OBJECTIVES

Scaphoid fractures are the most common carpal fracture, representing 70% of carpal bone fractures. The diagnostic accuracy of physical examination findings and emergency medicine (EM) imaging studies for scaphoid fracture has not been previously described in the EM literature. Plain x-rays are insufficient to rule out scaphoid fractures in a patient with a suggestive mechanism and radial-sided tenderness on physical examination. This study was a meta-analysis of historical features, physical examination findings, and imaging studies for scaphoid fractures not visualized on plain x-ray in adult emergency department (ED) patients, specifically to address which types of imaging tests should be recommended in patients with persistent concern for acute fracture after ED discharge.

METHODS

A medical librarian and two emergency physicians (EPs) conducted a medical literature search of PUBMED and EMBASE. The original studies' bibliographies were reviewed for additional references and unpublished manuscripts were located via a hand search of EM research abstracts from national meetings. All abstracts were independently reviewed by the two physicians, and Quality Assessment Tool for Diagnostic Accuracy Studies (QUADAS-2) was used to assess individual study quality. When two or more qualitatively similar studies were identified, meta-analysis was conducted using Meta-DiSc software. Primary outcomes were sensitivity, specificity, and likelihood ratios (LRs) for predictors of scaphoid fracture detected on follow-up in patients with normal ED x-rays.

RESULTS

A total of 957 unique citations were identified, yielding 75 studies eligible for inclusion in this systematic review. Studies were significantly heterogeneous in design, study population, and criterion standard. The majority of studies were conducted in non-ED settings (e.g., orthopedic clinics). No studies used accepted diagnostic research publishing guidelines, and the overall QUADAS-2 methodologic quality was low, indicating an increased risk of bias in the estimates of diagnostic accuracy. The prevalence of scaphoid fractures ranged from 12% to 57% with the point estimate of 25% pretest probability for adult ED patients with concern for scaphoid injuries, nondiagnostic index x-rays, and scaphoid fractures on later imaging studies. Except for the absence of snuffbox tenderness (LR- = 0.15), physical examination findings lack accuracy to rule in or rule out scaphoid fractures, and no validated clinical decision rules exist. In patients with persistent concern for injury, magnetic resonance imaging (MRI) is superior to bone scan, computed tomography (CT), or ultrasound (US) to both rule in and rule out scaphoid fractures. Both MRI and CT share the added benefit of identifying alternative etiologies for posttraumatic wrist pain.

CONCLUSIONS

Except for the absence of snuffbox tenderness, which can significantly reduce the probability of scaphoid fracture, history and physical examination alone are inadequate to rule in or rule out scaphoid fracture. MRI is the most accurate imaging test to diagnose scaphoid fractures in ED patients with no evidence of fracture on initial x-rays. If MRI is unavailable, CT is adequate to rule in scaphoid fractures, but inadequate for ruling out scaphoid fractures.