This doctoral thesis investigates the effect of somatic disorders on dementia, Alzheimer s disease (AD) and brain aging in late-life. The data for the studies are provided by the Kungsholmen Project (Studies I and II) and the Faenza Project (Studies III and IV). The Kungsholmen Project is a population-based longitudinal study on aging and dementia carried out on 75+ years old people, living in Stockholm, Sweden. The Faenza Project is a cross-sectional population-based study on dementia and cognitive impairment involving persons over 60 years, living in a wealthy area in Northern Italy. The major findings are summarized below. Study I. The hypothesis that anemia may increase the risk of dementia and AD was investigated using three-year follow-up data. Subjects who fulfilled WHO criteria for anemia (hemoglobin < 130 g/L for men and < 120 g/L for women) were at higher risk of developing dementia (HR: 1.6, 95% CI: 1.1-2.4). In persons with good baseline cognition (Mini-Mental State Examination (MMSE) >= 26, n=1,139) the association was stronger, and still significant after adjustment for conditions potentially related to anemia and dementia such as chronic diseases, inflammatory markers, and indicators of nutritional status. Using different definitions for anemia, a dose-response relationship between low hemoglobin concentration and incident dementia was observed: the lower the cut-off, the higher the probability of dementia. Study II. The relationship between body weight and dementia (or AD) was investigated using three-, six-, nine-year follow-up data. No association was detected between being underweight and dementia. Even after extensive adjustment, subjects with Body Mass Index (BMI) >= 25 had a lower risk of developing dementia than normal-weight subjects over nine years (HR: 0.8, 95% CI: 0.6 0.96). These results were confirmed also when the analysis was restricted to the cases occurring between three and nine, or between six and nine years of follow-up. Severe changes in BMI between baseline and three years of follow-up were associated with incident dementia in the following three years. Study III. The independent and combined effect of medical and social factors on cognitive status was investigated in the Faenza Project. A person was diagnosed with Cognitive Impairment No Dementia (CIND) if he/she scored two or more standard deviations lower than non-demented subjects on the adjusted MMSE. The diagnostic procedure identified 402 CIND cases (5.4%). Diabetes (OR: 1.6, 95% CI: 1.2-2.2), stroke (OR: 1.9, 95% CI: 1.4-2.6), and depressive symptoms (OR: 1.9, 95% CI: 1.4-2.7) were the most relevant associated medical comorbidities. The strength of the association increased when these conditions occurred in combination. Low education (OR: 1.8, 95% CI: 1.1-2.9), low socioeconomic status (OR: 1.5, 95% CI: 1.1-2.1), and unmarried status (OR: 1.7, 95% CI: 1.2-2.5) were also independently associated with CIND. Medical and social factors had a strong synergistic effect. Persons with at least one medical and one social factor were six times more likely to be diagnosed as CIND. Study IV. To investigate the relation between perceived cognitive decline and mental and somatic comorbidity we considered the non-demented, cognitively unimpaired cohort of the Faenza Project (n=6,825). According to the Global Deterioration Scale (GDS), 20.1 % and 8.3% of participants reported very mild or mild cognitive decline, respectively. Diseases significantly associated with perceived cognitive decline were stroke (OR: 1.8; 95% CI: 1.3-2.3), malignancy (OR: 2.7, 95% CI: 1.3-5.7), cardiovascular disorders (OR: 1.7, 95% CI: 1.2-2.3), diabetes (OR: 1.6; 95% CI: 1.2-2.2), and depressive and anxiety symptoms (OR: 3.7, 95% CI: 2.4-4.2; OR: 1.6, 95% CI: 1.2-2.1). Mental, cardiovascular, and respiratory diseases accounted for the discrepancy between perceived cognitive decline and cognitive performance. Conclusions. The findings support the hypothesis that several somatic disorders might affect brain aging and dementia. This thesis suggests that late-life anemia is a risk factor for dementia, that diabetes, stroke, and depressive symptoms are clinical correlates of CIND, and that mental and somatic morbidity accounts for the discrepancy between perceived cognitive decline and cognitive performance. Finally, we do not confirm being overweight in late-life as a risk factor for dementia, but we suggest that weight loss is a good predictor of incipient dementia.

ISBN: 978-91-7409-425-1

© Anna Rita Atti, 2009