Fibula allograft in complex three-part and four-part proximal humeral fractures in active patients, a matched case-control study.

Background: About 20% of proximal humerus fractures (PHFs) are unstable and/or markedly displaced and therefore require surgery. Locking plate fixation after anatomical reduction has become the current treatment of choice for these fractures in the active population. However, studies have shown complication rates up to 36%, such as loss of reduction and avascular necrosis. To date, data from literature are inconclusive on outcomes following the use of an intramedullary fibula allograft in PHFs, possibly due to the case mix. It is hypothesized that the use of a fibula allograft is beneficial to prevent secondary displacement of the fracture in cases where the medial hinge is markedly displaced and unstable, resulting in better clinical and patient reported outcomes.

Methods: In this multicenter matched cohort study, patients with an unstable, displaced PHF, including anatomic neck fractures and significantly displaced surgical neck fractures, were included. Patients that were treated with a locking plate augmented with a fibula allograft were matched to patients who had undergone locking plate reconstruction without the allograft. The matches were made based on fracture characteristics, age, and performance status. Functional outcomes, Patient Reported Outcome Measures, complications, and radiographic results were compared.

Results: Twelve patients with fibula allograft augmented osteosyntheses were included and matched to 12 control patients. The mean age was 58 years in the fibula allograft group compared to 62 years in the control group. Minimum follow-up was 12 months. Disability of the Arm Shoulder and Hand score, Constant Shoulder score, abduction, and external rotation were significantly better in the fibula allograft group (17.4 ± 8.6 vs. 26.1 ± 19.2,  = .048; 16.5 ± 11.5 vs. 19.8 ± 16.5  = .040; mean 127° ± 38° vs. mean 92° ± 49°  = -.045; 50° ± 21° vs. mean 26° ± 23°,  = .004). There was no statistically significant difference in the Oxford Shoulder score between groups ( = .105). The Visual Analog Scale was not significantly different between groups (3.1 ± 1.8 vs. 1.6 ± 1.9,  = .439). Radiographic union was reached in 11 patients of the fibula allograft group compared to 8 in the control group ( = .317). The complication rate was twice as high in the control group (3 vs. 7).

Conclusion: Additional support of the medial hinge in unstable PHFs with a locking plate in combination with a fibula allograft appears to create a more stable construct without compromising the viability of the articular surface of the head. The use of a fibula allograft in selected complex cases could therefore result in better clinical outcomes with lower complication rates.