Sensitivity and specificity of 5 min cognitive screening tests in patients with acute coronary syndrome

Abstract

Abstract Aims This study aimed to determine the sensitivity and specificity of the National Institute of Neurological Disorders and Stroke (NINDS) and the Canadian Stroke Network (CSN) brief (5 min) screen composed of three items of the Montreal Cognitive Assessment (MoCA), in acute coronary syndrome (ACS) patients during hospital admission, relative to the full MoCA and potential alternative combinations of other items. Methods and results Participants were consecutively recruited during ACS admission and administered the MoCA before discharge. The three NINDS–CSN screen items were extracted, collated and compared to the full MoCA. Receiver operator characteristic (ROC) curves were created to determine the sensitivity, specificity, and appropriate cut-off scores of the screens. The mean age of the sample (n = 81) was 63.49 [standard deviation (SD) 10.85] years and 49.4% screened positive for cognitive impairment. The NINDS–CSN mean score was 9.22 (SD 2.09 of the potential range 0–12). Area under the ROC (AUC) indicated high accuracy levels for screening for cognitive impairment (AUC = 0.89, P < 0.01, 95% confidence interval 0.82, 0.96) with none of the alternative combination screens performing better on both sensitivity and specificity. A cut-off score of ≤10 on the NINDS–CSN protocol provided 83% sensitivity and 80% specificity for classifying cognitive impairment. Conclusion The NINDS–CSN protocol presents an accurate, feasible screen for cognitive impairment in patients following ACS for use at the bedside and potentially also for telephone screens. Diagnostic accuracy should be confirmed using a neurocognitive battery.