Predictors of mortality after isolated proximal humeral fractures in elderly patients

Background: The overall incidence of proximal humeral fractures is steadily growing, but little is known about the mortality rates after these fractures in elderly patients. As life expectancy continues to increase, knowledge of mortality rates and predictors of mortality for these fractures will become progressively more important. The purpose of this study was to present the 30-day and 1-year mortality of elderly patients who sustained isolated proximal humeral fractures and to evaluate potential risk factors for increased mortality. Methods: We identified 205 patients older than 60 yr with isolated proximal humeral fractures. Mean age was 74.3 yr (range, 60-99 yr) and mean follow-up was 3.9 yr (range, 5 days-8 yr). Thirty-day and 1-year mortality rates were calculated. Age, gender, fracture classification, patient disposition, initial treatment, Charlson Comorbidity Index (CCI) and van Walraven modified Elixhauser Comorbidity Index (ECI) were analyzed in Cox Proportional Hazards models to estimate effect on mortality. Results: Fifty-six percent of patients sustained Orthopaedic Trauma Association (OTA) 11-A fractures, 43% were OTA 11-B, and 1% were OTA 11-C. One hundred and sixty-six patients (81%) were treated nonoperatively. The 30-day mortality rate was 2.0%; 1-year mortality rate was 4.4%. Each additional year of patient age increased mortality 12% (hazard ratio [HR] = 1.12, 95% confidence interval [CI]: 1.07, 1.17). Women did not differ from men (HR = 1.74, 95% CI: 0.80, 3.77). Controlling for age and gender, there was no increased mortality in those admitted at time of injury (HR = 1.89, 95% CI: 0.92, 3.88), or those with operative fixation (HR = 0.78, 95% CI: 0.27, 2.29). After adjusting for age, neither CCI (HR = 1.07, 95% CI: 0.95, 1.23) nor ECI (HR = 1.01, 95% CI: 0.97, 1.06) affected mortality. Conclusions: The 1-year mortality for elderly patients with isolated proximal humeral fractures was 4.4%, lower than previously reported. Age increased mortality but comorbidity indices did not.