specific “gold standard” when using an MRI to detect fractures, we ask the reasonable question: Will any fractures be missed by an MRI, and what will the clinical consequences be? With the traditional approach, there are theoretically no consequences to missing an early fracture because the patient would be treated empirically with definitive treatment for a nondisplaced fracture. Can the same be said for an early MRI approach in which a patient with a negative MRI would not be casted? We were unable to find any outcome studies that addressed this issue.

Second, the availability of immediate MRI is variable, and practical system issues need to be addressed when introducing a new “standard.” For example, at our institution it can take up to a week to get an outpatient MRI, and MRIs are not available at night or on weekends, so not all patients can be offered an early MRI approach.

We are certainly in agreement that an early MRI approach is reasonable and preferred under many clinical situations, but would not recommend completely abandoning the standard approach.