Back to Annual Meeting Abstracts

Hand and Upper Limb Centre, Western University, London, ON, Canada

Hypothesis: We hypothesize that by using serial computed tomography (CT) scans, we can more accurately predict the natural history of acute isolated proximal pole fractures treated non-operatively and identify risk factors which may lead to non-union and delayed union.

Methods: A radiology database at a tertiary care centre was used to identify all proximal pole scaphoid fractures between 2006 and 2012. Inclusion criteria included acute fractures with CT scans performed within 6 weeks of injury. Exclusion criteria included associated perilunate dislocations or scapholunate ligament injuries and fractures treated with acute surgical management. A retrospective chart review was performed collecting demographic, injury, and treatment characteristics. CT images were reviewed to assess for fracture orientation, displacement (translation between fragments and/or humpback deformity), comminution, cysts, and sclerosis. The association between imaging characteristics and risk of non-union was determined using odds ratios and time to union using a student’s t-test.

Results: There were 40 patients identified with isolated acute proximal pole fractures treated non-operatively. Two patients were lost to follow-up, leaving 38 analyzable patients (mean age 27.3 years, 84.2% male). Based on initial CT scan, there were 9/38 (23.7%) displaced fractures, 7/38 (18.4%) comminuted fractures, 3/38 (7.9%) with sclerosis present, and 10/38 (26.3%) with cystic changes present. All 38 patients were treated non-operatively with a short arm thumb spica cast. The overall union rate of proximal pole fractures in this cohort was 84.2% (32/38, SD 0.37) and the mean overall time to union (defined as 50% union based on CT scan) was 14.1 weeks. Of the 6 fractures that were considered non-unions, one patient was non-compliant with casting and is awaiting further management, 1 patient declined further surgical intervention, and 4 patients healed successfully after ORIF with bone grafting. We were unable to identify any factors that increased the risk of non-union based on initial CT scan. However, presence of comminution and displacement were found to be significant factors contributing to delayed union.

Conclusions: In a cohort of isolated, acute proximal pole scaphoid fractures treated with casting, based on serial CT scans, we confirmed that the mean union time for casting was 14.1 weeks and the overall union rate was 84.2%. This is a much higher union rate than the previous literature reports. There were no radiographic characteristics identified that could be used to predict development of a non-union. However, comminuted and displaced fractures were found to significantly increase time to union.