Risk factors for mortality in geriatric hip fractures: a compressional study of different surgical procedures in 785 consecutive patients.

Background: Only a little is known about whether type of surgical intervention has an effect on mortality of these patients. Our primary objective was to assess whether different type of surgical procedures has an effect on mortality among elderly patients with hip fracture. A secondary objective was to examine factors that are related to mortality in our patient population. Our hypothesis is that type of surgical procedure, especially external fixation, should have an influence on mortality outcomes.
Methods: We included 785 patients age 65 years or older, with hip fractures. Operative treatment consisted of external fixation, internal fixation, total hip arthroplasty and hip hemiarthroplasty. Age, gender, type of fracture, type of surgery performed, American Society of Anesthesiology (ASA) grade, clinical comorbidities, anesthesia type, blood transfusion requirement, time to surgery, intensive care unit requirement, operation length and length of hospital stay and number of comorbidities were documented.
Results: During the study period, 785 patients (262 male, 523 female) were included to study, Overall mortality rate was 37.2 % (292/785). Their age ranged between 65 and 100 years (mean 81). Surgery type Kaplan-Meier cumulative mortality curves suggested no significant difference between four different types of surgery groups (p = 0.064). Transfusion requirement was significantly lower in external fixation group comparing to other groups (p = 0.014). Cox regression analysis showed the number of comorbidities 2 and ≥ 3 (p = 0.0027, p = 0.015), transfusion requirement (p = 0.0001), ASA 4 (p = 0.016) to be significant predictors of mortality.
Conclusions: Transfusion requirement, ASA grade 4 and having more than two comorbidities are risk factors for mortality in geriatric hip fractures. Type of surgical intervention and fracture type had similar mortality rates in our patient population.