Safety of human serum albumin infusion in heart failure patients with hypoproteinemia: a propensity score-matched analysis

Figures

• 

  Fig. 1. Study flow diagram in the present study.

• 

  Fig. 2. Performance in predicting in-hospital mortality using albumin concentration. Prediction performance was based on ROC. The area under curve was 0.643. The cut-off value was 2.9 g/dL. ROC, Receiver Operator Curves.

• 

  Fig. 3. Effects of variables on hospital mortality in stepwise multivariate logistic regression analysis in heart failure patients with albumin > 2.9 g/dL (A) and with albumin ≤ 2.9 g/dL (B). ALB_infu, Albumin Infusion; SOFA Sequential Organ Failure Assessment; WBC_max, maximum White Blood Cell count during hospital stay; PLT_max, maximum blood Platelet count during hospital stay; Hematocrit, Initial hematocrit during hospital stay; Potassium_max, maximum Potassium during hospital stay; Cre_max, maximum Creatinine during hospital stay; BUN_max, maximum Urea Nitrogen during hospital stay; Sodium_max, maximum Sodium during hospital stay; COPD, Chronic Obstructive Pulmonary Disease.

• 

  Fig. 4. Characteristics of covariate between albumin administration and non-albumin administration with albumin ≤ 2.9 g/dL (*p < 0.05; ***p < 0.001). ALB_infu, Albumin infusion; Sofa, Sequential Organ Failure Assessment; WBC_max, maximum White Blood Cell count during hospital stay; PLT_max, maximum Blood Platelet count during hospital stay; Hematocrit, Initial hematocrit during hospital stay; Potassium_max, maximum Potassium during hospital stay; Cre_max, maximum Creatinine during hospital stay; BUN_max, maximum Urea Nitrogen during hospital stay; Sodium_max, maximum Sodium during hospital stay; ALB_C, initial Albumin concentration.

• 

  Fig. 5. Covariates including initial albumin concentration, SOFA score, maximum white blood cell count, maximum blood platelet count, maximum creatinine level, maximum serum urea nitrogen level, maximum serum sodium level, chronic obstructive pulmonary disease was adjusted by Propensity Score Matching (PSM) between albumin administration and non-albumin administration groups with albumin ≤ 2.9 g/dL and the caliper was 0.05. After PSM, 982 patients with albumin ≤ 2.9 g/dL were stayed.

• 

  Fig. 6. The relationship between amount of human albumin infusion and in-hospital mortality in the HF group (A) and HF plus albumin ≤ 2.9 g/dL group (B).

• 

  Fig. 7. Survival during 90 days in albumin administration and non-albumin administration subgroups before (A) and after (B) PSM when albumin ≤ 2.9 g/dL. PSM, Propensity Score Matching.