Is the iScore Tool for Stroke Outcome Prediction Any Good?
Yes, the iScore is a validated and clinically useful tool for predicting stroke outcomes, including mortality, functional disability, and response to thrombolytic therapy, though it performs best in low-to-medium risk patients and has important limitations in high-risk populations. 1, 2
Evidence for iScore Effectiveness
Validated Prediction Capabilities
The iScore demonstrates strong predictive performance across multiple outcomes:
- Mortality and functional outcomes: The tool reliably predicts death at 30 days and poor functional outcomes (modified Rankin Scale 3-6) with high correlation between observed and predicted outcomes (correlation coefficients 0.940-0.993) 3
- Discrimination ability: The iScore shows good discrimination for poor outcomes at discharge (area under ROC curve 0.767-0.775) and at 3 months (0.801-0.810) 4
- External validation: Performance has been validated across multiple cohorts including the Registry of the Canadian Stroke Network (n=3,818) and Ontario Stroke Audit (n=4,635) 3
Predicting Response to Thrombolytic Therapy
The iScore's most clinically valuable application is stratifying which patients will benefit from tPA treatment:
- Low-risk patients (iScore <200): tPA therapy provides significant benefit with 47% higher odds of favorable outcome at 3 months (OR 1.47,95% CI 1.30-1.67) and improved rates of modified Rankin Scale 0-2 (47.5% vs 38.9% without tPA, p<0.001) 1, 2
- High-risk patients (iScore ≥200): No significant benefit from tPA (7.6% favorable outcome with tPA vs 5.5% without, p=0.45), with substantially higher hemorrhagic complication rates 1
Hemorrhagic Risk Stratification
The iScore effectively predicts bleeding complications after thrombolysis:
- Symptomatic ICH risk: Patients with iScore ≥200 have 15.4% symptomatic ICH rate with tPA versus 3.9% with placebo (p=0.04) 5
- Any ICH: Risk increases to 30.8% versus 11.5% in placebo (p=0.014) in high-risk patients 5
- ICH mortality: Dramatically higher at 69.2% versus 23.8% in high iScore patients 5
Important Limitations and Caveats
Context-Specific Use
The iScore is NOT designed for large vessel occlusion detection - other tools like NIHSS, RACE, LAMS, and CPSSS are more appropriate for this purpose, though none achieve both high sensitivity and specificity 6
Practical Implementation Issues
- Original version requires TOAST classification: The stroke subtype determination may require extensive investigations that delay risk assessment 4
- Revised version available: Substituting TOAST with the simpler OCSP classification maintains comparable discrimination (area under ROC 0.767 vs 0.775, p=0.06) and allows faster bedside application 4
Broader Context of Stroke Risk Tools
Current guidelines acknowledge significant gaps in stroke risk stratification:
- The American Heart Association/American Stroke Association notes that no risk stratification systems have been generally recommended for use after stroke in existing secondary prevention guidelines 6
- Even "low-risk" groups defined by various stroke prediction instruments have annual stroke recurrence rates of 3.2-4%, well above the 2% threshold for risk equivalents 7, 6
- Current risk stratification schemes ignore clinically important outcomes including functional decline, disability, and dementia 7, 6
Clinical Application Algorithm
For acute ischemic stroke patients being considered for tPA:
- Calculate iScore using available clinical variables (age, stroke severity, comorbidities, stroke subtype)
- If iScore <200: Proceed with tPA if otherwise eligible - expect significant benefit with acceptable hemorrhagic risk 1, 2
- If iScore ≥200: Exercise extreme caution with tPA - minimal benefit with substantially elevated hemorrhagic risk (15.4% symptomatic ICH) 5, 1
- Use revised iScore with OCSP classification if TOAST subtype unavailable for faster bedside assessment 4
For general outcome prediction: