The validity of the Addenbrooke's Cognitive Examination-Revised (ACE-R) in acute stroke.

PURPOSE: The purpose was to examine the validity of the Addenbrooke's Cognitive Examination Revised (ACE-R) as a screening measure to detect cognitive impairment after stroke.

METHODS: Stroke patients in hospital were recruited and the ACE-R, which includes the Mini-Mental Status Examination (MMSE), was administered, followed by a battery of neuropsychological tests, which served as the 'gold standard' for classification of cognitive impairment. The diagnostic validity of the ACE-R was determined by ROC analysis.

RESULTS: Of the 101 patients who completed the ACE-R, 61 also completed the neuropsychological assessment. Both the MMSE and the ACE-R were found to have inadequate diagnostic validity for the detection of overall cognitive impairment (MMSE AUC = 0.53, p > 0.05; ACE-R AUC = 0.53, p > 0.05). The ACE-R subscales predicted impairment in specific cognitive domains significantly better than chance; Visuospatial (AUC = 0.71, p < 0.05), Fluency (AUC = 0.72, p< 0.05) and Attention and Orientation (AUC = 0.80, p < 0.05). However, no cut-off score for any subscale gave both adequate levels of sensitivity and specificity for the detection of impairment in specific areas of cognitive functioning.

CONCLUSIONS: The ACE-R was not a suitable measure to screen for overall cognitive impairment in acute stroke patients, but was able to detect impairment in visuospatial, attention and executive domains.