Locked plating of 3- and 4-part proximal humerus fractures in older patients: the effect of initial fracture pattern on outcome.

OBJECTIVES The use of locked plates in repairing osteopenic 3- and 4-part proximal humerus fractures remains controversial. The purpose of this article was to report the outcomes of open reduction and internal fixation in low-energy proximal humerus fractures treated with locked plating in patients older than 55 years and stratify risk of failure or complication based on initial radiographic features. DESIGN Retrospective. SETTING Level I Trauma Center. METHODS Seventy patients older than 55 years undergoing locked plate fixation for Neer 3- or 4-part proximal humerus fractures were studied retrospectively. All patients had standardized, true size digital radiographs of the injured and normal shoulder in the axillary, scapular Y, and 20-degree external rotation views with a minimum of 18 months' clinical follow-up. Two groups were identified based on the initial direction of the humeral head deformity: varus or valgus impaction. There were no statistical differences between treatment groups with regard to age, sex, Neer classification, follow-up, or dislocation. Radiographic measurements included humeral head angulation, tuberosity displacement, and length of the intact metaphyseal segment. Clinical outcomes measured Constant scores (CS) using active range of motion at latest follow-up. RESULTS Twenty-four patients with initial varus fracture patterns healed with an average of 16-degree varus head angulation and an overall CS of 63 at an average of 34 months' follow-up. Forty-six patients with initial valgus fracture patterns healed with an average of 6 degrees of varus angulation and an overall CS of 71 at an average of 37 months' follow-up (P < 0.01). Complications of avascular necrosis, humeral head perforation, loss of fixation, tuberosity displacement >5 mm, and varus subsidence >5 degrees were encountered in 19 of 24 (79%) in the varus group compared with 9 of 46 (19%) in the valgus group (P < 0.01). Final CSs for 3-part fractures were 65 versus 72 (P < 0.01) for varus and valgus groups, respectively, and 61 versus 69 (P = 0.19) for 4-part fractures. CONCLUSIONS Neer 3- and 4-part proximal humeral fractures in older patients with initial varus angulation of the humeral head had a significantly worse clinical outcome and higher complication rate than similar fracture patterns with initial valgus angulation. Two factors had significant influence on final outcome in these fracture patterns: initial direction of the humeral head angulation and length of the intact metaphyseal segment attached to the articular fragment. The best clinical outcomes were obtained in valgus impacted fractures with a metaphyseal segment length of greater than 2 mm, and this was independent of Neer fracture type. Humeral head angulation had the greatest effect on final outcomes (P < 0.001), whereas metaphyseal segment length of less than 2 mm was predictive of developing avascular necrosis (P < 0.001).