OTA 1997 Posters - Hip Fractures

Five-Year Mortality after Hip Fracture: An Analysis of Predictive Factors

Kenneth A. Egol, MD, Kenneth J. Koval, MD, Gina B. Aharonoff, MPH, Joseph D. Zuckermann, MD

New York, New York, USA

Purpose: The reported overall mortality in elderly patients one year following hip fracture ranges from 14% to 36%; much less information is known about the mortality rate after hip fracture with longer follow-up. The purpose of this study was to determine the two-, three-, four-, and five-year mortality rates following hip fracture in an ambulatory, community-dwelling, cognitively intact elderly population and to examine the predictors of five-year mortality.

Methods: Four hundred fifty-one patients age 65 or older, who were previously ambulatory and home dwelling, and who had femoral neck or intertrochanteric hip fracture of nonpathological origin were admitted between 1/1/88 and 12/31/91. Patients with moderate to severe dementia were excluded. All patients were identified at the time of admission and were prospectively followed. All patients had operative fracture treatment and followed a similar postoperative protocol consisting of patient mobilization out of bed on day 1 with full unrestricted weightbearing. All patients were contacted yearly postfracture or until death by one of two trained interviewers and follow-up information obtained. The predictor variables examined in this study were patient age, sex, number of associated comorbidities, ASA rating of operative risk, fracture type, type of surgery, and the development of an in-hospital postoperative complication. To determine the association between potential predictors and mortality, multiple logistic regression was used, with p <.05 considered significant.

Results: Of the 451 patients studied, at five-year follow-up, 137 patients (30.4%) were known dead and 25 (5.5%) were lost to follow-up; the mortality rate at 1, 2, 3, 4, and 5 years was 13.3%, 18.8%, 22.6%, 27.7%, and 30.4%, respectively. The mortality rate was almost twice as high for males (43.5%) as for females (27.0%). The mortality rate of patients who were 85 years or older (50.8%) was more than twice that of the younger age group (23. 1 %). In multivariate analyses, the factors predictive of increased five-year mortality were: patient age of 85 or older (p <.001), ASA rating 3 or 4 (p <.001), male sex (p =.04), and the development of one or more in hospital postoperative complications (p < .001).

Discussion: Although many studies have documented the one-year mortality following hip fracture, this study provides information about mortality rate for the first five years after hip fracture. By excluding institutionalized patients and those with severe dementia, our patient population represents a lower "at-risk" group than other studies. Nevertheless, a 30% mortality rate at five years was identified. In addition, the identification of factors predictive of mortality provided important information to stratify hip fracture patients based upon the presence of specific risk factors.