Multifragmentary proximal humeral fractures-precise fracture anatomy of the tuberosities identified with three-dimensional computed tomography.

Background: Although the Neer classification of proximal humeral fractures typifies fracture anatomy, clinical practice often presents cases in which distinguishing between fracture types can be challenging. Moreover, many atypical fractures, including shield fractures, have been documented. These indicate that the fracture anatomy of the tuberosities is not as simple as Neer's description, particularly in multifragmentary fractures. We hypothesized that three-dimensional computed tomography (3DCT) could provide a comprehensive view of the fractured tuberosities in multifragmentary proximal humeral fractures. Methods: A retrospective study was conducted on 80 patients, aged 34 to 94 years, who sustained multifragmentary proximal humeral fractures. The initial diagnosis of the surgeons identified 37 three-part fractures, 16 four-part fractures, 19 four-part valgus-impacted fractures, and 8 fracture-dislocations, according to the Neer classification. Pretreatment volume rendering 3DCT scans, including a superior view, were reviewed by 2 orthopedic surgeons. Tuberosity fractures were categorized as single tuberosity pattern (displacement of either the greater or the lesser tuberosity), dual tuberosity pattern (separate displacement of both tuberosities), or shield pattern (displacement of both tuberosities as well as the bicipital groove, encircling the humeral head). Results: The interobserver reliability for categorizing tuberosity fracture patterns was 0.725 in terms of the Cohen's kappa. Tuberosity fractures were identified as a single tuberosity pattern in 41% of cases, a dual tuberosity pattern in 13%, and a shield pattern in 36%. Most of the three-part fractures had a single tuberosity pattern, while all four-part valgus impacted fractures had a shield pattern. Detailed observation of the single tuberosity patterns revealed that, in 68% of cases, the fracture line was located 5 to 10 mm posterior to the anterior margin of the greater tuberosity, leaving the anteriormost portion of the greater tuberosity unfractured. Conclusion: 3DCT clearly demonstrated 3 patterns of tuberosity fracture: single, dual, and shield patterns. A shield pattern, an anatomical neck fracture laterally, was common among multifragmentary proximal humeral fractures. In greater tuberosity fractures, the fracture line was located posterior to the supraspinatus tendon insertion in many instances, while a fracture of the entire greater tuberosity was uncommon.