Clock drawing test: correlation with linear measurements of CT studies in demented patients.

OBJECTIVES: To investigate a presumed correlation between clock drawing ratings and linear measurements of computerized tomography (CT) studies in demented patients. DESIGN: Blinded evaluations of clock drawing tests and CT studies of elderly dementia patients were conducted by a geriatric psychiatrist and a neuroradiologist. SUBJECTS: Fifty-one community-dwelling elderly subjects meeting the criteria for DSM-IV diagnosis of dementia (Alzheimer's type dementia: N=31; vascular dementia: N=15; "mixed" type dementia: N=5). MATERIALS: Mini-Mental State Examination (MMSE), Cambridge Cognitive Examination (CAMCOG), Clinical Dementia Rating (CDR). CAMCOG derived scored clock drawings were evaluated using adaptations of Shulman et al.'s and Freedman et al.'s methods. CT studies were evaluated using six different linear measurements of brain atrophy described in the literature. RESULTS: Of the CT linear measurements, only the Cerebro-Ventricular Index-2 (CVI-2; bicaudate index) significantly correlated with clock drawing ratings (CAMCOG's clock r=-0.407, p=0.003; Shulman's method r=0.357, p=0.01, Freedman's method r=-0.413, p=0.003) in the dementia group. There was no significant correlation between CVI-2 with demographic (age), cognitive (MMSE, CAMCOG) and clinical (duration of illness, CDR) ratings. Alzheimer's patients generally maintained a significant correlation between CVI-2 and clock drawings, but vascular dementia patients did not; CVI-2 also correlated significantly with the Praxis subtest of the CAMCOG in dementia and Alzheimer's patients but not in the vascular dementia group. Similarly, multiple stepwise regression analysis showed that only CVI-2 but not the other radiological measures studied, was selected as the significant variable to correlated with clock drawing test ratings in the dementia group and Alzheimer's patients. Partial correlation analysis controlling for demographic and clinical variables shows that controlled variables had no significant effect on the relationship between clock drawing ratings and CVI-2. CONCLUSION: A single and easy to perform measure of caudate atrophy correlates specifically and consistently with impairments revealed in the clock drawing test and with a Praxis subtest, suggesting possible caudate involvement with clock drawings in dementia in general and of the Alzheimer's type in particular.