Open reduction and internal fixation of humeral nonunions - A biomechanical and clinical study

Background: Several studies have compared different methods for fixation of the midpart of the humeral shaft, but there are only scattered data regarding which type of plate construct provides the best fixation for humeral nonunion. The objectives of this study were (1) to obtain objective data on the performance of four different plate constructs used for fixation of humeral nonunion, and (2) to report our clinical experience with plate fixation of thirty-seven nonunions of the midpart of the humeral shaft. Methods: In the first part of the study, four plate constructs were compared in a Sawbones model. The groups consisted of (1) a posterior limited-contact dynamic compression plate alone; (2) a posterior limited-contact dynamic compression plate and an interfragmentary screw; (3) a posterior limited-contact dynamic compression plate, a lateral 3.5-mm reconstruction plate, and an interfragmentary screw; and (4) a posterior limited-contact dynamic compression plate and a lateral 3.5-mm reconstruction plate. Tests were performed with use of an MTS Bionix machine in anterior-posterior four-point bending, medial-lateral four-point bending, and external rotation torque. In the second part of the study, the charts of thirty-seven consecutive patients in whom a nonunion of the midpart of the humeral shaft had been treated with plate fixation were reviewed retrospectively. The average age of the patients was forty-eight years (range, thirteen to seventy-eight years). Nineteen patients were treated with a single posterior plate, and eighteen were treated with a two-plate construct with the plates parallel and lying at 900 to each other. All of the nonunions were treated with bone-grafting, and an interfragmentary screw was used in thirty-six of the thirty-seven patients. Radiographs and the clinical status were evaluated at an average of thirteen months postoperatively. Results: The biomechanical testing showed that the two-plate constructs were significantly stiffer than the single-plate constructs in all test modes (p < 0.05). In the clinical part of the study, thirty-four (92%) of the nonunions healed without complications at an average of 4.8 months. Two nonunions treated with the two-plate construct and one treated with one plate failed to heal. Conclusions: No significant difference in the healing rate was found between the two clinical groups (p = 0.4, 0 0.9), and the overall healing rate was 92%. However, a two-plate construct with the plates at right angles is mechanically stiffer than a single-plate construct, which might be helpful if rigid stabilization of the humerus at the midshaft level is needed.