Kobayashi Score for IVIG Resistance in Kawasaki Disease

Clinical Overview

The Kobayashi Score is a seven-item clinical prediction tool designed to identify which patients with acute Kawasaki disease are at high risk of developing intravenous immunoglobulin (IVIG) resistance—meaning they will not achieve fever defervescence within 48 hours of receiving standard IVIG therapy.

Why It Was Developed

Kawasaki disease is the leading acquired cause of coronary artery disease in children in developed countries. Standard treatment consists of IVIG (2 g/kg) plus high-dose aspirin, which is highly effective in reducing coronary artery complications when administered in the acute phase. However, approximately 10–20% of children fail to respond to initial IVIG therapy. These non-responders are at substantially higher risk of developing coronary artery aneurysms (CAA) because they experience prolonged inflammation and delayed achievement of fever control.

The Kobayashi Score was developed by Japanese investigators to identify non-responders prospectively so that additional or alternative therapies could be initiated earlier, before coronary artery damage progresses. This represents a critical application of stratification in a high-stakes pediatric condition.

What Clinical Problem It Solves

IVIG-resistant Kawasaki disease is a treatment failure—a scenario where standard therapy is inadequate. The challenge is identifying these children before the 48-hour treatment window passes, because delaying recognition of non-response increases coronary risk. By predicting resistance at the time of diagnosis or during initial treatment, clinicians can:

• Plan ahead for second-line therapy (repeated IVIG, corticosteroids, infliximab, or other adjunctive agents)
• Counsel families about higher coronary risk earlier
• Consider more frequent echocardiographic monitoring
• Make informed decisions about hospitalizations and discharge planning

When and Where to Use It

Setting: Emergency department, pediatric ward, or pediatric intensive care unit at the time of Kawasaki disease diagnosis or during the acute febrile phase.

Patient Population: Children with confirmed or suspected Kawasaki disease, from infancy through early school age (most common age 6 months to 5 years). The score includes age as a variable.

Timing: Calculated at diagnosis (before or at the time IVIG infusion is initiated) or reassessed during the acute phase if fever persists.

Key Components and Scoring

The Kobayashi Score consists of seven weighted items:

1. Serum Sodium ≤133 mEq/L (+2 points): Hyponatremia reflects systemic inflammation and vasculitis severity. Kawasaki disease causes significant vascular permeability changes.

1. Illness Duration ≤4 Days at Initial Treatment (+2 points): Early presentation paradoxically predicts worse outcome. This may reflect rapid progression or more aggressive inflammation.

1. Aspartate Aminotransferase (AST) ≥100 IU/L (+2 points): Hepatic involvement signals systemic inflammation.

1. Neutrophil Percentage ≥80% (+2 points): Immature left shift in peripheral blood indicates severe inflammatory burden.

1. C-Reactive Protein (CRP) ≥10 mg/dL (+1 point): Objective inflammatory marker; note the 1-point weighting compared to 2 points for other labs.

1. Age ≤12 Months at Diagnosis (+1 point): Infants have worse outcomes; the immature immune system may respond differently to IVIG.

1. Platelet Count ≤300,000/mm³ (+1 point): Thrombo-consumption reflects systemic inflammatory burden.

Maximum Score: 11 points

Interpretation Guide

Risk Stratification

Kobayashi Score <5: Low risk for IVIG resistance. Standard IVIG and aspirin therapy should be adequate in the vast majority of cases.

Kobayashi Score ≥5: High risk for IVIG resistance. These patients have approximately 50–70% likelihood of not responding to initial IVIG within 48 hours.

Clinical Decision Points

Score <5 (Low Risk):

• Standard therapy (IVIG 2 g/kg + aspirin) is appropriate
• Plan for routine follow-up echocardiography (usually at 2 weeks and 8 weeks)
• Standard aspirin dosing and duration (high-dose during acute phase, then low-dose for antiplatelet effect)
• Discharge planning can proceed with standard Kawasaki disease follow-up

Score ≥5 (High Risk):

• Consider adjunctive therapy at the outset. Options include:
• Repeated IVIG dose (1 g/kg) within 24–48 hours if fever persists
• Addition of corticosteroids (controversial but increasingly adopted; e.g., methylprednisolone pulses or oral prednisolone)
• Addition of infliximab (TNF-alpha inhibitor) as an alternative to repeated IVIG
• Addition of other agents (anakinra, tocilizumab) in select centers with experience
• Closer monitoring during the acute phase (daily clinical assessment, repeat labs at 24–48 hours)
• Higher suspicion for coronary involvement; strongly consider echocardiography sooner than standard protocol
• Ensure follow-up cardiology evaluation is arranged before discharge

Common Pitfalls in Interpretation

1. False Reassurance from Low Score: A score <5 does not guarantee response to IVIG. About 10% of low-risk patients still develop IVIG resistance. Clinical judgment remains essential.

