The objective of this study was to obtain external validation of the only available midlife dementia risk score cardiovascular risk factors , aging and dementia study (CAIDE) constituting age, education, hypertension, obesity, and hyperlipidemia in a larger, more diverse population. Our second aim was to improve the CAIDE risk score by additional midlife risk factors.
This retrospective cohort study was conducted in an integrated health care delivery system. A total of 9480 Kaiser Permanente members who participated in a health survey study (age range, 40–55 years) from 1964 to 1973 were included in this study. Dementia diagnoses from primary care and medical specialist visits were collected from January 1, 1994 to January 16, 2006, using International Classification of Diseases 9 codes 290.0, 290.1 for “possible dementia,” and 331.0 and 290.4 for “specialist confirmed dementia.” Risk model prediction and validation were examined with the C statistic, net reclassification improvement, and integrated discrimination improvement. Dementia risk per sum score was calculated with Kaplan-Meier estimates.
A total of 2767 participants (25%) were diagnosed with any type of dementia, of which 1011 diagnoses (10.7%) were specialist-confirmed diagnoses. Average time between midlife examination and end of follow-up was 36.1 years. The CAIDE risk score replicated well with a C statistic of 0.75, quite similar to the original CAIDE C statistic of 0.78. The CAIDE score also predicted well within different race strata. Other midlife risk factors (central obesity, depressed mood, diabetes mellitus, head trauma, lung function, and smoking) did not improve predictability. The risk score allowed stratification of participants into those with 40-year low (9%) and high (29%) dementia risk.
A combination of modifiable vascular risk factors in midlife is highly predictive of the likelihood of dementia decades later. Possible dementia prevention strategies should point to a life course perspective on maintaining vascular health.
2013. Vol. 10, no 5, 562-570 p.