European Journal of Cardiovascular Medicine

Stroke is one of the leading causes of adult disability and mortality worldwide, with ischemic stroke accounting for the majority of cases. Early and accurate assessment of stroke severity plays a pivotal role in determining prognosis, guiding treatment decisions, and streamlining triage in emergency settings [1].

The National Institutes of Health Stroke Scale (NIHSS) is one of the most widely used tools for stroke assessment due to its standardized format and validation in both clinical and research settings [2]. Initially designed for use by neurologists, the NIHSS has since been extended to non-neurologist healthcare providers, demonstrating acceptable reliability and ease of application in acute care environments [2,3]. Despite its widespread use, the NIHSS has limitations—particularly in detecting less overt symptoms such as aphasia, cognitive deficits, or right-hemisphere syndromes [4].

Alternative tools, such as the Scandinavian Stroke Scale (SSS), offer simplified scoring structures and a more balanced assessment of motor, consciousness, and speech functions. Although less commonly used than the NIHSS, the SSS has shown promise in outcome prediction and may be more practical in certain emergency or non-specialist settings. It also allocates more explicit scoring for speech and eye movement, potentially increasing sensitivity to deficits like aphasia [8].

Aphasia, a common and often underrecognized consequence of ischemic stroke, has been linked to poor short- and long-term outcomes [5]. While comprehensive batteries such as the Western Aphasia Battery (WAB-R) [6] and the Comprehensive Aphasia Test (CAT) [7] exist for in-depth language assessment, their use is impractical in emergency settings. Therefore, the choice of an initial severity scale that adequately captures language impairment is crucial. Screening tools like the Language Screening Test (LAST) have been developed for this purpose [8], but are often used alongside broader severity scales, not in isolation.

Given the evolving epidemiology of stroke and the increasing emphasis on time-sensitive interventions, there is a need to compare practical, bedside scales that can be administered at the point of care by frontline physicians [8]. While the NIHSS continues to serve as the gold standard in many institutions, real-world data on its comparative accuracy with alternatives like the SSS in emergency department (ED) triage settings remains limited.

Aims and Objectives

Aim

To evaluate and compare the diagnostic accuracy of the National Institutes of Health Stroke Scale (NIHSS) and the Scandinavian Stroke Scale (SSS) in assessing early stroke severity at the point of entry into the emergency medicine department.

Objectives

1. To assess the distribution of stroke severity using NIHSS and SSS among adult patients presenting with suspected ischemic stroke.
2. To determine the level of agreement between the NIHSS and SSS in classifying stroke severity (mild, moderate, severe).
3. To compare the diagnostic performance of NIHSS and SSS against CT brain imaging as the reference standard, in terms of:
• Sensitivity
• Specificity
• Positive Predictive Value (PPV)
4. To evaluate the potential of each scale as a practical bedside tool for early stroke triage in emergency medicine settings.

This cross-sectional analytical study was conducted at the Emergency Medicine Department of Bharati Hospital and Research Centre, Pune, over a period of one year, from January 2024 to January 2025. The study aimed to evaluate and compare the diagnostic accuracy of the National Institutes of Health Stroke Scale (NIHSS) and the Scandinavian Stroke Scale (SSS) in assessing stroke severity at the time of patient presentation.

All adult patients aged 18 years and above presenting to the Emergency Medicine Department with symptoms suggestive of acute ischemic stroke were considered for inclusion. Patients were excluded if they had already been diagnosed or treated for stroke at another facility, were intubated prior to arrival, or were sedated or paralyzed, thereby limiting accurate neurological assessment.

Eligible patients were evaluated immediately upon arrival using the departmental stroke assessment protocol, which included initial vital signs (blood pressure, pulse, respiratory rate, SpO₂, body temperature), neurological status (Glasgow Coma Scale), blood sugar measurement, and electrocardiography. Stroke severity was assessed using both the NIHSS and SSS scoring systems. Each patient was evaluated by two trained and certified emergency medicine residents independently, both of whom were specifically trained in stroke scale application to minimize inter-observer variability.

Non-contrast computed tomography (CT) brain imaging was performed for all patients as per standard emergency stroke protocol. CT reports were reviewed and categorized into mild, moderate, or severe stroke based on radiological findings including infarct size, location, presence of haemorrhagic transformation, mass effect, or midline shift.

