Introduction The scaphoid bone is the most commonly fractured carpal bone. MRI is now recognised to be more sensitive and specific than serial plain films in the diagnosis of occult scaphoid fracture. What is not known is whether it is safe to discharge, without follow-up, those with a normal limited extremity MRI. The primary objective of this study was to determine whether a normal limited scaphoid MRI safely excludes clinically significant injury. The secondary objectives were to determine morbidity and patient experiences of those without fracture on MRI.

Methods This is a cohort study with retrospective case note review and a written follow-up questionnaire conducted in a South of England emergency department soft tissue review clinic.

Results 214 limited extremity MRIs were performed between January 2006 and July 2008 for suspected scaphoid injury. 152 participants were included in the study. 122 (80%) MRIs showed no fracture. In 72 (47%), there were no traumatic findings. The most significant finding in 33 (22%) was bony bruising and joint oedema in 17 (11%). One (0.9%) clinically important injury was missed. There was significant morbidity in those with normal MRIs. Worst-case scenario analysis shows that at least 12.5% of patients with a normal MRI still had symptoms and 14% still had abnormal function a year or more after injury.

Discussion This study shows that limited extremity MRI can safely exclude clinically important injury. Significant symptoms do persist, however, for many patients with a normal MRI.