A Simple 3-Item Stroke Scale
Comparison With the National Institutes of Health Stroke Scale and Prediction of Middle Cerebral Artery Occlusion
Background and Purpose— The purpose of the study was to design a simple stroke scale that requires minimal training but reflects initial stroke severity and is predictive of middle cerebral artery (MCA) occlusion.
Methods— The new stroke scale assessed 3 parameters: (1) level of consciousness, (2) gaze, and (3) motor function. Each item was graded 0 to 2, where 0 indicated normal findings and 2 severe abnormalities (ie, profound drowsiness or worse, forced gaze deviation, and severe hemiparesis, respectively). During a study period of 11 months, patients presenting with acute stroke symptoms (onset ≤6 hours) were examined by a stroke neurologist assessing the new scale as well as the National Institutes of Health Stroke Scale (NIHSS). In addition, 83 patients received acute magnetic resonance angiography (MRA; as part of an acute stroke protocol).
Results— The new stroke scale was strongly associated with the NIHSS. Interobserver reliability of the new scale was high (intraclass correlation coefficient 0.947). Using post hoc analysis, a score of ≥4 predicted proximal vessel occlusion (T-segment or M1-segment occlusion of the MCA on MRA) almost as accurately (overall accuracy 0.86) as an NIHSS score of ≥14 (overall accuracy 0.93).
Conclusions— The new stroke scale reflects acute stroke severity well and predicts proximal MCA occlusion with reasonable accuracy. However, the clinical scale needs further evaluation before it can be recommended as a tool for the triage of acute stroke patients.
Modern imaging techniques allow the rapid assessment of vessel status in acute ischemic stroke and can be used for the triage of patients for (invasive) recanalization strategies. However, computed tomography (CT) and magnetic resonance angiography (MRA) as well as transcranial Duplex ultrasound are not yet broadly available 24 hours per day. In addition, for some severely ill patients, MRI is not an ideal imaging modality because of limitations concerning patient monitoring.1 Particularly for hospitals with limited imaging resources, surrogate markers of vessel occlusion would be helpful.
Recently, a relatively strong relationship between the neurological deficit as measured by the National Institutes of Health Stroke Scale (NIHSS) and vessel status has been established.2,3 However, although the NIHSS has proven useful in the context of clinical trials, it remains a complex scoring system necessitating regular training for accurate application. Furthermore, there is evidence that the NIHSS is performed rarely in clinical routine.4 Recent studies investigated the utility of shorter stroke scales to assess stroke severity and to differentiate stroke from other conditions.5–9 However, their predictive value as to large vessel status has not been investigated yet. The purpose of this study was to design a simple scale that: (1) reflects acute stroke severity, and (2) may serve as a predictor of middle cerebral artery (MCA) occlusion.
Materials and Methods
The 3-Item Stroke Scale
Table 1 displays the scoring sheet of the new 3-item stroke scale (3I-SS), assessing the items: (1) level of consciousness, (2) gaze, and (3) motor function. Each item was graded 0 to 2, where 0 indicated normal, 1 mild, and 2 severe abnormalities. Consciousness was classified as being mildly disturbed if the patient was drowsy. If vigorous (painful) stimuli were necessary for arousal, or the patient was unarousable, consciousness was classified as being severely disturbed. Gaze deviation was classified as being incomplete if crossing of the midline was possible by visual or acoustic stimuli; otherwise it was classified as being fixed. Hemiparesis was classified as moderate if the patient was able to elevate both upper and lower extremities but with a drift in <10 seconds in either or both extremities of one side. Hemiparesis was rated as severe if no or only minimal movement against gravity was possible in at least one limb. The 3I-SS was developed and evaluated prospectively.
During a study period of 11 months from January through November 2002, all patients (n=180) presenting to our stroke unit with the clinical suspicion of an acute (symptom onset ≤6 hours cerebrovascular event (hemorrhagic or ischemic stroke) were prospectively enrolled in the study. Patients were examined by experienced stroke neurologists assessing the new 3I-SS as well as the NIHSS before initial imaging procedures. Furthermore, the final diagnosis at discharge was recorded. For assessment of the inter-rater reliability, the 3I-SS and the NIHSS were assessed by 2 neurologists in a subset of 20 patients.
