Umeå universitets logga

Objectives: To determine whether dementia could explain the association between poor olfactory performance and mortality risk within a decade-long follow-up period. Design: Prospective cohort study. SettingBetula Study, Umea, Sweden. ParticipantsA population-based sample of adult participants without dementia at baseline aged 40 to 90 (N = 1,774). Measurements: Olfactory performance using the Scandinavian Odor-Identification Test (SOIT) and self-reported olfactory function; several social, cognitive, and medical risk factors at baseline; and incident dementia during the following decade. Results: Within the 10-year follow-up, 411 of 1,774 (23.2%) participants had died. In a Cox model, the association between higher SOIT score and lower mortality was significant (hazard ratio (HR) = 0.74 per point interval, 95% confidence interval (CI) = 0.71-0.77, P < .001). The effect was attenuated, but remained significant, after controlling for age, sex, education, and health-related and cognitive variables (HR = 0.92, 95% CI = 0.87-0.97, P = .001). The association between SOIT score and mortality was retained after controlling for dementia conversion before death (HR = 0.92, 95% CI = 0.87-0.97, P = .001). Similar results were obtained for self-reported olfactory dysfunction. Conclusion: Poor odor identification and poor self-reported olfactory function are associated with greater likelihood of future mortality. Dementia does not attenuate the association between olfactory loss and mortality, suggesting that olfactory loss might mark deteriorating health, irrespective of dementia.