Is arthroplasty preferable to internal fixation for the treatment of extracapsular fracture of the upper femur in the elderly?

INTRODUCTION: Although internal fixation is the reference treatment for extracapsular fracture of the upper femur, indications for arthroplasty are broadening, especially in unstable comminutive fracture in fragile bone. The present study hypothesis was that arthroplasty reduces early mortality and morbidity and provides better recovery of autonomy in over-80 year-old patients than does internal fixation. MATERIAL AND METHODS: A prospective multicenter study was conducted on 8 sites. Internal fixation was systematically used in 5 centers; arthroplasty was used systematically in 1 center, and reserved for unstable fracture in 2 centers. A total of 697 patients aged over 80 years (mean age, 85±5 years), presenting with extracapsular fracture, were included; 521 were treated by internal fixation and 176 by arthroplasty. Results were studied on multivariate analysis of ASA score, blood loss, transfusion, and also of treatment modality as an independent factor for early (first 6 months) mortality and morbidity (mechanical, general and nutritional complications) and functional outcome (autonomy and dependence). RESULTS: Overall mortality was 19.2%. Autonomy deteriorated in 56% of patients alive at 6 months and dependence worsened in 44%. Two percent of those managed by internal fixation underwent revision for disassembly (n=8) or infection (n=1). Eight percent of those managed by arthroplasty underwent revision for dislocation (n=4), implant loosening (n=3) or infection (n=7). On univariate analysis, mortality was higher in the arthroplasty group (25%) than with internal fixation (17%; P=0.002), as were blood loss (425±286mL versus 333±223mL; P<0.0001), transfusion rate (61% versus 32%; P<0.0001) and infection (4% versus 0.2%; P<0.001). On multivariate analysis, however, treatment modality no longer showed impact on mortality or on morbidity and autonomy at 6 months. Nutritional status was better conserved at 6 months following arthroplasty, but dependence worsened. Poor preoperative autonomy, ASA score, and nutritional status and time to treatment were independent factors for mortality. Transfusion, associated with onset of mechanical complications, significantly increased dependence. CONCLUSION: Type of treatment had little impact on mortality, morbidity or functional outcome. Differences seemed more related to preoperative functional and nutritional status. LEVEL OF EVIDENCE: III, prospective case-control study.