Although patients undergoing HA for femoral neck fractures are older and have more medical comorbidities, we found that the rate of adverse events in the perioperative period for patients undergoing THA was higher than that in the HA group. Consistent with prior studies, patients who underwent THA had higher rates of blood transfusion, pulmonary embolism, and induced mental disorders.34 Multivariable regression analysis demonstrated that after controlling for age, medical comorbidity, and type of surgery performed, THA emerged as an independent risk factor for any adverse event in the perioperative period. Increased anesthesia time, reaming of the acetabulum, and increased complexity of surgery probably account for these changes.
Our study results are consistent with those of Jain and colleagues,30 which showed a decrease in utilization of THA for femoral neck fractures between 1990 and 2001. Since THA is generally indicated for younger, more active patients in relatively good health, this would explain why changes in baseline health in this cohort over the last 20 years would lead to fewer THA procedures being performed. Surgeons in the US may be finding there are fewer patients who are candidates for THA. Miller and colleagues5 reported conflicting results and showed an increase in THA utilization in this patient population. However, their study evaluated treatment trends based on data from the ABOS part II of recently graduated orthopedic surgeons and may not be an accurate representation of national practice trends in the US. The trend toward increased subspecialization may explain their findings. As the authors noted, although they found an increase in the use of THA for femoral neck fractures by new adult reconstruction surgeons, the percentage of new surgeons treating femoral neck fractures has declined.5
Our analysis showed very concrete trends in treatment management at the extremes of the age ranges. There were substantial increases in the use of ORIF for patients <50 years (from 63.9% in 1990–1995 to 81.4% in 2002–2007, P < .001) and in the use of HA for patients >80 years (from 74.4% in 1990–1995 to 84.6% in 2002–2007, P < .001). This trend parallels recent studies that purport better outcomes for young patients undergoing ORIF and elderly patients undergoing HA.30 Our analysis did not demonstrate a large shift in surgeon preference for treatment of patients between 50 and 80 years, although there was a statistically significant decrease in ORIF and THA usage and a reflective increase in HA usage in this population as well. The fact that there has not been as substantial a shift in treatment trends for this large age group is potentially due to the wide variations in comorbid conditions and the functionality that abounds in this age group.1
The limitations of the current study are those inherent with a retrospective database analysis. The reliance on accurate coding brings up a potential for error; however, it is unlikely that comorbidities and outcomes are undercoded as hospitals are incentivized to input values that increase the acuity and thus reimbursement for each hospital stay.35 The database also relies on the ICD-9 procedural and diagnostic codes, which are not as specific as the currently adopted ICD-10 codes; hence, we are unable to distinguish between different forms of internal fixation, for example intramedullary nailing versus dynamic hip screw. This also precludes us from including other critical data such as degree of fracture displacement, cemented versus uncemented implantation, surgical approach for arthroplasty, and functional outcomes of individual patients. Moreover, the database used, although the largest inpatient sample available for analysis, represents only approximately 20% of hospitals nationwide. In addition, as patients cannot be tracked over time within the database, we are limited to outcomes in the perioperative period captured in a single hospital stay and cannot identify readmissions. Finally, our analysis is limited to the years 1990 to 2007 because of an increase in the relative standard error of the database in more recent years. Although this results in data that are not the most current, we believe that this study provides valuable insight regarding the trends in surgical treatment and acute postoperative outcomes of these injuries that have hitherto not been reported. To limit the inherent biases and the limitations within this study, prospective, randomized studies with long-term follow-up comparing outcomes across modes of treatment are needed to definitively determine the optimum form of treatment for this fracture type.
CONCLUSION
This is the largest study to date reporting on national trends in the surgical treatment and outcomes of the femoral neck fracture population. Orthopedic surgeons performing THA should be aware that the femoral neck fracture population is changing and at higher risk for perioperative complications. The advent of bisphosphonate therapy has been suggested as a possible reason for the decrease in fragility fractures and why a larger proportion of the femoral neck fracture population is now >80 years.36,37 With an aging population at a higher risk for perioperative complications, clinicians must take special care in choosing the appropriate surgical intervention that will give their patients the best functional outcome while minimizing the risk of surgical complications. Orthopedic surgeons should weigh the added risk associated with THA in this population.