Abstract
BACKGROUND: Intravenous immunoglobulin (IVIG) is the primary treatment for Kawasaki disease (KD). However, 10-20% of KD patients show no response to IVIG treatment, making the early prediction of IVIG resistance a key focus of KD research. Our aim is to explore the application of the C-reactive protein to albumin ratio (CAR) for predicting IVIG resistance in children with KD through meta-analysis. METHODS: Cochrane Library, PubMed, MEDLINE, EMbase, CNKI, WanFang, the Chinese Biomedical Database, and CQVIP were searched up to November 2023 for cohort studies on predicting IVIG-resistant KD using the CAR. Articles were selected based on pre-established inclusion and exclusion criteria after extracting literature data and assessing them using the QUADAS-2.0 tool for evaluating the accuracy of diagnostic tests. Stata 15.0 software was used for meta-analysis. RESULTS: Four Chinese and English literature reports were included in this meta-analysis. The results revealed the presence of a threshold effect and high heterogeneity among the included studies. The combined sensitivity for CAR predicting IVIG-resistant KD was calculated as 0.65 (95% CI 0.58-0.72), specificity as 0.71 (95% CI 0.57-0.81), and the area under the curve (AUC) as 0.70 (95% CI 0.66-0.74) using the random-effects model. The combined positive likelihood ratio was 2.22 (95% CI 1.35-3.65), the combined negative likelihood ratio was 0.49 (95% CI 0.35-0.69), and the diagnostic odds ratio was 5 (95% CI 2-10). CONCLUSION: CAR is an auxiliary predictive indicator with moderate diagnostic value that provides guidance in the early treatment of the disease, demonstrating a certain predictive value that warrants further investigation. However, CAR cannot yet be considered as a definitive diagnostic or exclusionary marker for IVIG-resistant KD. Therefore, multi-center, large sample, and high-quality long-term follow-up trials are warranted to confirm the current findings.