Fractures as predictors of excess mortality in the aged-a population-based study with a 12-year follow-up.

INTRODUCTION AND OBJECTIVE: The association between fractures and excess mortality in old age is ambiguous. The objective of this study was to analyze the long-term gender-specific association between fractures and mortality among older persons by controlling several survival related confounders.

METHODS: A population-based prospective cohort study in the municipality of Lieto, south-western Finland. Data on health, health behaviour, fractures, and mortality in 482 men and 695 women aged 65 or older was collected from 1991 until 2002. The Cox Proportional Hazards regression model with fractures as time-dependent variables was used in the analyses.

RESULTS: During the 12-year follow-up, 295 (25%) persons sustained at least one fracture. Sustaining any kind of fracture was related to excess mortality both in men (age-adjusted Hazards Ratio, HR 2.2, 95% confidence intervals, CI 1.6-3.1) and in women (HR 1.6, 95% CI 1.3-2.1). In the multivariate analyses, hip fractures in men (HR 8.1, 95% CI 4.4-14.9) and in women (HR 3.0, 95% CI 1.9-4.9), and proximal humerus fractures in men (HR 5.4, 95% CI 1.6-17.7) were related to increased mortality.

CONCLUSION: A hip fracture was a powerful independent predictor of long-term excess mortality in both genders but the risk in men was more than 2-fold compared to women. Proximal humerus fractures were associated with increased mortality in men. Actions to improve prevention, acute care and rehabilitation of fractures are needed in order to reduce excess mortality in older people.