Objective: Calculation of incidence of dementia and AD, including eases in
the earliest phases of the diseases. Background: Establishment of incidence
estimates is important for the future planning of the health care system,
and incidence studies can offer insights into risk factors. Methods: A tota
l of 5,237 persons age 65 to 84 years were randomly drawn among people livi
ng in the municipality of Odense, Denmark. Of this sample 3,086 persons wer
e eligible for the incidence study. All participants were examined with CAM
COG, the cognitive section of The Cambridge Examination for Mental Disorder
s of the Elderly (CAMDEX), and the follow-up period was 2 years. Using mult
iple linear regression, the CAMCOG cutoff score was individualized to detec
t even minor cognitive decline with optimal precision. Possibly demented pe
rsons were further examined with the remaining part of the CAMDEX and neuro
psychological tests. AD was diagnosed according to National Institute of Ne
urological and Communicative Disorders and Stroke-Alzheimer's Disease and R
elated Disorders Association criteria for probable AD, and vascular dementi
a and dementia of other types were diagnosed according to Diagnostic and St
atistical Manual of Mental Disorders (3rd ed., revised) criteria for dement
ia. Finally, the severity of dementia was determined according to the Clini
cal Dementia Rating scale. Results: The incidence rate for very mild to sev
ere dementia was 29.5 per 1,000 person-years and 20.9 for AD, and the rates
were similar for men and women. Conclusion: Application of an individualiz
ed cutoff for the screening instrument resulted in detection of a substanti
al number of cases with very mild dementia, which subsequently resulted in
higher incidence rates than those reported in most other studies.