Incidence of dementia related to medical, psychological and social risk factors: A longitudinal cohort study during a 25-year period

This study is based on an entire cohort (n = 192) of 67-year-old persons born in 1902 and 1903 and living in a community in Southern Sweden. All subjects participated in interviews, psychological tests, and medical examinations. All contacts with primary health care and social services were recorded, as were death diagnoses. The cohort has been followed since 1969, with nine examinations until age 92. The incidence rate (per 1000 person years) of dementia between 67 and 92 years of age, including those alive at age 92 as well as those deceased prior to 92, was 2.8 in the first five-year period, increasing up to 44.5 between ages 87 and 92. Altogether 16 % developed dementia during the period. The mean age for diagnoses of dementia was 80 years. Fifteen presumptive social, medical and psychological risk factors for dementia have been applied (Cox regression analyses). Neither gender nor education were significant risk factors for dementia. Non-smokers tended to have a higher risk for dementia, but not a significantly higher risk. Normal blood pressure tended to increase the risk for dementia nearly significantly (p = 0.07). Diabetes increased the risk for dementia, however not significantly (p = 0.08). Using Cox regression analysis with dementia as a time-dependent covariate, it was found that dementia significantly increased the relative risk for death more than three times (CI 1.9-5.7), controlling for gender, smoking, diabetes, blood pressure, and education. Four cognitive tests were not found to be risk factors for dementia; neither were any of the medical parameters or social network variables.