PT  - JOURNAL ARTICLE
AU  - Lixia Wang
AU  - Yinan Yang
AU  - Quanmiao Cui
AU  - Ya Cui
AU  - Qiaoe Li
AU  - Xinyao Che
AU  - Cong Wang
AU  - Peiqin Quan
AU  - Xiaobin Hu
TI  - Evaluating the added predictive ability of MMP-9 in serum for Kawasaki disease with coronary artery lesions
AID  - 10.1136/jim-2020-001281
DP  - 2021 Jan 01
TA  - Journal of Investigative Medicine
PG  - 13--19
VI  - 69
IP  - 1
4099  - http://hw-f5-jim.highwire.org/content/69/1/13.short
4100  - http://hw-f5-jim.highwire.org/content/69/1/13.full
SO  - J Investig Med2021 Jan 01; 69
AB  - To investigate the predictive ability of serum matrix metalloproteinase-9 (MMP-9) in the acute phase of Kawasaki disease (KD) with coronary artery lesions (CALs). Patients with KD hospitalized in Lanzhou University Second Hospital, Northwest China, from November 2015 to January 2018 were retrospectively reviewed, and clinical trial indicators and peripheral blood specimens were collected before intravenous immunoglobulin therapy treatment. The independent risk factors were determined using multivariate regression analysis. The net reclassification improvement (NRI) and integrated discrimination improvement (IDI) were used to quantitatively evaluate the ability of MMP-9 to improve the efficiency of predicting KD with CALs. The white cell, neutrophil percentage, C-reactive protein (CRP), and erythrocyte sedimentation rate (ESR) were higher in patients with higher MMP-9, and the monocyte percentage was higher in patients with lower MMP-9. Logistic regression analysis revealed that long-term fever; elevated CRP, ESR, platelets (PLT), and MMP-9; and low albumin (ALB) levels were independent predictors of KD with CALs. A predictive model of KD with CALs using fever duration, CRP, ESR, PLT, and ALB showed significantly improved predictive ability when MMP-9 was added to the model (the area under the curve increased by 0.02; no change in sensitivity; specificity increased from 81.48% to 87.04%; NRI value: 13.46%; IDI value: 5.00%, p<0.05). Adding MMP-9 to traditional risk factors may improve prediction of CALs, the overall predictive ability of model 2 was increased by 5%.