Neuroradiological Imaging and MRA
During the study period, multimodal stroke MRI was the preferred imaging modality, being available 24 hours per day. Patients not eligible for MRI (n=58) because of contraindications to MRI, such as cardiac pacemaker, psychomotoric agitation, or only minimal neurological deficit (NIHSS≤2), received a conventional CT scan. Our standard MRI protocol included a time-of-flight (TOF) MRA with the following parameters: repetition time, 35 ms; echo time, 6.6 ms; flip angle, 20°; field of view, 142×190 mm; matrix, 200×512; slab thickness, 70 mm; number of partitions, 40; acquisition time, 3 minutes and 45 s; and maximum intensity projection reconstruction. MRA analysis was performed blinded to clinical data except for the side of clinically suspected ischemia.
For statistical analysis, SPSS 10.0 software (SPSS) was used. κ-statistics were used for assessment of inter-rater reliability of the single items of the new stroke scale, and an intraclass correlation coefficient was calculated for the total scores. In a post hoc analysis, Cross tables for different cut-off values were used to evaluate sensitivity and specificity as well as positive predictive value (PPV) and negative predictive value (NPV), and overall accuracy of different NIHSS and 3I-SS scores for the presence of MCA occlusion.
Figure 1 and Table 2 summarize basic patient characteristics. In total, n=171 patients were analyzed, including n=14 stroke mimics. Mean time between symptom onset and neurological examination was 132±78 minutes.
Inter-rater reliability of the new scale (3I-SS) was high; the intraclass correlation coefficient for the total score was 0.947 compared with 0.953 for the NIHSS. κ-values for inter-rater agreement for the single items were 0.77, 0.77, and 0.84 for disturbance of consciousness, gaze/head deviation, and hemiparesis, respectively.
A strong association between the 3I-SS and the NIHSS was observed when all patients (n=171, including 28 patients with intracerebral hemorrhage [ICH]) as well as when only patients receiving MRA (n=83) were analyzed. Supplemental Figure I (available online at http://www.strokeaha.org) gives an overview of the NIHSS range for each 3I-SS score.
The relationship between clinical scores and vessel status (as assessed on MRA) is shown in Table 3 and Figure 2. NIHSS and 3I SS scores were highest in patients with distal internal carotid artery (T) or proximal MCA (M1) occlusion, moderately elevated in patients with MCA branch occlusion and lowest in patients without vessel occlusion. All patients with a score ≥5 (n=8) had proximal MCA occlusion, in contrast to only 2 of 45 patients (4.4%) with a score ≤1. However, there were also substantial interindividual differences in stroke severity despite similar vessel occlusion sites on MRA (Table 3).
Furthermore, best-fitting cutoff values of the 3I-SS for predicting proximal vessel occlusion (T or M1 occlusion of the MCA) were evaluated (Table 4). Best overall accuracy was achieved for a score of ≥4, with a sensitivity and specificity of 0.67 and 0.92 for detecting proximal MCA occlusion. The PPV and NPV were 0.74 and 0.89, respectively. Concerning the NIHSS, best overall accuracy was achieved for a score of ≥14, with a sensitivity of 0.86, a specificity of 0.95, and a PPV and NPV of 0.86 and 0.95, respectively, for detecting vessel occlusion (supplemental Table I, available online at http://www.strokeaha.org). These cutoff values need to be confirmed in a separate prospectively enrolled patient population.
In this study, a simple stroke scale consisting of only 3 items (level of consciousness, gaze, and motor function) was found to be well associated with the NIHSS and to predict MCA occlusion.
The 3 items of the new stroke scale are selected items (and, in part, modified) of the NIHSS, which comprises 15 neurological examination test items. “Motor function” (ie, the degree of hemiparesis) in the new scale is an aggregate item of the NIHSS items “best motor arm” and “best motor leg” function. It was included in the new score as one of the most robust features of the NIHSS, which can be scored reliably and is a good indicator of stroke severity. The parameters “level of consciousness” and “gaze” were selected because we assumed that abnormal findings would indicate ischemia involving cortical tissue. This was done to separate patients with major vessel obstruction from patients with lacunar stroke.
With the new scale, the separation of mild (scores 0 to 1) from severe (scores ≥4) strokes is straightforward. Together, these 2 subgroups comprised >80% of our study population. This corresponds to very low and high rates of proximal vessel occlusion in these 2 groups, respectively. For patients with moderate scores (2 and 3) on the 3I-SS, the range of corresponding NIHSS scores and the findings on MRA were more variable. In part, the relatively wide range of NIHSS values for scores of 2 to 4 on the 3I-SS may also be attributable to peculiarities of the NIHSS, in which at least 4 items are speech dependent and may be influenced by aphasia.10,11 In contrast, the items tested by the 3I-SS are not influenced by aphasia.
