This study reviews eleven patients treated for a humeral periprosthetic fracture after a total elbow arthroplasty by the senior author between 1991 and 1999. Treatment included cementation of a Coonrad-Moorey revision stem past the fracture site with one to three strut allografts wired around the humeral shaft for reinforcement. Six fractures occurred after a primary arthroplasty and five occurred after a revision arthroplasty. Six elbows required revision of the ulnar component. Seven elbows had complete distal humeral bone loss at least to the level of the olecranon fossa. Impaction grafting was used in three cases. Final follow-up ranged between 9 months to 7.8 years (average 3 years) with age ranges from 42 to 85 years. Overall, 10 of 11 patients achieved union. One patient developed aseptic loosening at 8 years post-operatively which required revision arthroplasty. Seven of eight patients with an intact reconstruction had a functional arc of motion ranging between 16 degrees of extension (range, 0 to 30 degrees) and 131 degrees of flexion (range, 110-140 degrees). There were four excellent results, four good and three poor as rated by the Mayo Elbow Performance Score. Seven patients were satisfied with their result. Six patients had at least one complication. Two patients had transient nerve dysfunction involving the radial and ulnar nerves that recovered completely. One patient had a permanent ulnar nerve injury. One patient who underwent a revision of the ulnar component sustained a fracture of the olecranon at 6 months after treatment. There was one patient each with a nonunion, aseptic loosening, new humeral periprosthetic fracture and triceps insufficiency.
Cemented humeral implant revision with strut allograft augmentation results in a high union rate for loose periprosthetic humeral shaft fractures. This treatment parallels recommendations for lower extremity periprosthetic fractures; however, cement was uniformly used considering that press fit stems are not available for the elbow and bone stock is often quite deficient (especially in the rheumatoid population). As expected, the complication rate was quite high occurring in over 50% of patients. This included 3 nerve injuries resulting in patient satisfaction in only 7 of 11 cases. The results of this paper may help the clinician provide realistic expectations to the patient confronted with revision arthroplasty due to loosening and fracture.