All clinical and radiological data were entered prospectively into a structured Excel spreadsheet and later analyzed using Python and statistical tools for descriptive analysis, confusion matrices, and diagnostic performance metrics. Stroke severity scores were categorized into mild, moderate, or severe based on established thresholds for both NIHSS and SSS. CT imaging findings were used as the reference standard to calculate sensitivity, specificity, and positive predictive value (PPV) for each scale.

The study was approved by the Institutional Ethics Committee prior to initiation. Informed consent was obtained from all patients or their legally authorized representatives prior to participation.

Baseline Characteristics of the Study Population

A total of 306 patients were included in the study. The mean age was 59.7 years (SD ± 14.9), with a range from 29 to 91 years. The mean Glasgow Coma Scale (GCS) on presentation was 13.6 (SD ± 3.1).

Vital parameters at the time of emergency department arrival included:

• Mean systolic blood pressure: 162.9 mmHg (SD ± 30.7)
• Mean diastolic blood pressure: 95.4 mmHg (SD ± 20.1)
• Mean pulse rate: 84.4 bpm (SD ± 13.6)
• Mean random blood sugar: 160.1 mg/dL (SD ± 59.4)

Of the 306 patients initially included, 221 patients (145 males and 76 females) had valid demographic data for analysis. Among them:

• Hypertension was present in 59.3% of patients (n = 131).
• Diabetes mellitus was documented in 25.6% (n = 56).
• Smoking history was noted in 17.6% (n = 39), and alcohol use in 25.8% (n = 57).
• Tobacco chewing was reported by 20.8% (n = 46).
• Dyslipidaemia was relatively uncommon, recorded in only 3.2% (n = 7).

These findings reflect a population with high cardiovascular risk factors, aligning with known ischemic stroke profiles.

Table1a: Numerical Baseline Characteristics

Table1b: Categorical Baseline Characteristics

2. Stroke Severity Score Overview

All 306 patients underwent stroke severity assessment at the time of presentation using both the National Institutes of Health Stroke Scale (NIHSS) and the Scandinavian Stroke Scale (SSS).

The mean NIHSS score was 8.6 ± 7.0, with a minimum of 0 and a maximum of 42. Based on NIHSS classification:

• 121 patients (39.5%) had mild stroke (≤4),
• 124 (40.5%) had moderate stroke (5–15),
• and 61 (19.9%) were classified as having severe stroke (≥16).

The mean SSS score was 40.1 ± 15.4, with values ranging from 0 to 58. Based on SSS severity cut-offs:

• 117 patients (38.2%) were categorized as mild (≥50),
• 125 (40.8%) as moderate (25–49),
• and 64 (20.9%) as severe (<25).

The distribution of stroke severity categories across the two scoring systems was broadly similar, though small differences were noted in the classification of borderline cases.

Figure1: Stroke Severity Category Distribution: NIHSS vs SSS

Figure 1. Distribution of stroke severity categories classified by the NIHSS and Scandinavian Stroke Scale (SSS). The bar chart compares the number of patients categorized as mild, moderate, or severe by each scoring system.

3. Agreement Between NIHSS and SSS

To assess the level of agreement between the NIHSS and Scandinavian Stroke Scale (SSS) in classifying stroke severity, both scores were categorized into mild, moderate, and severe categories using validated thresholds. A cross-tabulation was performed to evaluate classification concordance between the two scales.

While a high degree of overall agreement was observed, certain discrepancies emerged in the classification of borderline moderate-to-severe cases. The detailed distribution is presented in the confusion matrix below.

Table2: NIHSS vs SSS Confusion Matrix

The level of agreement between NIHSS and SSS severity classifications was assessed using Cohen’s Kappa coefficient. The analysis yielded a Kappa value of 0.567, which indicates a moderate level of agreement beyond chance.

This interpretation aligns with the Landis and Koch classification, where a Kappa value between 0.41 and 0.60 is considered moderate agreement.

Among the 306 patients assessed, classification agreement between the NIHSS and Scandinavian Stroke Scale (SSS) was evaluated using a three-level severity categorization (mild, moderate, severe). The highest concordance was observed in the mild and severe categories, with 89 patients consistently labelled as mild and 48 as severe by both scales. However, discrepancies were noted particularly in the moderate group, where classification varied between the two scoring systems.

