OBJECTIVETo improve the accuracy of diagnosis of heart failure (HF) has been the focus of research for a long time. The diagnosis for HF with congenital heart disease, however, is more difficult. The aim of the study was to evaluate the diagnostic criteria for HF in children and examine the value of plasma brain natriuretic peptide (BNP) and NT-proBNP for diagnosing HF in pediatric patients with congenital heart disease, and to look for the most valuable index for the diagnosis according to the multifactor analysis.

METHODSTotally 118 children with congenital heart disease were enrolled. They were diagnosed using modified Ross score, Qingdao criteria, NYU PHFI, and plasma BNP and NT-proBNP. According to modified Ross score as the referent criteria, other diagnostic criteria and plasma BNP and NT-proBNP were studied. The sensitivity, specificity and area of the ROC curve were examined. Logistic regression analysis was used to select the valuable index for diagnosing HF.

RESULTS(1) The value of each clinical criteria: 1 The sensitivity of Qingdao criteria for diagnosing HF was 47.9%. The specificity was 100% and the accuracy was 57.6%. 2 There were 52 patients younger than six months in whom 27 (51.9%) were breast fed. Only 25 children were measured with Ross score. The Ross score was positively correlated with the modified Ross score (r = 0.948). The area under the ROC curve of Ross score diagnosing HF was 0.985, and the sensitivity was 88%, while the specificity was 100%. 3 NYU PHFI score was positively correlated with the modified Ross score. The area under the ROC curve of the NYU PHFI diagnosing HF was 0.964, and the sum of sensitivity and specificity was favorite when > or = 8 was set as the cut-off point. If > 2 was set as cut-off point, it had a high sensitivity but a low specificity. The sensitivity of NYU PHFI was 100% > was set 2 as cut-point for diagnosing HF, but the specificity was 4.5%. (2) Plasma BNP and NT-proBNP were positively correlated with the modified Ross score, and increased with the severity of congestive HF. The area under the ROC curve of BNP was 0.880, and the cut-off line was > or = 349 pg/ml. The area under the ROC curve of NT-proBNP was 0.981, and the cut-off line was > or = 499 fmol/ml. (3) Logistic regression analysis showed that in multifactor analysis, only plasma concentration of NT-proBNP, dyspnea, tachycardia, tachypnea, failure to thrive were the independent predictors for diagnosing HF. (4) Plasma concentration of NT-proBNP incorporated with clinical criteria would improve its accuracy.

CONCLUSIONAll the clinical criteria commonly used were valuable for diagnosing HF in children with congenital heart disease, but each has its own limits, such as the low sensitivity of Qingdao, the low adaptation of Ross score because of the high breast-feeding rate in our country and the low specificity of NYU PHFI when > 2 was set as the cut-off point. Plasma concentrations of BNP and NT-proBNP were valuable for diagnosing HF in children with congenital heart disease, and NT-proBNP was the independent predictor for HF.