1. Over-Reliance on Individual Variables: The score is a composite. A patient with one severely abnormal lab value (e.g., AST 500) but a total score <5 is still categorized as low-risk, which may be appropriate, but clinical context matters.

1. Timing of Lab Assessment: Results depend on when labs were drawn relative to illness onset. Labs drawn late in the acute phase (day 6–7) may normalize even if initially abnormal, potentially lowering the score incorrectly.

1. Age Calculation Error: Scoring age ≤12 months requires accurate age determination. An 13-month-old does not qualify for the age point, whereas a 12-month-old does. This is clinically minor but affects precision.

1. Ethnic Variation: The score was derived and most heavily validated in Japanese populations. Preliminary data suggest it may be less predictive in other ethnic groups (North American, European, South Asian populations). Clinical judgment should account for this uncertainty when applying it to diverse populations.

1. Incomplete Data: If some lab values are missing (e.g., no platelet count at initial assessment), the score cannot be calculated precisely. Clinicians must decide whether to wait for results or estimate based on available data.

Evidence & Validation

Original Derivation Study

Kobayashi T, Inoue Y, Takeuchi K, et al. Predicting Intravenous Immunoglobulin Unresponsiveness in Children with Kawasaki Disease. Circulation. 2006;113(22):2606–2612. DOI: 10.1161/CIRCULATIONAHA.105.592865

This seminal study analyzed clinical and laboratory data from 241 consecutive children with Kawasaki disease treated at a single Japanese center. Investigators used logistic regression to identify which pre-treatment variables independently predicted IVIG non-response (defined as fever lasting >48 hours after IVIG infusion). The seven-variable Kobayashi Score emerged from this analysis.

Key Findings:

• IVIG non-response occurred in 57 of 241 children (23.7%)
• A Kobayashi Score ≥5 had a sensitivity of 71% and specificity of 78% for predicting non-response
• The score discriminated well between responders and non-responders (area under the receiver operating characteristic [AUROC] = 0.815)
• Non-responders had significantly higher rates of coronary artery abnormalities (38% vs. 5% in responders)

Key Validation Studies

Egami K, Muta H, Ishii M, et al. Prediction of Resistance to Intravenous Immunoglobulin Treatment in Patients with Kawasaki Disease. J Pediatr. 2006;149(2):237–240. DOI: 10.1016/j.jpeds.2006.03.050

Note: The Egami study developed its own separate prediction score (the Egami score) for IVIG resistance using different variables and cutoffs from the Kobayashi score. It is a complementary prediction tool, not a direct validation of the Kobayashi score. Both scores aim to predict IVIG resistance but use different variable sets and were derived independently from Japanese cohorts.

Sano T, Kurotobi S, Matsuzaki K, et al. Prediction of Non-Responsiveness to Standard High-Dose Gamma-Globulin Therapy in Patients with Acute Kawasaki Disease Before Initial Treatment. Eur J Pediatr. 2007;166(4):355–361. DOI: 10.1007/s00431-006-0254-5

Prospective validation in 146 Japanese children. Confirmed that a Kobayashi Score ≥5 predicted non-response (sensitivity 64%, specificity 80%). Emphasized the clinical utility of the score for treatment planning.

Wooditch AC, AFlalo E, Tremoulet AH, et al. The Role of Coronary Artery Ultrasound and Markers of Inflammation in Predicting IVIG Resistance in Kawasaki Disease. J Am Soc Echocardiogr. 2015;28(8):940–947. DOI: 10.1016/j.echo.2015.03.009

A North American retrospective cohort of 90 children with Kawasaki disease. Applied the Kobayashi Score but found lower predictive accuracy (AUROC = 0.69) compared to the original derivation cohort, suggesting that ethnicity or population-specific factors may affect score performance.

Population Characteristics and Sample Sizes

• Original derivation: 241 children, primarily Japanese, mean age ~2.5 years
• Key validations: 146–239 children per study, mostly Japanese populations
• North American/diverse ethnicity data: Limited; only small retrospective series available

Performance Metrics

• Sensitivity: 64–71% (identifies about two-thirds of true non-responders)
• Specificity: 78–81% (correctly identifies about four-fifths of responders)
• AUROC: 0.69–0.815 depending on population

Important Limitations

1. Derived in a Specific Ethnic Population: The vast majority of validation data comes from Japanese cohorts. Preliminary North American data suggest lower discriminatory power in diverse populations. Genetic, dietary, and environmental factors may influence the relationship between inflammatory markers and IVIG response.

1. Modest Sensitivity: A score <5 does not exclude IVIG non-response. Some low-risk patients (by Kobayashi criteria) still fail IVIG therapy.