The level of agreement between the two scales was further quantified using Cohen’s Kappa, which yielded a value of 0.567, indicating moderate agreement beyond chance.

4. Diagnostic Accuracy Compared to CT Imaging

To evaluate the diagnostic performance of the NIHSS and Scandinavian Stroke Scale (SSS) in assessing early stroke severity, both scoring systems were compared against radiological findings from non-contrast CT brain imaging, considered the reference standard in acute stroke evaluation.

CT brain impressions were reviewed and categorized into three levels of severity—mild, moderate, and severe—based on radiologist-reported findings such as infarct size, presence of haemorrhage, midline shift, or mass effect.

For diagnostic accuracy analysis, severity levels were further dichotomized into non-severe (mild) and clinically significant (moderate to severe) categories. Sensitivity, specificity, and positive predictive value (PPV) were calculated for each scale using CT imaging as the gold standard.

Table3a: NIHSS vs CT Imaging Confusion Matrix

Table3a: NIHSS vs CT Imaging shows how many patients were classified as having clinically significant stroke (1) or not (0) by NIHSS compared to CT results.

Table3b: SSS vs CT Imaging Confusion Matrix

Table3b: SSS vs CT Imaging shows how many patients were classified as having clinically significant stroke (1) or not (0) by SSS compared to CT results.

Table4:  Diagnostic Accuracy of NIHSS and SSS Compared to CT Brain Imaging

Note: CT imaging was used as the reference standard to define stroke severity. “Positive” indicates moderate-to-severe stroke on CT; “Negative” indicates mild stroke. Sensitivity refers to the scale’s ability to correctly identify patients with clinically significant stroke, while specificity refers to its ability to correctly identify those without. PPV indicates the probability that patients classified as positive truly had moderate/severe findings on imaging.

To assess the diagnostic accuracy of NIHSS and SSS, CT imaging findings were used as the reference standard. Among patients with interpretable CT results, NIHSS demonstrated a sensitivity of 37.5% and a specificity of 80.0%, with a positive predictive value (PPV) of 75.0%.

In comparison, the Scandinavian Stroke Scale (SSS) showed superior sensitivity at 62.5%, the same specificity (80.0%), and a higher PPV (83.3%).

These findings suggest that while both tools are specific, the SSS may be more sensitive in detecting moderate-to-severe stroke severity as identified by early CT brain imaging.

5. Summary of Diagnostic Performance

To facilitate direct comparison between the two scoring systems, a summary table was compiled to present the key diagnostic performance metrics of the NIHSS and Scandinavian Stroke Scale (SSS), using CT imaging as the reference standard.

This summary highlights that while both tools demonstrate similar specificity (80.0%), the SSS outperformed NIHSS in sensitivity (62.5% vs. 37.5%) and in positive predictive value (83.3% vs. 75.0%).

Table5:  Summary of Diagnostic Performance

Table 5 shows Summary of sensitivity, specificity, and positive predictive value (PPV) for NIHSS and Scandinavian Stroke Scale (SSS), using CT imaging as the reference standard.

In summary, both the NIHSS and Scandinavian Stroke Scale demonstrated comparable ability to classify stroke severity in patients presenting to the emergency department. While agreement between the two tools was moderate, the SSS showed higher sensitivity and positive predictive value when benchmarked against CT imaging findings. These results suggest that the SSS may offer a more sensitive point-of-care assessment tool for early detection of clinically significant ischemic stroke in acute emergency settings.

This study evaluated and compared the diagnostic accuracy of the National Institutes of Health Stroke Scale (NIHSS) and the Scandinavian Stroke Scale (SSS) in assessing early stroke severity in patients presenting to the Emergency Medicine Department. Both scales are widely used tools designed for rapid clinical evaluation; however, their comparative effectiveness in acute emergency settings continues to be a subject of investigation [10].

Patient Profile and Stroke Severity Trends

The baseline characteristics of the study population reflect a typical ischemic stroke demographic, with a mean age of 59.7 years and high prevalence of vascular risk factors including hypertension (59.3%) and diabetes mellitus (25.3%). These findings align with established stroke epidemiology and reinforce the importance of early risk stratification in emergency settings [9].

