Abstract: Objective  To analyze the predictors of infantile bronchopneumonia with acute heart failure in pediatric intensive care unit (PICU). Methods  Totally 381 infants with bronchopneumonia in our hospital were retrospectively analysed,  among them, 94 cases had acute heart failure (AHF) and 287 cases had no acute heart failure (nonAHF). The univariate and multivariate logistic regression analysis was conducted,  and the ROC curve was applied to evaluate the predictive efficacy of each factor,  then the optimal cutoff value was selected. Results  There were statistical differences in malnutrition,  arrhythmia,   NTproBNP,  CTNI,  CKMB,  neutrophil to lymphocyte ratio(NLR),  DDimer  between AHF group and nonAHF group (P<0.05). The variables that eventually entered the regression model were NTproBNP,  CTNI,  CKMB and DDimer (P<0.05). The areas under the ROC curves (AUC) of NTproBNP,  CTNI,  CKMB and DDimer were 0.957,  0.855,  0831 and 0.823,  respectively,  and the best cutoff values were 233.2  pg/ml,  0.035  ng/ml,  31.15  U/L and 1.005  mg/L,  respectively. Conclusion  NTproBNP,  CTNI,  CKMB and DDimer are independent predictors of infantile bronchopneumonia with acute heart failure.

Key words: bronchopneumonia, heart failure, prediction, intensive care units; , pediatric