Clinical prediction of large vessel occlusion in anterior circulation stroke: mission impossible?

Abstract

Simple clinical scores to predict large vessel occlusion (LVO) in acute ischemic stroke would be helpful to triage patients in the prehospital phase. We assessed the ability of various combinations of National Institutes of Health Stroke Scale (NIHSS) subitems and published stroke scales (i.e., RACE scale, 3I-SS, sNIHSS-8, sNIHSS-5, sNIHSS-1, mNIHSS, a-NIHSS items profiles A-E, CPSS1, CPSS2, and CPSSS) to predict LVO on CT or MR arteriography in 1085 consecutive patients (39.4 % women, mean age 67.7 years) with anterior circulation strokes within 6 h of symptom onset. 657 patients (61 %) had an occlusion of the internal carotid artery or the M1/M2 segment of the middle cerebral artery. Best cut-off value of the total NIHSS score to predict LVO was 7 (PPV 84.2 %, sensitivity 81.0 %, specificity 76.6 %, NPV 72.4 %, ACC 79.3 %). Receiver operating characteristic curves of various combinations of NIHSS subitems and published scores were equally or less predictive to show LVO than the total NIHSS score. At intersection of sensitivity and specificity curves in all scores, at least 1/5 of patients with LVO were missed. Best odds ratios for LVO among NIHSS subitems were best gaze (9.6, 95 %-CI 6.765-13.632), visual fields (7.0, 95 %-CI 3.981-12.370), motor arms (7.6, 95 %-CI 5.589-10.204), and aphasia/neglect (7.1, 95 %-CI 5.352-9.492). There is a significant correlation between clinical scores based on the NIHSS score and LVO on arteriography. However, if clinically relevant thresholds are applied to the scores, a sizable number of LVOs are missed. Therefore, clinical scores cannot replace vessel imaging.

Response by Maïer and Turc to Letter Regarding Article, “Clinical Scales Do Not Reliably Identify Acute Ischemic Stroke Patients With Large-Artery Occlusion”

“…2 Our results are consistent with those of other publications on this topic, including a recent study by Hastrup et al themselves. 3,4 However, their interpretation differs from ours. Hastrup et al underline that several clinical scales have a good accuracy (70%-79%) to predict LAO and argue that such scales could improve prehospital triage.…”

“…We agree that patients with LAO should be sent to a comprehensive stroke center rather than to a primary stroke center lacking neurointerventional facilities. However, we, like others, 4 remain skeptical about the impact of clinical scales to improve triage in the field.…”

“…However, the optimal National Institute of Health Stroke Scale score cutoff maximizing the sum of sensitivity and specificity varies greatly across studies 2 and depends on the time since stroke onset, but also on the infarct territory (anterior or posterior circulation). 4 Besides, the best cutoff for patient triage may differ depending on the geographical setting. Indeed, in urban areas with several comprehensive stroke centers, a cutoff associated with a higher sensitivity (lower falsenegative rate) may be preferred to avoid increasing the time-tothrombectomy by inappropriately sending patients with LAO to a primary stroke center.…”

Cognitive Rehabilitation for Memory Deficits After Stroke

“…Важно отметить, что многие из этих шкал были разработаны и валидизированы в строго отобранной выборке пациентов, например шкала PASS (Prehospital Acute Stroke Severity Scale) [2], для валидизации которой включали только пациентов с ишемическим инсультом, которым проводили систем-ный тромболизис и которым выполнили визуализацию сосудов, или шкала CPSSS (Cincinnati Prehospital Stroke Severity Scale) [3], которая была разработана с исполь-зованием данных пациентов, включенных в испыта-ния NINDS и IMS-III (National Institute of Neurological Disorders and Stroke and Interventional Management of Stroke-III). Более того, в последующих исследованиях для одновременной ва лидизации этих шкал использо-вали ретроспективные выборки тщательно отобран-ных пациентов с ишемическим инсультом, госпитали-зированных в специализированные отделения [4][5][6]. Следовательно, пациентов с масками инсульта, малыми инсультами, лакунарными инсультами или геморраги-ческими инсультами в большинстве этих исследований, по всей вероятности, пропустили.…”

Occurrence of intracranial large vessel occlusion in consecutive, non-referred patients with acute ischemic stroke

“…However, these studies did not include unselected, consecutive patients [9], did not use appropriate neuro-imaging in all patients [10], included patients who were transferred from other hospitals [11,12], or used a restricted time window [13].…”