360 - Performance of Scoring Systems to Predict Coronary Artery Aneurysms in Patients with Kawasaki Disease in Hawai‛i

Publication Number: 2349.360

Sung Hwan Choi, University of Hawaii, John A. Burns School of Medicine, Honolulu, HI, United States; Rodolfo Bégué, University of Hawaii, John A. Burns School of Medicine, Honolulu, HI, United States

Background: Identifying patients with Kawasaki disease (KD) who are at higher risk of developing coronary artery aneurysms (CAAs) remains challenging. Although multiple predictive models have been developed for both Asian and non-Asian populations to stratify risk for refractory KD and/or CAAs, these models have not demonstrated consistent performance across different ethnic groups. With its high incidence of KD and notable ethnic diversity, Hawai‛i provides a unique setting to validate existing predictive models and assess risk factors for CAAs.

Objective: To validate previously developed scoring systems (Kobayashi, Sano, Egami, and Son scores) in patients with KD in Hawai‛i and to identify predictive factors associated with CAAs.

Design/Methods: A total of 234 patients with KD diagnosed between January 1, 2016, and December 31, 2024 were included. Demographic, clinical, laboratory, and echocardiographic data were collected and analyzed. Sensitivity, specificity, positive and negative predictive values (PPV and NPV) were calculated for the four scoring systems. Statistical analyses were performed to identify predictive factors for CAAs at two time points: at KD diagnosis and after treatment through 6 weeks.

Results: Of the 234 patients, 95 (40.6%) were Asian, 88 (37.6%) were multiracial patients with Asian ancestry, 14 (6.0%) Pacific Islander, 1 (0.4%) Hispanic, and 8 (3.4%) White. In this population, a Kobayashi score ≥ 4 demonstrated a sensitivity of 53.4%, specificity of 63.9%, PPV of 34.1%, and NPV of 79.7%. For a Sano score ≥ 2, sensitivity was 17.2%, specificity 89.2%, PPV 35.7%, and NPV 75.5%. An Egami score ≥ 3 showed a sensitivity of 36.2%, specificity 72.2%, PPV 30.9%, and NPV 76.7%. A Son score of 3 to 5 had the highest sensitivity (88.2%) and NPV (97.0%), but low PPV (41.1%). Male sex was associated with increased risk of CAAs at diagnosis (odds ratio [OR] 2.76; 95% confidence interval [CI] 1.40-5.42; p = 0.003). Patients with CAAs at diagnosis were significantly more likely to have persistent CAAs after treatment (OR 18.91; 95% CI 8.33-42.93; p < 0.001).

Conclusion(s): Among patients with KD in Hawai‛i, the Son score demonstrated the highest sensitivity (88.2%) for predicting CAAs, but its low PPV (41.1%) limits its utility for identifying patients who may benefit from intensified initial therapy. Male sex was associated with a higher risk of CAAs at diagnosis, and the presence of CAAs at diagnosis strongly predicted post-treatment CAAs, consistent with previous studies.

Table 1.
Performance of existing predictive scoring systems in patients with Kawasaki disease in Hawai‛i.