Periprosthetic fractures around hip hemiarthroplasty performed for hip fracture.

UNLABELLED: Hip fracture is associated with considerable morbidity and mortality and occurs in an elderly and infirm group of patients. Periprosthetic fracture after hip hemiarthroplasty is a serious complication. In this study, we have reviewed our experience of this injury. The outcome measures used were fracture union, mortality, infection and requirement for further surgery.
METHOD: We identified a cohort of 79 patients who sustained periprosthetic fractures after hip hemiarthroplasty from a prospective hip fracture database of 8354 patients (3611 were treated with hemiarthroplasty). Seventy-two percent were female and the mean age was 86 years at time of periprosthetic fracture.
RESULTS: Sixty-two fractures occurred around uncemented prostheses (Austin Moore n=61); the remainder occurred around cemented prostheses. The mean time from hip fracture surgery to periprosthetic fracture was 35 months (median time 5 months). Fractures were classified according to the Vancouver system. Fifteen percent (n=12) were type A fractures, 26% (n=21) were type B1 fractures, 41% (n=32) were type B2 fractures, 9% (n=7) were type B3 fractures and 9% (n=7) were type C fractures. Twenty-eight patients underwent open reduction internal fixation (ORIF), 36 required revision surgery, one required fixation and simultaneous revision and 14 were treated non-operatively. Eleven percent (n=9) died within 1 month of periprosthetic fracture, 23% had died by 3 months, 34% by 1 year and 49% by 2 years. Nineteen patients (24%) died before fracture union had occurred. Fracture union occurred in 97% of the remaining cases (58/60). Two patients developed nonunion requiring revision surgery (3%), and three patients developed deep infection requiring debridement (4%), one patient had an infection at the time of the periprosthetic fracture requiring a planned two-stage revision, one patient sustained a second periprosthetic fracture and two patients underwent superficial wound debridement (3%). The incidence of periprosthetic fracture at our institution since 1999 is 1.7% (62 of 3611 patients). The incidence rate after uncemented Austin Moore stem was 2.3% (54/2378) and cemented Exeter stem was 0.5% (4/812); Fisher's exact test p=0.004.
CONCLUSIONS: This article reports satisfactory outcomes in this complex group of patients. We have established the incidence of 1.7%, with relatively low rates of nonunion, infection and other complications. The mortality rate has been established, and survivorship analysis has identified an increased rate of fracture around the Austin Moore prosthesis.