Prevalence of non-union and delayed union in proximal humeral fractures.

BACKGROUND Little is known about the prevalence of proximal humeral non-union. There is disagreement on what constitutes union, delayed union and non-union. Our aim was to determine the prevalence of these complications in proximal humeral fractures (PHFs) admitted to trauma hospitals. METHODS The Victorian Orthopaedic Trauma Outcomes Registry identified 419 cases of PHFs, of which 306 were analysed. Three upper limb orthopaedic surgeons used X-rays to classify fractures according to the Neer classification and determine union. Twelve-item Short Form Health Survey scores were used to assess patient health and wellbeing. RESULTS Of 306 cases, 49.4% reached union. Median time to union was 100 days (confidence interval 90-121). Of these, 17.0% united by 60 days, 8.5% united by 89 days and 23.9% united after 90 days, demonstrating 'prolonged delayed union'. There were 25 non-unions with a prevalence of 8.2%, most occurring in two-part surgical neck fractures. CONCLUSION Our cohort of largely displaced PHFs admitted to trauma hospitals had a non-union prevalence of 8.2% and an overall delayed union prevalence of 32.4%. Consensus is required on definitions of non-union and delayed union timeframes.