Does the Hertel criteria correlate with postoperative avascular necrosis and failure in proximal humerus fractures? A retrospective cohort study.

PURPOSE: Avascular necrosis and failure are significant complications following surgical fixation of proximal humerus fractures. The Hertel criteria are a widely used morphological classification system intended to predict humeral head ischaemia. This study aimed to evaluate the correlation of Hertel criteria with the development of postoperative avascular necrosis and failure and to assess the prognostic value of other radiological parameters. MATERIALS AND METHODS: In this retrospective cohort study, 94 patients with proximal humerus fractures treated with locking plate fixation and a minimum 2-year follow-up were included. Preoperative radiographs and computed tomography scans were used to assess Hertel criteria, Neer classification, and proximal humeral cortical thickness. Postoperative radiographs were evaluated for avascular necrosis and failure, defined as a composite of nonunion, severe avascular necrosis, screw penetration, severe deformity, or significant arthrosis. Clinical outcomes were assessed using the Visual Analog Scale and Disability of Arm, Shoulder and Hand scores. RESULTS: The presence of an anatomical neck fracture and tubercle displacement greater than 10 mm were significantly associated with both avascular necrosis and failure (p < 0.05). In contrast, other factors in the classical Hertel criteria, such as metaphyseal head extension less than 8 mm, medial hinge displacement greater than 2 mm, and humeral head angle greater than 45°, did not show a significant correlation with avascular necrosis and failure. More severe Neer fracture types were also significantly associated with worse outcomes. Cortical thickness was not a significant predictor. CONCLUSION: Anatomical neck fracture and tubercle displacement according to Hertel criteria are critical radiological risk factors for postoperative avascular necrosis and failure. However, other parameters do not show the same correlation. LEVEL OF EVIDENCE: Level III, Retrospective cohort study.