To determine whether deficient early callus formation can be defined objectively based on association with an eventual nonunion and specific patient, injury, and treatment factors. Final healing outcomes were documented for 160 distal femur fractures treated with locked bridge plate fixation. Radiographic callus was measured on postoperative radiographs until union or nonunion had been declared by the treating surgeon. Deficient callus was defined at 6 and 12 weeks based on associations with eventual nonunion via receiver operator characteristic analysis (ROC). A previously described computational model estimated fracture site motion based on the construct employed. Univariable and multivariable analyses then examined the association of patient, injury, and treatment factors with deficient callus formation. There were 26 nonunions. Medial callus area at 6 weeks < 24.8 mm2 was associated with nonunion (12 of 39, 30.8%) vs. (12 of 109, 11.0%), p = 0.010. This association strengthened at 12 weeks with medial callus area < 44.2 mm2 more closely associated with nonunion (13 of 28, 46.4%) vs (11 of 120, 9.2%), p < 0.001. Multivariable logistic regression analysis found limited initial longitudinal motion (OR 2.713 (1.12 to 6.60), p = 0.028)) and Charlson Comorbidity Index (1.362 (1.11 to 1.67), p = 0.003) were independently associated with deficient callus at 12 weeks. Open fracture, mechanism of injury, smoking, diabetes, plate material, bridge span, and shear were not significantly associated with deficient callus. Deficient callus at 6 and 12 weeks is associated with eventual nonunion and such assessments may aid future research into distal femur fracture healing. Deficient callus formation was independently associated with limited initial longitudinal fracture site motion derived through computational modeling of the surgical construct, but not more routinely discussed parameters such as plate material and bridge span. Given this, improved methods of in vivo assessment of fracture site motion are necessary to further our ability to optimize the mechanical environment for healing.
Read full abstract