Design
Patients were divided into 4 grades of increasing extent of vascular lesions on CT. The frequency of VaD was compared between these grades.

Setting
A university department of neurology.

Patients
Forty-two consecutive patients who underwent neuropsychological assessment for possible dementia and who had a CT scan performed within 6 months following any stroke causing dementia. Patients with delirium, severe aphasia, and motor and/or sensory deficits that impaired neuropsychological testing and patients with mass lesions or nonvascular white matter disease shown on CT were excluded.

Main Outcome Measure
The National Institute of Neurological Disorders and Stroke and the Association Internationale pour la Recherche et l'Enseignement en Neurosciences criteria for probable VaD.

Results
The frequency of VaD was greater in patients with grade 1 (7 [50%] of 14, P=.01), grade 2 (2 [50%] of 4, P=.2), and grade 3 (7 [78%] of 9, P=.002) scans than the frequency of VaD with grade 0 scans (1 [7%] of 15). There was a linear association of the frequencies of VaD between imaging grades (P=.0008). In a subgroup of patients with neuropsychological deficits caused by cerebrovascular disease, there was a linear association of the severity of the deficits between imaging grades (P=.007).

Conclusions
We conclude that our criteria can separate CTs into increasing levels of support for a diagnosis of VaD. The extent of vascular lesions on CT reflects the severity of associated neuropsychological deficts.