The distribution of stroke severity across both scales was relatively balanced, with NIHSS classifying 39.5% of patients as mild, 40.5% moderate, and 19.9% severe. The SSS showed a similar pattern (38.2% mild, 40.8% moderate, 20.9% severe), underscoring both tools’ utility in identifying varying levels of stroke severity at presentation.

Agreement Between NIHSS and SSS

The agreement between the two scales was moderate, with a Cohen’s Kappa of 0.567, indicating consistency in severity classification but also highlighting discrepancies, particularly in moderate stroke cases. The highest agreement was observed in the mild and severe categories, which is clinically significant given that these groups often guide treatment urgency and imaging prioritization.

These findings are in line with prior research showing that both tools have acceptable inter-rater reliability, but may diverge in cases where neurological deficits are borderline or more complex to interpret [10,12]. Askim et al. [12] previously reported comparable predictive value for 3-month outcomes between the two scales, reinforcing the clinical relevance of both tools despite moderate classification differences.

Diagnostic Accuracy Compared to CT Imaging

Using CT brain imaging as the reference standard, the SSS demonstrated superior sensitivity (62.5%) compared to the NIHSS (37.5%), while both showed equal specificity (80.0%). Additionally, the positive predictive value (PPV) was higher for SSS (83.3%) than NIHSS (75.0%). These findings suggest that the SSS may be more reliable in identifying patients with moderate-to-severe radiological involvement, potentially making it a more sensitive bedside tool in emergency stroke triage.

Similar observations were made in comparative studies evaluating prognostic accuracy, where the CSS (an earlier version of SSS) demonstrated strong predictive capacity for infarct extent and outcomes [13]. Furthermore, Pedersen et al. emphasized that stroke severity correlates closely with lesion size and type on imaging, supporting the validity of using CT as a reference for scale evaluation [9].

NIHSS Limitations in Detecting Aphasia and Specific Deficits

One explanation for the lower sensitivity of NIHSS in this study may be its limited sensitivity to certain deficits, such as aphasia, especially in early or subtle presentations. Studies have noted that NIHSS may underdiagnose language impairments or misclassify patients with non-dominant hemisphere strokes [11,14,15]. Grönberg et al. [11] specifically highlighted NIHSS’s limitations in detecting aphasia with high precision, a finding echoed in systematic reviews [15,16].

Given the clinical and psychosocial impact of aphasia on recovery and quality of life [17], early recognition is vital. This limitation may partially explain why SSS, which distributes scoring more evenly across motor, speech, and consciousness parameters, performed better in capturing moderate-severe radiologic involvement.

Clinical Implications

In busy emergency departments, where time-to-treatment is critical, the availability of a sensitive and simple severity scale can significantly enhance early triage decisions. While NIHSS remains widely used, this study supports the SSS as a practical alternative or complement, especially when neurological findings are subtle or CT access is delayed.

Given its simplicity and rapid administration, SSS may be particularly useful in non-specialist or resource-limited settings, without compromising diagnostic value. These insights can inform both clinical practice and training protocols in emergency stroke care.

Strengths and Limitations

A key strength of this study is the direct comparison of two validated stroke scales against radiological findings in real-world emergency presentations. The structured data collection by trained emergency medicine residents adds to the reliability of the assessments.

However, the diagnostic accuracy analysis was limited to a smaller subset (n = 13) with clearly interpretable CT severity classifications, due to variability in the CT report language. Additionally, long-term functional outcomes (e.g., mRS at 3 months) were not assessed and should be included in future studies for outcome correlation.

This cross-sectional analytical study compared the diagnostic accuracy of the National Institutes of Health Stroke Scale (NIHSS) and the Scandinavian Stroke Scale (SSS) in assessing early stroke severity at the point of entry into the emergency medicine department.

Both tools demonstrated comparable classification patterns; however, the SSS exhibited higher sensitivity and positive predictive value when benchmarked against CT imaging, the radiological gold standard. While NIHSS remains a well-established and widely used scale, these findings suggest that the SSS may serve as a more sensitive tool for identifying clinically significant ischemic strokes—particularly in time-sensitive or resource-constrained environments.

