Radiographs and Computed Tomography in the Diagnosis of Open Fractures: Is the Presence of Air in the Soft Tissues an Accurate Diagnostic Sign?

Open fractures are catastrophic injuries that are associated with high rates of infection and the need for additional interventions. The presence of air on radiograph and computed tomography (CT) is taught to be an indicator of open fracture. The purpose of our study was to retrospectively review preoperative radiographs and CT in patients with closed and open fractures to determine the sensitive and specificity of soft tissue free air in diagnosing an open injury.

All confirmed open fractures treated at our institution from 2009 to 2017 were identified using irrigation and debridement CPT codes 11010, 11011, 11012. A separate cohort of surgically treated and confirmed closed fractures was also evaluated. The closed fracture cohort was matched by fracture location to the open group and had similarly complete radiographic and CT imaging. Radiographs were evaluated for presence of air within the soft tissues surrounding the fracture. The CT was evaluated for the presence of free air utilizing three separate positive definitions: 1) presence of any air in the soft tissues; 2) air located within the subfascial soft tissues; and 3) air located only in the juxtacortical area. The sensitivity and specificity of radiographs and the three different CT definitions were calculated and compared.

The open and closed fracture cohorts each contained 110 patients. The former was composed of 18.2% (20/110) Type I, 31.8% (35/110) Type II and 50% (55/110) Type III open fractures.

Radiographs were 94.5% sensitive and 100% specific in diagnosing open fractures. CT was found to be more sensitive for all three positive definitions but less specific. CT sensitivity was 97-100% and the specificity was found to be 77.5-80%. Overall, the positive predictive value (PPV) for radiographs was 100%. CT had a PPV range from 92.4-93.1%. No significant difference in the PPV for radiographs or CT was found when subclassifying the fractures based on their Gustilo-Anderson Type (p>0.05). The negative predictive value (NPV) for radiographs was 86.9%. The NPV for CT ranged from 91.4-100% with the presence of any air having an NPV of 100%; this was not significantly different from radiographs (p>0.05).

We conclude that orthogonal radiographs are a sufficient initial imaging modality to confirm the clinical suspicion of open fracture based on physical exam findings. If clinical exam and radiographs are inconclusive, the absence of any air on CT scan is highly reliable for ruling out the presence of an open fracture.