Incidence of dementia in a Munich community sample of the oldest old.

In a two-wave community study a representative sample of 402 very old people (older than 85 years) was restudied 1 year later. Four instruments in the clinical examination were used for case identification: (a) the Geriatric Mental State Interview (GMS-A); (b) the Structured Interview for the Diagnosis of Dementia (SIDAM); (c) the Global Deterioration Scale (GDS); and (d) the Mini Mental State Examination (MMSE). The clinical examination was performed by the interviewing physician who made a diagnosis according to DMS-III-R. The focus of the present study is on the (true) incidence of dementia in a representative community sample. The establishment of incidence rates is particularly important for dementia because the prevalence of dementia is affected by the length of survival, which is reduced in dementia and with increasing age. The annual incidence rates per 1000 person years on the basis of the SIDAM DSM-III-R were 116.6 for all cases at risk, 113.6 for those aged 85-89 years, 112.5 for those aged 90-94 years and 235.7 for those aged 95 years and older at first assessment (t1). Incidence rates based on the other methods of assessment are reported. In order to obtain the most meaningful estimate of incidence rates a compound dementia diagnosis was defined. According to this the annual incidence rate per 1000 person years was 144.1 for all persons at risk. The incidence rate tended to be higher in the older-age cohorts: It was 126.2 for those aged 85-89 years, 193.1 for those aged 90-94 years and 295.5 for those aged 95 years and older. In comparison with the literature the incidence rates were high. The results are plausible when the very old age of the sample is taken into account.