Incidence of dementia: evidence for an effect modification by gender. The ILSA Study

BACKGROUND: Gender differences for incidence of dementia among elderly people have been usually investigated considering gender as a predictor and not as a stratification variable. METHODS: Analyses were based on data collected by the Italian Longitudinal Study on Aging (ILSA), which enrolled 5,632 participants aged 65-84 years between 1992 and 2000. During a median follow-up of 7.8 years, there were 194 cases of incident dementia in the participants with complete data. Cox proportional hazard models for competing risks, stratified by sex, were defined to determine risk factors in relation to developing dementia. RESULTS: The incidence rate of dementia increased from 5.57/1,000 person-years at 65-69 years of age to 30.06/1,000 person-years at 80-84 years. Cox proportional hazard models for competing risks of incidence of dementia and death revealed that, among men, significant risk factors were heart failure, Parkinson's disease, family history of dementia, mild depressive symptomatology and age, while triglycerides were associated with a lower risk of developing dementia. Significant risk factors in women were age, both mild and severe depressive symptomatology, glycemia ≥109 mg/dL, and a BMI < 24.1 kg/m². Even as little as three years of schooling was found to be a significant protective factor against the incidence of dementia only for women. CONCLUSIONS: Our results suggest that there is an effect modification by gender in our study population in relation to the association between low education level, lipid profile, BMI, and glycemia and dementi