Stroke severity as assessed by the 3I-SS and the NIHSS was well correlated with the presence of major vessel pathology. A direct correlation between the NIHSS score and findings on MRA was found in our study as it has been reported by others.3,12 The best cutoff value for the prediction of proximal vessel occlusion (T or M1 occlusion) was a score of ≥14, which yielded a high PPV (0.86) and NPV (0.95) resulting in a high overall accuracy (0.93), being in line with findings of Derex et al.2 For the 3I-SS, the highest overall accuracy (0.86) was achieved for a score ≥4, with a PPV of 0.74 and a NPV of 0.89 for the presence/absence of proximal vessel occlusion. However, these results need to be confirmed based on a separate data set.
For statistical analysis of the 3I-SS as a predictor of vessel occlusion, we dichotomized the study population into patients with and without proximal vessel occlusion. This was done partially for technical reasons because TOF MRA is most reliable for the detection of proximal vessel occlusion. On the other hand, patients with proximal vessel occlusion are possible candidates for aggressive recanalization procedures including intra-arterial thrombolysis.
The proposed 3I-SS has several limitations. First, the scale was not primarily designed to differentiate stroke from nonstroke or stroke mimics. Therefore, the 14 stroke mimics were not excluded from analysis. Second, it was not designed to serve as a screening tool for intravenous thrombolysis. Patients with a 3I-SS of 0 or 1 may well have a NIHSS score ≥4, which is commonly regarded as a lower limit for intravenous thrombolyis. Third, as a global scale with a limited range (0 to 6) of possible scores, the 3I-SS is insensitive to small differences between patients and to small changes in clinical status for individual patients. Because the aim of the study was to develop a screening tool for proximal MCA occlusion in the acute stroke setting, clinical outcome was not assessed systematically. Therefore, the predictive value of the 3I-SS concerning stroke outcome was not evaluated. Fourth, as a predictor of MCA occlusion, it is less accurate than the NIHSS, and therefore, it may be argued that it is of limited use. However, the complexity of the NIHSS is responsible for its infrequent use in clinical routine.4 In contrast, the 3I-SS is simple and quick to apply.
In summary, the strength of the new scale is its simplicity and the good overall accuracy for detection of proximal MCA occlusion, especially in patients with mild and severe symptoms, corresponding to scores of 0 to 1 or ≥4, respectively. However, even in these patients, the stroke scale should not replace thorough neuroimaging procedures before further confirmation of our results in an independent patient population. Additional studies concerning the utility of the 3I-SS in the paraclinical setting and its ability to predict stroke outcome are warranted.
This work was supported by a research grant of the Deutsche Forschungsgemeinschaft Ne 569/3-1 and by the J.W. Goethe University, Frankfurt.
- Received November 1, 2004.
- Revision received December 9, 2004.
- Accepted December 15, 2004.
Singer OC, Sitzer M, du Mesnil de Rochemont R, Neumann-Haefelin T. Practical limitations of acute stroke MRI due to patient-related problems. Neurology. 2004; 62: 1848–1849.
Lewandowski CA, Frankel M, Tomsick TA, Broderick J, Frey J, Clark W, Starkman S, Grotta J, Spilker J, Khoury J, Brott T. Combined intravenous and intra-arterial r-TPA versus intra-arterial therapy of acute ischemic stroke: Emergency Management of Stroke (EMS) Bridging Trial. Stroke. 1999; 30: 2598–2605.
Kidwell CS, Starkman S, Eckstein M, Weems K, Saver JL. Identifying stroke in the field. Prospective validation of the Los Angeles prehospital stroke screen (LAPSS). Stroke. 2000; 31: 71–76.
Tirschwell DL, Longstreth WT Jr, Becker KJ, Gammans RESr, Sabounjian LA, Hamilton S, Morgenstern LB. Shortening the NIH Stroke Scale for use in the prehospital setting. Stroke. 2002; 33: 2801–2806.
Nor AM, McAllister C, Louw SJ, Dyker AG, Davis M, Jenkinson D, Ford GA. Agreement between ambulance paramedic- and physician-recorded neurological signs with Face Arm Speech Test (FAST) in acute stroke patients. Stroke. 2004; 35: 1355–1359.
Fink JN, Selim MH, Kumar S, Silver B, Linfante I, Caplan LR, Schlaug G. Is the association of National Institutes of Health Stroke Scale scores and acute magnetic resonance imaging stroke volume equal for patients with right- and left-hemisphere ischemic stroke? Stroke. 2002; 33: 954–958.
Ernst R, Pancioli A, Tomsick T, Kissela B, Woo D, Kanter D, Jauch E, Carrozzella J, Spilker J, Broderick J. Combined intravenous and intra-arterial recombinant tissue plasminogen activator in acute ischemic stroke. Stroke. 2000; 31: 2552–2557.