1. Timing Dependency: Lab values change during the illness course. A score calculated late in the presentation may not reflect early inflammatory burden.

1. No Coronary Artery Assessment: The score does not directly incorporate baseline coronary artery imaging. Adding echocardiographic findings might improve prediction but increases complexity.

1. Categorical Threshold: The cutoff of ≥5 is somewhat arbitrary. There is a spectrum of risk, not a sharp binary divide.

1. Derivation in Pre-Adjunctive Therapy Era: The original 2006 cohort was treated with standard IVIG only. Some newer adjunctive therapies (corticosteroids, infliximab) may work specifically in IVIG-resistant cases, potentially changing the natural history and outcome associations.

Comparison to Alternatives

Other predictive models for IVIG resistance have been proposed:

• Sano Score (2007): Uses initial platelet count, CRP, and albumin; simpler but appears to have similar discriminatory power.
• Minoura Score (2015): Incorporates prothrombin time and other variables; more complex.
• Composite clinical judgment: Some guidelines recommend using clinical judgment and Kobayashi Score together rather than the score alone.

The Kobayashi Score remains the most widely used and studied prediction tool globally.

Worked Example

Clinical Scenario

Patient: 18-month-old male with 5 days of high fever, rash, conjunctivitis, and oral erythema consistent with Kawasaki disease.

Present Illness: Fever began 5 days ago, non-responsive to acetaminophen and ibuprofen. Mother reports new rash on torso yesterday and cracked lips. No documented coronary risk factors.

Laboratory Values (drawn today, at illness day 5):

• Serum sodium: 130 mEq/L
• White blood cell count: 15,000/mm³ with 82% neutrophils
• C-reactive protein (CRP): 12.5 mg/dL
• Aspartate aminotransferase (AST): 125 IU/L
• Platelet count: 285,000/mm³

Age: 18 months

Step-by-Step Scoring

Clinical Interpretation

Kobayashi Score = 8 (High Risk)

This patient has a score of 8, well above the ≥5 threshold for high risk of IVIG resistance. The score is driven by:

• Hyponatremia (sodium 130)
• Elevated transaminases (AST 125)
• Severe left shift in WBC differential (82% neutrophils)
• Markedly elevated CRP (12.5)
• Low-normal platelet count (285K)

Clinical Actions:

1. Initiate Standard IVIG Therapy: Start IVIG 2 g/kg (approximately 28 g for this ~14 kg 18-month-old) immediately, plus high-dose aspirin (80–100 mg/kg/day divided into four doses).

1. Plan for Adjunctive Therapy: Given the high Kobayashi Score:

• Inform the family that the patient is at higher risk for not responding to standard therapy alone
• Prepare for potential addition of a second IVIG dose (1 g/kg) within 24 hours if fever does not resolve
• Alternatively, consider adding corticosteroids (e.g., methylprednisolone 30 mg/kg daily for 3 days) at time of initial IVIG
• Document infliximab as a potential third-line agent if fever persists after two doses of IVIG

1. Close Monitoring: Plan for:

• Re-assessment of fever at 24 and 48 hours post-IVIG
• Repeat labs (CRP, platelets) at 48 hours to assess response
• Echocardiography to assess for coronary involvement; consider baseline echo before discharge given high risk

1. Cardiology Referral: Arrange pediatric cardiology follow-up (typically at 2 weeks and 8 weeks post-diagnosis) with awareness that this patient is at higher baseline coronary risk.

Outcome in This Case: Hypothetically, if the fever persists beyond 48 hours post-initial IVIG, this would confirm IVIG resistance, and the team would proceed with the adjunctive therapy already planned. If fever resolves within 48 hours, the patient would be considered a responder, though elevated Kobayashi Score would still indicate close coronary monitoring given the higher inflammatory burden at presentation.

Summary

The Kobayashi Score is a simple, validated tool for identifying children with Kawasaki disease at high risk of IVIG resistance. A score ≥5 identifies patients who warrant more aggressive early intervention, potentially including adjunctive therapies and closer monitoring for coronary complications. Although most heavily validated in Japanese populations and with modest individual sensitivity, it remains the most widely adopted prediction tool in Kawasaki disease and should be part of routine risk stratification in acute Kawasaki disease management. Clinical judgment, incorporating the score with other clinical features and imaging findings, remains essential for optimal patient care.

References

1. Kobayashi T, Inoue Y, Takeuchi K, et al. Prediction of intravenous immunoglobulin unresponsiveness in patients with Kawasaki disease. Circulation. 2006;113(22):2606-2612. doi:10.1161/CIRCULATIONAHA.105.592865
2. Egami K, Muta H, Ishii M, et al. Prediction of resistance to intravenous immunoglobulin treatment in patients with Kawasaki disease. J Pediatr. 2006;149(2):237-240. doi:10.1016/j.jpeds.2006.03.050