| Aims:To explore the clinical features, to indentify independent risk factors and/or to establish new predictive models for incomplete KD and transitional KD, infection-positive KD, KD with delayed initial IVIG treatment, IVIG resistant KD, KD with CAL at acute and subacute stages, and KD with persistent/transient CAL CAA/CAD at convalescent stage.Methods:A single-center retrospective study was performed to analyze data from patients with KD and echocardiographic follow-ups 1 month post-onset between January 2009 and December 2014 in our hospital. Univariable analysis was applied to screen for factors with significant differences, and further indentify independent predictors by multivariable logistic regression analysis so as to construct predictive models. Finally, ROC was used to determine the area under curve, sensitivity and specificity of these models.Results:1. Basic epidemiologic and clinical features. In our data, the male to female ratio was 1.62:1,71.3% of the total patients were ranged from six months to five years of age; KD hospitalization occurring in May and June accounted for 22.8%, and in December and January for 13.7%. Binary linear regression was used to analyze the relationship between KD hospitalization and local PM10 with R2=0.0342 and p=0.12, and local PM2.5 with R2= 0.0888 and p= 0.157. The percentages of IVIGRKD and KD with CAL at acute and subacute stages were 17.9% and 24.1%, respectively. The percentages of KD with transient CAL or persistent CAL, and with CAD, CAA or GCA at convalescent stage were 13.7%,12.6%,12.5%,2.3% and 0.5%, respectively.2. Compared to patients with iKD, the parameters of CBC (pre-and post-IVIG treatment ESR, CRP, WBC, NE% and absolute NE count) and biochemical tests (ALT, TBil, GGT and serum ferritin) were significantly elevated and the level of ALB was significantly decreased in patients with cKD. The ratio of delayed treatment with IVIG (delayed diagnosis) was increased in patients with iKD (3.8% vs 10.5%, p= 0). There were no significant differences in the frequencies of IVIGR (18.7% vs.17.3%, p= 0.392) and CAL (22.4% vs.25.3%, p= 0.115) between patients with iKD and cKD. Compared to patients with cKD and iKD,8 and 33 laboratory indicators were significant difference in patients with tKD. The rates of IVIGR among groups were no significant difference. The rate of delayed treatment with IVIG in patients with tKD was no significant difference compared to patients with cKD but was significantly lower than those with iKD (3% vs 3.8%.12.4%; p= 0.557 and 0, respectively). The rate of CAL in patients with tKD was significantly lower than those with cKD and iKD (15.5% vs 22.4%,22.7%; p= 0.013 and 0, respectively).3. Compared to patients with KD at typical age, the principal diagnostic criteria was significantly decreased in younger patients but simliar in older patients, the occurrence of delayed treatment with IVIG was significantly increased in older patients but similar in younger patients, and the parameters of CBC (pre-and post-IVIG treatment WBC, LY%, EO%, MO%, Plt, PCV and absolute counts for LY, EO, MO) were significantly elevated in younger patients but decreased in older patients, the rate of infection-positive KD and the levels of ALB and serum Na were significantly decreased in younger patients but increased in older patients. There was no difference with regard to gender, the occurrences of IVIGR and CAL between groups.4. Compared to infection-negative group, the ratio of male to female was significantly decreased and the percentage of patients with KD aged more than five years old was significantly increased in infection-positive group. There were no difference between groups with regard to the occurrences of delayed treatment with IVIG, IVIGR, and CAL, the parameters of CBC (pre-and post-IVIG treatment ESR, WBC, NE, absolute NE count, LY, and NLR) and biochemical tests (ALT, AST, ALB and TBil).5. Compared to KD without delayed diagnosis, the ratio of male to female, the percentage of patients with KD older than five years old and the frequency of CAL were significantly increased (2.23:1 vs 1.57:1,p=0.047; 19.6% vs 13.2%,p= 0.026; 44.3% vs 21.6%, p= 0), and the parameters of CBC (pre-and post-IVIG treatment CRP, NE, absolute NE count, and NLR) and biochemical tests (ALT, TBil, GGT, and serum ferritin) were significantly decreased in KD with delayed diagnosis. Multivariable regression analysis showed that illness days on admission longer than 9 days with an OR of 19.5, total fever duration longer than 10 days with an OR of 36.1 and the fewer the principal diagnostic criteria the higher the OR for delayed diagnosis.6. Multivariable regression analysis for IVIGRKD showed that the OR of total fever duration longer than 7 days, delayed dignosis, GGT≥25U/L, serum Na<135mmol/L, pre-IVIG NLR≥2.8, and pre-IVIG Plt<350×10^9/L were 2.94,1.64, 1.38,1.68,1.58, and 1.39, respectively. These