LAUSR

Abstract Background Three-part and four-part fracture-dislocations of the proximal humerus are characterized by severe soft tissue disruptions that can compromise the viability of the humeral head. As a result, nonunion and avascular necrosis are more common in these injuries. In such injuries, surgical treatment (internal fixation or arthroplasty) is performed in most patients who are determined to be fit for surgery to potentially restore shoulder function. Although the decision to preserve or replace the humeral head is simple in young patients or those > 65 years, in most other patients, the decision can be complicated, and little is known about which patient-related and injury-related factors may be independently associated with poor shoulder function or complications like avascular necrosis. Questions/purposes (1) What proportion of fractures united after internal fixation of a three-part or four-part fracture-dislocation of the proximal humerus, what is the mean Constant score at a minimum of 2 years after this injury, and what proportion had serious complications (such as loss of fixation, nonunion, reoperation, or avascular necrosis)? (2) After controlling for potential confounding variables, what factors are independently associated with poor shoulder function (defined as a Constant score < 55 out of 100) and occurrence of serious complications such as loss of fixation or reduction resulting in revision surgery, nonunion, or radiographic evidence of avascular necrosis of the humeral head? Methods Between 2011 and 2017, the senior author of this study (ASG) treated 69 patients with three-part or four-part proximal humerus fracture dislocations. During this time, indications for internal fixation in these patients were adequate humeral bone quality as determined by the surgeon on radiographs, adequate bone stock and volume available for fixation in the humeral head as determined on CT images, and the absence of a head split component as assessed on preoperative radiographs and CT images. On this basis, 87% (60 patients) underwent internal fixation with a locked plate and suture fixation of the tuberosities through a deltopectoral approach. Thirteen percent (nine patients) underwent either a hemiarthroplasty or a reverse total shoulder arthroplasty. Of the 60 patients who underwent internal fixation, four declined to participate in the study and two with brachial plexus palsy were not considered for inclusion. This study focused on the remaining 54 patients who were considered potentially eligible. To be included, a minimum follow-up of 2 years was required; 11% (6 of 54) were lost before that time, and the remaining 48 patients were analyzed at a mean of 48 months ± 17 months in this retrospective study, which drew data from longitudinally maintained institutional databases. Fracture union was assessed by obliteration of fracture lines and the presence of bridging trabecular bone on plain radiographs. Shoulder function was assessed using the Constant score, which is scored from 0 to 100 points, with 0 indicating the most disability and 100 the least disability. The anchor-based minimal clinically important difference for the Constant score is 9.8 points. Twelve patient-related and injury-related factors were analyzed using a multivariate regression model to identify factors that are independently associated with poor results after internal fixation as measured by shoulder function and the occurrence of serious complications. We categorized results as poor if patients had one or more of the following: Constant score < 55 out of 100 at the last follow-up examination (for patients who underwent revision surgery, the Constant score immediately before revision was considered) and loss of fixation or reduction resulting in revision surgery, nonunion, or avascular necrosis of the humeral head. Patients were screened for avascular necrosis at 6 and 12 months after surgery, then annually for another 2 years. Further assessments were made only based on symptoms. Results Seventy-nine percent of the fractures united within 18 weeks of surgery (38 of 48), and an additional 13% united by 24 weeks (6 of 48), while 8% did not unite (4 of 48). The mean Constant score at the last follow-up was 68 ± 12. Twenty-one percent (10 of 48) had a Constant score < 55, indicating poor shoulder function. Twenty-one percent (10 of 48) experienced avascular necrosis, and 15% (7 of 48) with either nonunion or avascular necrosis underwent revision shoulder arthroplasty. Two patients who underwent arthroplasty had both nonunion and avascular necrosis. After controlling for potentially confounding variables, we found that being a woman (odds ratio 1.7 [95% confidence interval 1.4 to 2.1]; p = 0.01), four-part fracture dislocations (OR 2.1 [95% CI 1.5 to 2.7]; p < 0.001), absence of a metaphyseal head extension (OR 2.4 [95% CI 1.8 to 3.3]; p < 0.001), absence of active back-bleeding from the humeral head (OR 3.4 [95% CI 2.3 to 5.1]; p < 0.001), height of the head segment < 2 cm (OR 2.3 [95% CI 1.8 to 2.8]; p < 0.001), and absence of capsular attachments to the head fragment (OR 2.2 [95% CI 1.6 to 2.9]; p < 0.001) were independently associated with poor shoulder function and the occurrence of complications such as nonunion and avascular necrosis. Conclusion Internal fixation of three-part and four-part proximal humerus fracture dislocations resulted in poor shoulder function and complications in a high number of patients, although fracture union was achieved in most patients. A nonunion proportion of 8%, 21% proportion of avascular necrosis, and 15% proportion of patients who underwent revision surgery suggests this is a fairly terrible injury. Being a woman and injury factors such as four-part fracture dislocation, absent metaphyseal head extension and back-bleeding from the head, height of the fractured head segment < 2 cm, and absence of capsular attachments to the head were independently associated with poor function and complications. Our findings can help surgeons decide between internal fixation and arthroplasty for the surgical treatment of these injuries in patients across different age groups and functional demands. Level of Evidence Level III, therapeutic study