Postoperative periprosthetic fractures of the humerus can be challenging for both the patient and the surgeon given that these fractures are difficult to treat and have high complication rates. The incidence of periprosthetic fractures after anatomic shoulder replacements has been reported to be 0.6%1. A recent series included four patients with a periprosthetic fracture after a reverse-geometry shoulder arthroplasty2, but we do not know of any other report or data on periprosthetic fractures after reverse shoulder arthroplasty. As the numbers of anatomic and reverse total shoulder replacements increase, orthopaedic surgeons will more commonly encounter these difficult fractures. Wright and Cofield proposed a classification system for periprosthetic humeral fractures3, although this system has not been validated and its ability to guide treatment has not been evaluated.
In their study, Andersen and colleagues described the operative treatment and surgical outcomes of thirty-six periprosthetic fractures after anatomic (nineteen) or reverse (seventeen) shoulder arthroplasty. Seventeen patients had a well-fixed prosthesis and were treated with open reduction and internal fixation (ORIF), whereas nineteen patients underwent revision arthroplasty because of a loose prosthesis and/or bone loss. In the ORIF group, all fractures healed, with a mean time to union of 6.8 months. In the group treated with revision arthroplasty, the original arthroplasty was reverse-geometry in sixteen cases and anatomic in three. Eighteen of the nineteen fractures in this group healed, with a mean time to union of 7.7 months. In both groups, shoulder motion at the time of the latest follow-up was similar to prefracture motion. The authors identified complications in fourteen (39%) of the thirty-six patients, which is consistent with rates reported in the literature. Campbell et al. reported a 43% complication rate after operative management of periprosthetic humeral fractures4.
Andersen and colleagues evaluated the validity of the classification system proposed by Wright and Cofield3. Three surgeons classified thirty of the thirty-six fractures according to the system of Wright and Cofield and, although the intraobserver reliability was good (mean kappa coefficient of 0.69), the interobserver reliability was poor (mean kappa coefficient of 0.37). Therefore, this classification system may not adequately direct periprosthetic fracture treatment.
The strengths of this study include its number of patients, particularly the number with periprosthetic fracture after reverse shoulder arthroplasty, and the evaluation of a previously unvalidated classification system. Although the authors found that the location of the fracture (described with use of the classification system of Wright and Cofield) was not helpful in guiding treatment of these fractures, they paid particular attention to other factors that may assist treatment decision-making. Andersen and colleagues are to be commended for their discussion of preoperative evaluation and planning, particularly with regard to stem loosening. They utilized the radiographic signs of loosening reported by Sperling et al.5 and Sanchez-Sotelo et al.6 and found only one false-positive and one false-negative result. This is particularly important for preoperative planning with regard to the surgical approach and implants. In particular, the authors noted the utility of the posterior approach for ORIF in the setting of a well-fixed stem as this approach allows the surgeon to achieve greater distal fixation. However, the deltopectoral approach is essential for revision arthroplasty. Thus, it is necessary to understand the implant stability preoperatively.
The limitations of this report include those weaknesses inherent in retrospective studies. The surgical techniques were not standardized, and some aspects of the variable technique, such as the decision to include a strut graft, have the potential to affect outcome. Additionally, the study did not include a control arm with nonoperative management.
In summary, periprosthetic humeral fractures after shoulder arthroplasty are a difficult problem to treat and are associated with a high complication rate. As shoulder arthroplasty numbers increase, so too may these troubling fractures. Although the classification proposed by Wright and Cofield may not reliably guide treatment, other measures such as stem loosening can be assessed and be used preoperatively in surgical decision-making. This study demonstrates that, with well-planned operative treatment of these fractures, a high union rate and a return to prefracture function are possible.