Session VI - Geriatrics

Fractures of the Femur Following Hip Fracture Surgery

C.M. Robinson, FRCS; C.I. Adams; M. Craig; W. Doward; M.C.C. Clarke; J. Auld, Royal Infirmary of Edinburgh, Edinburgh, Scotland

Purpose: To examine the epidemiology of ipsilateral fractures of the femur following hip fracture surgery, with regard to the incidence, presentation, risk factors, treatment and prognosis.

Methods: Over the 10-year period from January 1988 to December 1997, 6,230 patients (median age 82 years, male:female ratio 1,247:4,983) who sustained 6,696 hip fractures were admitted to the Orthopaedic Trauma Unit of the Royal Infirmary of Edinburgh and prospectively coded onto the Edinburgh Orthopaedic Trauma database. Only permanent Edinburgh residents undergoing surgical treatment for an intracapsular neck, basi-cervical or intertrochanteric femoral neck fracture were included in this study. Patients with subtrochanteric fractures, pathological fractures and fractures of the greater or lesser trochanter were excluded, as were the 132 patients who died pre-operatively and the 35 medically unfit patients who were treated non-operatively.

The Unit is the only hospital treating fractures for a well-defined estimated catchment population of 595,591 people aged over 13 years during the period of the study. All patient injury events were checked using other prospectively collected data sources, including the Scottish Trauma Audit Group and Scottish Hip Fracture Audit to provide a 98% cross-validation of individual cases. Type of fracture was verified by re-classification of initial radiographs with blinding to outcome. Mortality data were collected on all patients included in the study from the General Register Office for Scotland. All femoral fractures were coded on the same database and re-admissions with ipsilateral or contralateral fracture of the femur following previous hip surgery were re-coded, linked to the previous fracture event. The treatment used to stabilise the femoral fractures was recorded together with any subsequent complications.

Results: One hundred forty-one patients (2.1%) sustained a subsequent ipsilateral re-fracture of the femur by the end of 1998. The median age of these patients was identical to that of the hip fracture population at 82 years (interquartile range 74.75 to 86 years) with a male to female ratio of 17:124. Refractures presented either after a further simple fall (in 124 patients) or after sudden onset of thigh pain while walking (in 17 patients).

The incidence of ipsilateral fracture of the femur increased with age and mirrored the incidence of hip fractures in the general population in that the incidence in ipsilateral re-fracture was consistently much higher in post-menopausal females than males. Analysis of prosthetic fractures by type of implant revealed considerable differences in both incidence and configuration of fracture. The highest incidence of re-fracture was seen in the 210 patients who required revision of their primary implant to a cemented arthroplasty. The incidence of prosthetic fracture was also relatively high in an intermediate group consisting of patients treated by Gamma nail and uncemented hemiarthroplasty and was low in patients treated using either a compression hip screw, cannulated screws or primary cemented hemiarthroplasty.

Two-thirds of the fractures in the series propagated from the tip of the prosthesis. The risk of ipsilateral versus contralateral fracture of the femur was independently greater both for fractures occurring in the diaphysis (relative risk [RR] = 5.20, 95% CI = 3.10-8.72, P = 0.001) and in the distal metaphysis (RR = 2.33, 95% CI = 1.25-4.33, P = 0.005). Multiple logistic regression analysis of patients with extracapsular and intracapsular fractures revealed that in both sub-groups the presence of severe osteoporosis was independently predictive of later femoral fracture. In addition, among extracapsular fractures, the use of the Gamma nail was independently predictive of femoral fracture, increasing the risk by almost threefold compared with the DHS (OR = 2.98, 95% CI = 1.45-6.12, P = 0.003). Following an intracapsular fracture, revision to cemented arthroplasty following failed primary treatment increased the risk of femoral fracture by fourfold (OR = 4.00, 95% CI =2 .24-7.15, P < 0.001). The combination of factors (high-risk implant and severe osteoporosis) was associated with subsequent femoral fracture in 4.7% of patients with extracapsular fracture and 16.2% of patients with intracapsular fracture.

A policy of operative stabilization of prosthetic fractures was adopted for all patients who were medically fit for a general anaesthetic. Of the 135 patients treated operatively, postoperative complications occurred in 12 (8.8%), and most were managed with further operative intervention. The outcome in terms of level of mobility and requirement for institutional care was significantly worse in patients with ipsilateral femoral fracture when compared with the hip fracture population as a whole (P = 0.001 and P = 0.004 respectively). Survival analysis to death demonstrated that the median survival following ipsilateral re-fracture of the femur was worse than the hip fracture population during the first 3 years after fracture (P = 0.005).

Discussion: Prosthetic fractures of the femur following hip fracture represent a different spectrum of injuries from prosthetic fractures associated with hip or knee replacement for arthritis. They occur earlier, in a more elderly population, with clearly recognizable risk factors for fracture. Furthermore these patients are often more medically frail and osteoporotic resulting in difficult revision surgery to stabilize the fracture. The risk of re-fracture was much greater in the presence of severe osteoporosis and also showed considerable variation depending on the type of primary implant used to stabilise the hip fracture. The difference in incidence of fracture can, therefore, largely be explained by the patient selection process that determined the primary treatment and the configuration of implant used. In all multivariate environments the severity of the osteoporosis was the greatest predictor of subsequent risk of re-fracture. In general, the use of implants with an intramedullary stress riser increased the risk of later femoral fracture, whereas implants without an intramedullary stress riser had a lower incidence of later fracture.

Conclusion: The Gamma nail should not be used in the treatment of extracapsular hip fractures. In patients with severe osteoporosis and a displaced intracapsular fracture who are treated by hemiarthroplasty, attempts should be made to exclude a preventable cause for osteoporosis such as osteomalacia, as well as instituting secondary preventive treatment for osteoporosis. The highest incidence of prosthetic fractures followed revision to cemented hip arthroplasty following failed primary treatment. Theoretically, the use of a long-stem arthroplasty may reduce the risk of fracture at the tip of the prosthesis in these high risk patients. However, the risk of dislocation and infection is increased in revision procedures following failed primary treatment; the use of a long-stem implant in these circumstances may significantly hamper any attempts at further revision surgery. We advise that in the presence of multiple screw holes at the lower end of the implant, a long-stem prosthesis be used, based on our findings. Patients undergoing revision following previous arthroplasty surgery may be adequately treated with a short-stem implant.