Determining scaphoid waist fracture union by conventional radiographic examination: an analysis of reliability and validity

INTRODUCTION: Conventional radiographic imaging is the first imaging tool of choice in scaphoid fractures. The majority of undisplaced scaphoid waist fractures unite after 6 weeks of cast immobilization. We hypothesized that conventional radiographic imaging at 6 weeks after injury can both accurately and reliably predict union in undisplaced scaphoid waist fractures. MATERIALS AND METHODS: Fleiss' kappa statistics were used concerning the opinions of four observers reviewing 47 sets of good-quality scaphoid radiographs of undisplaced scaphoid waist fractures. As reference standard for union, radiographs were taken at a minimum of 6 months after injury to determine validity. RESULTS: Overall agreement was defined as moderate. (kappa = 0.583) "No consolidation" (kappa = 0.816), "full consolidation" (kappa = 0.517) and "partial consolidation" (kappa = 0.390) were defined as good, moderate and fair agreement, respectively. The average sensitivity and specificity of diagnosing scaphoid waist fracture union on standard scaphoid radiographs were 0.65 and 0.67, respectively. The positive predictive value for diagnosing union was 0.93 and the negative predictive value was 0.22. CONCLUSIONS: Conventional radiographic imaging is accurate and moderately reliable in diagnosing union, and reliable but inaccurate in diagnosing nonunion of scaphoid waist fractures at 6 weeks follow-up.