Defining the Lower Limit of a “Critical Bone Defect” in Open Diaphyseal Tibial Fractures

Authors:
  • WLWilliam D. Lack
    William D. Lack
  • RSRachel B. Seymour
    Rachel B. Seymour
  • MBMichael J. Bosse
    Michael J. Bosse

Abstract
To determine healing outcomes of open diaphyseal tibial shaft fractures treated with reamed intramedullary nailing (IMN) with a bone gap of 10-50 mm on ≥50% of the cortical circumference and to better define a "critical bone defect" based on healing outcome. Retrospective cohort study. Forty patients, age 18-65, with open diaphyseal tibial fractures with a bone gap of 10-50 mm on ≥50% of the circumference as measured on standard anteroposterior and lateral postoperative radiographs treated with IMN. IMN of an open diaphyseal tibial fracture with a bone gap. Level-1 trauma center. Healing outcomes, union or nonunion. Forty patients were analyzed. Twenty-one (52.5%) went on to nonunion and nineteen (47.5%) achieved union. Radiographic apparent bone gap (RABG) and infection were the only 2 covariates predicting nonunion outcome (P = 0.046 for infection). The RABG was determined by measuring the bone gap on each cortex and averaging over 4 cortices. Fractures achieving union had a RABG of 12 ± 1 mm versus 20 ± 2 mm in those going on to nonunion (P < 0.01). This remained significant when patients with infection were removed. Receiver operator characteristic analysis demonstrated that RABG was predictive of outcome (area under the curve of 0.79). A RABG of 25 mm was the statistically optimal threshold for prediction of healing outcome. Patients with open diaphyseal tibial fractures treated with IMN and a <25 mm RABG have a reasonable probability of achieving union without additional intervention, whereas those with larger gaps have a higher probability of nonunion. Research investigating interventions for RABGs should use a predictive threshold for defining a critical bone defect that is associated with greater than 50% risk of nonunion without supplementary treatment. Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.

This publication has 21 references indexed in Scilit: