Abstract
Background
Systematic reviews disagree, but some recent studies have shown better function and less pain after operation with bipolar hemiarthroplasty compared to fixation by two screws in elderly patients operated for displaced femoral neck fractures. There is still uncertainty regarding the mortality associated with both procedures.
Aim of the study
To investigate mortality and the risk factors for death among patients with displaced femoral neck fractures within the first three years after surgery, comparing operation with bipolar hemiarthroplasty (HA) and internal fixation (IF) by two screws.
Methods
12,313 patients (65 years or older) with displaced femoral neck fractures (Garden 3 and 4) operated either with IF by two screws (n = 3,436) or a bipolar HA (n = 8,877) were selected from the files of The Norwegian Hip Fracture Register 2005–2009. Mortality was assessed using Kaplan-Meier survival analysis and risk factors for death were studied using Cox-regression analysis. A power analysis showed the study sample to be sufficient to detect a difference in mortality of 3% at one and three years postoperatively.
Results
The overall one-year mortality was 28% and three-year mortality was 50%. High age, male gender, cognitive impairment, increasing ASA score and delay in surgery >48 hours after injury were all associated with increased risk of death. For patients operated with a bipolar HA there was a statistically significant increased risk of death at day 30 (RR = 1.30, p<0.01), day 120 (RR = 1.15, p = 0.01), and day 240 (RR = 1.10, p<0.01) postoperatively compared to IF by two screws. At one year (RR = 1.04, p = 0.28) and three years (RR = 1.05, p = 0.16) postoperatively, we found no difference in the risk of death when comparing operation with bipolar HA to IF by two screws.
Interpretation
The mortality for patients with displaced femoral neck fractures was considerable and dependent on age, gender, ASA score, cognitive status, and timing of surgery. Given the small magnitude of the difference in short-term mortality and no difference in one- and three-year mortality (or at least less than 3%) between the two treatment groups, we do not believe our findings should influence the choice of treatment method.