Given its ease of use, balanced scoring across neurologic domains, and stronger correlation with imaging-defined stroke severity, the Scandinavian Stroke Scale may be considered a practical and reliable alternative to NIHSS for early stroke assessment in emergency settings.

1. Aked J, Delavaran H, Norrving B, Lindgren A. Temporal trends of stroke epidemiology in Southern Sweden: a population‐based study on stroke incidence and early case‐fatality. 2018;50(3-4):174‐182. doi:10.1159/000489938.
2. Goldstein LB, Samsa GP. Reliability of the National Institutes of Health Stroke Scale. Extension to non‐neurologists in the context of a clinical trial. 1997;28(2):307‐310. doi:10.1161/01.STR.28.2.307.
3. Zandieh A, Kahaki ZZ, Sadeghian H, et al. The underlying factor structure of National Institutes of Health Stroke scale: an exploratory factor analysis. Int J Neurosci. 2012;122(3):140‐144. doi:10.3109/00207454.2011.631658.
4. El Hachioui H, Visch‐Brink EG, de Lau LM, et al. Screening tests for aphasia in patients with stroke: a systematic review. J Neurol. 2017;264(2):211‐220. doi:10.1007/s00415-016-8314-3.
5. Berthier ML. Poststroke aphasia: epidemiology, pathophysiology and treatment. Drugs Aging. 2005;22(2):163‐182. doi:10.2165/00002512-200522020-00006.
6. Kertesz A, Raven JC. WAB‐R: Western Aphasia Battery‐Revised. Texas, USA: PsychCorp; 2007.
7. Swinburn K, Porter G, Howard D. Comprehensive Aphasia Test. New York, NY: Psychology Press; 2004.
8. Flamand‐Roze C, Falissard B, Roze E, et al. Validation of a new language screening tool for patients with acute stroke: the Language Screening Test (LAST). 2011;42(5):1224‐1229. doi:10.1161/STROKEAHA.110.606103.
9. Pedersen PM, Vinter K, Olsen TS. Aphasia after stroke: type, severity and prognosis. The Copenhagen aphasia study. Cerebrovasc Dis. 2004;17:35‐43. doi:10.1159/000073896.
10. Budinčević H, Meštrović A, Demarin V. Stroke Scales as Assessment Tools in Emergency Settings: A Narrative Review. Medicina (Kaunas). 2022;58(11):1541. doi:10.3390/medicina58111541.
11. Grönberg A, Henriksson I, Lindgren A. Accuracy of NIH Stroke Scale for diagnosing aphasia. Acta Neurol Scand. 2021;143(4):375-382. doi:10.1111/ane.13388.
12. Askim T, Bernhardt J, Churilov L, Indredavik B. The Scandinavian Stroke Scale is equally as good as the National Institutes of Health Stroke Scale in identifying 3-month outcome. J Rehabil Med. (Exact journal name not specified in source, assumed based on context).
13. Zhao XJ, Li QX, Liu TJ, Wang DL, An YC, Zhang J, et al. Predictive values of CSS and NIHSS in the prognosis of patients with acute cerebral infarction: A comparative analysis. Medicine (Baltimore). 2018;97(39):e12419. doi:10.1097/MD.0000000000012419.
14. Engelter ST, Gostynski M, Papa S, et al. Epidemiology of aphasia attributable to first ischemic stroke: incidence, severity, fluency, etiology, and thrombolysis. 2006;37(6):1379‐1384. doi:10.1161/01.STR.0000221815.64093.8c.
15. Flowers HL, Skoretz SA, Silver FL, et al. Poststroke aphasia frequency, recovery, and outcomes: a systematic review and meta‐analysis. Arch Phys Med Rehabil. 2016;97(12):2188‐2201. doi:10.1016/j.apmr.2016.03.006.
16. Hilari K. The impact of stroke: are people with aphasia different to those without? Disabil Rehabil. 2011;33(3):211‐218. doi:10.3109/09638288.2010.488711.
17. Nesi M, Lucente G, Nencini P, Fancellu L, Inzitari D. Aphasia predicts unfavorable outcome in mild ischemic stroke patients and prompts thrombolytic treatment. J Stroke Cerebrovasc Dis. 2014;23(2):204‐208. doi:10.1016/j.jstrokecerebrovasdis.2013.03.026.