Abstract 2322: Validity of the NINDS-CSN VCI Neuropsychological Protocols is Supported in a Sample of Ischemic Stroke Patients

Introduction: The NINDS and the Canadian Stroke Network co-sponsored a 2005 meeting to harmonize data collection methods for patients with suspected vascular cognitive impairment (VCI). At this meeting, the neuropsychology group proposed 60 Minute (60M), 30 Minute (30M) and 5 Minute (5M) neuropsychological protocols. This study tests the validity of the protocols by comparing protocol performance of an ischemic stroke patient sample to matched controls. Hypotheses: 1. The 60M, 30M and 5M protocol scores will be more impaired in the ischemic stroke group than the matched controls; 2. The 60M Executive subscale score will be a more sensitive and specific factor in differentiating the two groups than the 60M Memory, Language and Spatial subscale scores, as measured by ROC area under the curve scores; and 3. The 60M protocols will be inversely related to ADL function, as measured by the Functional Rating Scale (FRS) and the Disability Assessment for Dementia (DAD) Methods: 1. ANCOVA and ROC analyses are used to compare 46 ischemic stroke patients (mean age: 66.4 (12.8), 62% male, 14.4 (3.0) years of education, 60% LH strokes, mean Rankin score: 2.14 (0.69), mean MMSE score: 27.2 (2.9)) with 76 controls of similar age, gender ratio, education, and Barona (estimate of premorbid IQ) score on the 60M, 30M and 5M protocols. Because patients reported more CESD-depression symptoms than controls (12.33 vs. 7.3, p < .01), it is entered as a covariate. 2. 60M protocol scores are entered for correlation with FRS and DAD scores Results: 1. All three protocol scores are significantly lower in patients than in matched controls (F statistics range from 15.7 to 50.5; all p values < .000; partial eta2 values range from .14 to .31). 2. ROC analyses show the 60M Executive subtest to be the most sensitive and specific, followed by the Memory, Language, and Spatial subtests (AUC values: .86, .75, .70, .67, respectively). 3. The 60M protocol score correlates with DAD Total (p < .05) and DAD IADL (p < .01) scores and is inversely correlated with FRS Total (p < .05) score 4. Domain specific cognitive and functional score relationships are noted between 60M Executive score and FRS Problem Solving rating score (p < .05), and between 60M Memory and FRS Memory rating score (p < .05). The correlation between 60M Language and FRS Language rating is not significant. Conclusions: 1. The validity of the VCI Harmonization protocols is supported by their ability to differentiate an ischemic stroke group from a matched normal control group. 2. The 60M Executive subscale shows higher sensitivity and specificity than the Memory, Language and Spatial subscales, which supports the important role of executive function in VCI and further validates the scale 3. The 60M scale and its Executive and Memory subscales relate to functional disability in a post-ischemic stroke sample, further supporting the validity of these scales.