Page 503

Karger_ESC London_2013

London, United Kingdom 2013 Correlation (r2) -0.451 -0.583 -0.158 -0.421 Sensitivity 0.42 0.321-0.524 0.373 0.261-0.5 0.043 0.016-0.111 0.242 0.155-0.358 Poster Session Red Cerebrovasc Dis 2013; 35 (suppl 3)1-854 503 410 Behavioral disorders and post-stroke dementia Cog-4 has limited diagnostic test accuracy and validity in stroke survivors R. Lees1, S. Corbet2, L. Lua3, C. Johnston4, E. Melling5, Y. Miao6, E. Moffitt7, G. Shaw8, J. Tan9, T.J. Quinn10 Institute of Cardiovascular and Medical Sciences, School of Medicine, University of Glasgow, Glasgow, UNITED KINGDOM1, Institute of Cardiovascular and Medical Sciences, School of Medicine, University of Glasgow, Glasgow, UNITED KINGDOM2, Institute of Cardiovascular and Medical Sciences, School of Medicine, University of Glasgow, Glasgow, UNITED KINGDOM3, Institute of Cardiovascular and Medical Sciences, School of Medicine, University of Glasgow, Glasgow, UNITED KINGDOM4, Institute of Cardiovascular and Medical Sciences, School of Medi-cine, University of Glasgow, Glasgow, UNITED KINGDOM5, Institute of Cardiovascular and Med-ical Sciences, School of Medicine, University of Glasgow, Glasgow, UNITED KINGDOM6, Insti-tute of Cardiovascular and Medical Sciences, School of Medicine, University of Glasgow, Glasgow, UNITED KINGDOM7, Institute of Cardiovascular and Medical Sciences, School of Medicine, University of Glasgow, Glasgow, UNITED KINGDOM8, Institute of Cardiovascular and Medical Sciences, School of Medicine, University of Glasgow, Glasgow, UNITED KINGDOM9,Institute of Cardiovascular and Medical Sciences, School of Medicine, University of Glasgow, Glasgow, UNIT-ED KINGDOM10 Introduction: Guidelines suggest all stroke survivors should be screened for cognitive impairment. The National Institutes of Health Stroke Scale (NIHSS) is a routinely used impairment scale. Cer-tain elements of NIHSS (orientation, executive function, language and inattention) have been sug-gested as a short cognitive screening test - Cog-4. We aimed to describe accuracy of Cog-4 against a cognitive assessment recommended for stroke survivors (Montreal Cognitive Assessment MoCA). Method: We assessed consecutive stroke unit admissions in two urban, teaching hospitals over 10 weeks. Independent, blinded assessors performed NIHSS and MoCA. We described sensitivity, specificity and correlation for total scores and for individual cognitive domains (executive function, language, orientation and combined visuospatial/executive function). We used conventional diag-nostic thresholds for MoCA (<26/30) and Cog-4 (≥1). Results: We assessed 129 participants; 122 had Cog-4 data and 105 MoCA; 50% male (n=65), median age:74 (IQR:64-85). Median Cog-4 was 0 (IQR:0-1.5); median MoCA 19 (IQR: 13-24). Cog-4 had a sensitivity of 0.4 (95% CI:0.3-0.5); specificity of 0.97 (95% CI:0.78-0.997) Positive Predictive Value:0.99. Negative Predictive Val-ue: 0.26 for MOCA defined cognitive impairment. Individual Cog-4 domains correlated with corre-sponding MoCA values but strength of association was modest (r2:-0.58 orientation; -0.42 language; -0.19 for combined visuospatial/executive function and no significant correlation for executive func-tion p=0.074). Conclusions: Many stroke survivors with cognitive problems would not be picked up by Cog-4 testing. Other brief cognitive screening tests may be better suited to acute stroke. The subtest correlations suggest Cog-4 may not be a true measure of the cognitive domains it claims to assess. Table 1: Accuracy comparisons between the Cog-4 and MoCA. -0.188 0.173 0.106-0.269 0.193 0.033 0.122-0.29 0.933 0.702-0.988 0.941 0.73-0.99 Estimate 95% CI Executive function/ inattention 0.412 <0.001 0.313-0.518 0.941 0.73-0.99 0.972 0.858-0.995 Estimate 95% CI Language 0.182 0.074 0.118-0.271 0.875 0.396-0.987 0.972 0.781-0.997 Estimate 95% CI Executive Function 0.532 <0.001 0.423-0.638 0.88 0.7-0.958 0.933 0.821-0.977 Estimate 95% CI Orientation 0.257 <0.001 0.168-0.372 0.986 0.833-0.999 0.972 0.781-0.997 Estimate 95% CI Cog-4 Significance (p value) Negative Predictive Value Positive Predictive Value Test Specificity


Karger_ESC London_2013
To see the actual publication please follow the link above