indicators were combined to predict IVIGR with AUC of 0.685, and the sensitivity and specificity were 60.7% and 66.5%, respectively. Multivariable regression analysis for IVIGRKD younger than 6 months showed that the OR of total fever duration longer than 7 days, serum Na<135mmol/L, pre-IVIG NLR≥1.9, and pre-IVIG Plt<350x 10^9/L were 3.63,2.02,3.11, and 2.20, respectively. These indicators were combined to predict IVIGRKD younger than 6 months with AUC of 0.746, and the sensitivity and specificity were 82.1% and 53.8%, respectively.7. Multivariable regression analysis for KD with CAL at acute and subacute stages showed that the OR of male, total fever duration longer than 8 days, IVIGR, ALB<35.9g/L and pre-IVIG MO>5.9% were 1.45,1.78,1.42,1.53, and 1.36, respectively. These indicators were combined to predict CAL with AUC of 0.634. Multivariable regression analysis for KD younger 6 months with CAL showed that the OR of total fever duration longer than 8 days, delayed treatment with IVIG, and ALB<35.9g/L were 3.61,3.49, and 2.07, respectively. These indicators were combined to predict KD younger than 6 months with CAL with AUC of 0.731, and the sensitivity and specificity were 64.7% and 80.9%, respectively.8. Multivariable regression analysis for KD with persistent CAL at convalescent stage showed that the OR of male, total fever duration longer than 8 days, delayed diagnosis, IVIGR, ALB<35.9g/L and pre-IVIG EO>2.2% for KD-NCAL were 1.50, 1.91,3.52,2.24,1.62, and 1.51, respectively, which combined to predict persistent CAL with AUC of 0.707, and the sensitivity and specificity were 67.8% and 63.1%, respectively; and also showed that the OR of IVIGR, GGT≥25U/L, and pre-IVIG Plt>400x 10A9/L for KD with transient CAL were 1.70,1.55, and 2.09, respectively, which combined to predict persistent CAL with AUC of 0.631. Multivariable regression analysis for KD with transient CAL showed that the OR of male, total fever duration longer than 8 days, serum Na≤135mmol/L and pre-IVIG NE<59% for KD-NCAL were 1.45,1.61,1.46, and 1.38, respectively, which combined to predict transient CAL with AUC of 0.601.9. Multivariable regression analysis for KD with CAA at convalescent stage showed that the OR of male, total fever duration longer than 8 days, delayed treatment with IVIQ and IVIGR for KD-NCAL were 3.17,5.16,3.71, and 5.70, respectively, which combined to predict CAA with AUC of 0.845, and the sensitivity and specificity were 70.2% and 87.5%, respectively, and also showed the OR of age at onset older than 5 years old, total fever duration longer than 8 days, and IVIGR for CAD were 4.44,3.41, and 4.10, respectively. Multivariable regression analysis for KD with CAD showed that the OR of male and pre-IVIG ESR for KD-NCAL were 1.61 and 1.59, respectively.Conclusions:1. Male and age at onset of illness ranged from 6 months to 5 years old were more frequent in KD. KD hospitalization peaked at May and June and had a nadir at December and January, and had no linear relationship with PM2.5 and PM10. More than half of CAL regressed to normal at convalescent stage.2. Patients with cKD had a more severe extent of inflammation and patients with iKD were more likely to be delayed diagnosis. tKD was one of transitional status of KD and early recognition contributed to avoid delayed diagnosis and to decrease the occurrence of CAL.3. Patients with KD showed less severe inflammation and more infection-positive cases. Patients aged at the extreme of onset were more likely to be delayed diagnosis.4. Female and older patients with KD are more likely to be co-infection with pathogens and infection-positive cases showed similar inflammation, clinical presentations, treatment and outcomes with infection-negative cases.5. KD with delayed diagnosis showed more frequent in male and age older than 5 years, and more likely to have CAL. Independent risk factors for delayed diagnosis included longer illness days on admission and total fever duration, and less principal diagnostic criteria.6. Independent risk factors for IVIGRKD included longer total fever duration, delayed diagnosis, higher GGT and pre-IVIG NLR, and lower serum Na and pre-IVIG Plt. Further study in subgroup younger than 6 months could optimize predictive model.7. Independent risk factors for acute and subacute KD with CAL included male, longer total fever duration, IVIGR, lower ALB and higher pre-IVIG MO%. Further study in subgroup younger than 6 months could optimize predictive model.8. Independent risk factors for convalescent KD with persistent CAL included male, longer total fever duration, delayed treatment, IVIGR, lower ALB and higher pre-IVIG ESR and EO%, which generating a model to predict persistent CAL.9. Independent risk factors for convalescent KD with CAA included male, longer total fever duration, delayed treatment, and IVIGR, which producing a better model to predict CAA. |
PDF Full Text Request