Progression of cognitive impairment in stroke patients P.S. Sachdev, MD, PhD, FRANZCP; H. Brodaty, MD, FRACP, FRANZCP; M.J. Valenzuela, BSc (Hons); L. Lorentz, M Clin Psychol, MAPS; and A. Koschera, PhD Abstract—Objective: To examine the progression of neuropsychological deficits in stroke patients with and without cognitive impairment. Methods: The authors assessed the Sydney Stroke Study cohort 1 year after index assessment with detailed neuropsychological and medical–psychiatric assessments. The neuropsychological tests were classified into cogni- tive domains, and composite z-scores adjusted for age and education. Changes in cognitive test scores were compared between groups and predictors of cognitive change examined. Results: Patients (n = 128) had a mean decline of 0.83 (SD 2.2) points on the Mini-Mental State Examination (MMSE) compared to an increase of 0.76 (1.3) in controls (n = 78) (p  0.0001), and a small but significant decline in informant ratings of function and cognition. The decline on a composite index of cognitive function was not significantly different in the groups after correction for age, education, and index assessment cognitive function. Stroke/transient ischemic attack patients, however, had greater decline in verbal memory and visuoconstructive function. The occurrence of an interval stroke (n = 14) significantly increased the cognitive decline to a mean 2.0 points on the MMSE. The rate of change had a significant correlation (r = 0.24) with white matter hyperintensity volume at index assessment. On regression analysis the only predictor of cognitive change was years of education, which had a protective function. Conclusions: Subjects with cerebrovascular disease have a slow decline in cognitive functioning in the absence of further cerebrovascular events, although the occurrence of such an event accentu- ates the dysfunction. Education plays a protective role. NEUROLOGY 2004;63:1618 –1623 Cognitive impairment is commonly present in stroke patients, with about a quarter meeting criteria for vascular dementia (VaD). 1,2 Most published studies have assessed cognitive function in ischemic stroke patients at 3 months or more after the stroke, with the expectation that cognitive deficits have stabilized at this stage, after a period of initial recovery. 2 The few studies that have assessed stroke patients longi- tudinally over extended periods 3-6 report further re- covery in a few subjects but cognitive decline in the overall sample. The rate of change and the detailed neuropsychological profile of this decline, i.e., the cognitive domains most likely to change, have not been adequately investigated. The course of vascular cognitive impairment (VCI) has been recently reported. 7 VCI is referred to here as a superordinate construct that includes all levels of cognitive impairment of vascular origin. 8 Many but not all patients involved in the studies of pro- gression of VCI had a history of stroke or TIA. These studies suggest that deficits of VCI are progressive, although the changes are variable depending upon the sample being studied. While earlier studies sug- gested that mortality rates were higher in VaD com- pared to Alzheimer disease (AD), 9 the rate of cognitive decline in VaD appears to be slower than AD. 10 This is supported by the placebo arms of the recent trials of cholinesterase inhibitors in VaD. 11-13 A detailed examination of the progression of cogni- tive deficits in patients with ischemic vascular brain lesions is therefore of considerable interest. We report a longitudinal study of stroke patients who were initially assessed at 3 to 6 months after a stroke or TIA and followed up a mean 14.6 months later to determine the profile and determinants of cognitive decline over 1 year. Methods. Sample. Subjects were consecutive patients, admit- ted to two large teaching hospitals affiliated with the University of New South Wales, who had recently had an ischemic stroke or TIA and had agreed to participate in the Sydney Stroke Study. 1,14 Subjects were recruited over a 38-month period between May 1997 and June 2000. Subjects were aged 49 to 87 years, did not have a diagnosis of dementia or other neurologic disorder prior to the stroke/TIA, did not have severe aphasia as a limiting factor for assessment, and were well enough to consent to participate. Healthy control subjects were unpaid age-matched volunteers, re- cruited from the same neighborhood as the stroke/TIA subjects. They had no history of stroke, TIA, or other neurologic or psychi- atric disorder. In all, 1,050 patients were screened and 252 consid- ered eligible and recruited. At the time of detailed assessment 3 to 6 months later, 210 patients (176 stroke and 34 TIA) and 103 controls were included in the study. Of these, 170 patients and 96 From the School of Psychiatry (Drs. Sachdev and Brodaty, and M.J. Valenzuela and L. Lorentz), University of New South Wales; and Neuropsychiatric Institute (Dr. Sachdev and M.J. Valenzuela), Academic Department for Old Age Psychiatry (Drs. Brodaty and Koschera, and L. Lorentz), the Prince of Wales Hospital, Sydney, Australia. Supported by grants from the National Health and Medical Research Council of Australia, the Rebecca Cooper Foundation, and Fairfax Family Foundation, and a Fellowship from the NSW Institute of Psychiatry. Received March 16, 2004. Accepted in final form July 8, 2004. Address correspondence and reprint requests to Professor P. Sachdev, NPI, Prince of Wales Hospital, Barker Street, Randwick NSW 2031, Australia; e-mail: p.sachdev@unsw.edu.au 1618 Copyright © 2004 by AAN Enterprises, Inc.