Evaluation of various methods of assessing symptoms of cognitive impairment and dementia.

Background and Purpose: The effect of different diagnostic criteria for detecting dementia in both epidemiological and stroke cohort studies has been shown, but comparison between different assessment methods has only seldom been done. We compared both assessment methods and diagnostic criteria for dementia in a large well-defined stroke cohort.
Subject and Methods: A group of 227 of 486 patients aged 55 to 85 years who 3 months after ischemic stroke completed a comprehensive neuropsychological test battery, structured clinical mental status examination of defined cognitive domains with expanded Mini-Mental State Examination. The criteria for dementia were those of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III, DSM-III-R) and the National Institute of Neurological Disorders and Stroke-Associated Internationale pour la Recherche et l'Enseignement en Neurosciences (NINDS-AIREN).
Results: The main differences between clinical and neuropsychological examinations were seen in memory functions: clinically 24.7% and neuropsychologically 54.2% had impairment in short-term memory and 10.4% versus 5.3% in long-term memory. Accordingly, the prevalence of dementia varied greatly: It was clinically 14.1% by DSM-III, 9.7% by DSM-III-R and 8.4% by NINDS-AIREN criteria. The corresponding frequencies based on neuropsychological evaluation were 27.3%, 4.0% and 25.6%. Between these 3 diagnostic criteria the concordance varied in clinical testing between 59.4%-68.8% (kappa 0.72-0.79) and in neuropsychological testing between 14.5%-81.1% (kappa 0.20-0.86). The concordance between clinical and neuropsychological testing was 56.8% (kappa 0.42) by DSM-III, 31.6% (kappa 0.35) by DSM-III-R and 25.5% (kappa 0.24) by NINDS-AIREN.
Conclusions: The frequency of poststroke dementia and cognitive decline varied sharply when different systems of diagnostic classification and methods were used. This may have serious influences on investigation and treatment of patients. We underline the importance of further debate and studies to refine the categories of cognitive impairment used in the setting of CVD.