Objectives: To identify discrete construct characteristics related to overall construct rigidity that may be independent predictors of nonunion after lateral locked plate (LLP) fixation of distal femur fractures.
Design: Retrospective case-control study.
Setting: Three level-1 urban trauma centers.
Patients/participants: Two hundred and seventy-one supracondylar femoral fractures treated with LLP at 3 affiliated level 1 urban trauma centers between August 2004 and December 2010.
Methods: Nonunion was defined as a secondary procedure for poor healing. Construct variables included: (1) combined plate design and material variable, (2) Plate length, (3) # screws proximal to fracture, (4) total screw density (TSD), (5) proximal screw density (PSD), (6) presence of a screw crossing the main fracture, and (7) rigidity score multivariable analysis was performed using logistic regression to identify independent risk factors for nonunion.
Intervention: LLP fixation.
Main outcome measure: Nonunion.
Results: Nonunion rate was 13.3% (n = 36). There was a significant association between plate design/material and nonunion with 41% of stainless constructs and 10% of titanium constructs resulting in a nonunion (P < 0.001). Rigidity scores reached significance (P = 0.001) with constructs resulting in a nonunion having higher scores. No significant univariate differences with respect to number of proximal screws, plate length, total screw density, or proximal screw density were observed between healed fractures and those with nonunion. Results of the multivariate analysis confirmed that the primary significant independent predictor of nonunion was plate design/material (odds ratio, 6.8; 95% CI, 2.9-16.1; P < 0.001).
Conclusions: When treating distal femur fractures with LLP, combined plate design and material variable has a highly significant influence on the risk of nonunion independent of any other construct variable.
Level of evidence: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.