Different Classification Systems Yield Different Dementia Occurrence among Nonagenarians and Centenarians

Literature data consistently show different prevalence estimates of dementia when different classification systems are used in the same population. Very few data are available for the oldest old of the elderly. We investigated the occurrence of dementia among 34 nonagenarians and centenarians according to four classification systems: the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders, third edition revised (DSM-III-R) and fourth edition (DSM-IV), the World Health Organization's International Classification of Diseases, 10th revision (ICD-10), and the Cambridge Examination for Mental Disorders of the Elderly (CAMDEX). Cognitive functioning, work, social function and independence in activities of daily living were evaluated by using an extensive neuropsychological examination. The prevalence (95% CI) of dementia was the following: 47.1% (95% CI 30.3-63.8) with the DSM-III-R criteria, 41.2% (95% CI 24.6-57.7) with the DSM-IV criteria, 29.4% (95% CI 14.1-44.7) with the ICD-10 criteria and 38.2% (95% CI 21.9-54.6) with the CAMDEX. The factors that best predicted disagreement between DSM-III-R and DSM-IV were calculation impairment and the presence or absence of personality changes. DSM-III-R and ICD-10 were differentiated by the weight given to executive functions that all have to be impaired according to ICD-10, whereas progressive deterioration differentiated CAMDEX from DSM-III-R. It should be noted that although the DSM-III-R diagnoses differ by a factor of 1.6 times from the ICD-10 and 1.2 times from the CAMDEX diagnoses, we are speaking about dementia, which is very frequent in nonagenarians and centenarians. Moreover, with regard to public health, an estimation of the number of subjects who will lose their autonomy is rather more useful and informative than simple prevalence figures of dementia by itself. In this light, classification systems, such as the ICD-10, that do not include impairment of social function as a criterion for assessing dementia become less adequate.