Prognostic Value of B-type Natriuretic Peptide (BNP) and Its Potential Role in Guiding Fluid Therapy in Critically Ill Septic Patients

Zhongheng Zhang,Zhengguang Zhang,Yadong Xue,Xiao Xu,Hongying Ni
DOI: https://doi.org/10.1186/1757-7241-20-86
2012-01-01
Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine
Abstract:Background and objectives The prognostic role of B-type natriuretic peptide (BNP) in septic patients is controversial. The study aimed to investigate the prognostic value of BNP in critically ill septic patients. Furthermore, because BNP is primarily released from ventricles in response to myocardial stretch, the second aim of the study was to test whether the change in BNP was correlated to the amount of fluid balance. Methods This was a prospective observational study conducted in a tertiary 18-bed ICU. Patients fulfilled criteria of sepsis were enrolled. Exclusion criteria included renal dysfunction on ICU entry, age < 18 or >80 years old. BNP was measured on entry (BNP0) and day 3 (BNP1) and daily fluid balance over the three days were recorded. Diagnostic performances of BNP0 and ΔBNP (BNP1-BNP0) were analyzed. The correlation between fluid balance and ΔBNP were tested using Spearman’s correlation test. Results A total of 67 subjects were eligible for the study during study period. BNP0 was significantly higher in non-survivors than in survivors (738 vs 550 pg/ml, p < 0.01). The area under curves (AUCs) of BNP0 in predicting mortality, duration of mechanical ventilation (MV) > 7 d, length of stay in ICU (LOS ICU ) > 7 d and hospital (LOS hospital ) > 12 d were 0.71, 0.79, 0.66 and 0.71, respectively. The AUCs of ΔBNP in predicting duration of MV > 7 d, LOS ICU > 7 d and LOS hospital > 12 d were 0.80, 0.84 and 0.85, respectively. The amount of fluid balance was correlated to ΔBNP (Spearman’s rho = 0.63, p < 0.01), and the correlation remained statistically significant in multivariate model. Conclusions BNP measured on ICU entry is associated with mortality, duration of MV, LOS ICU and LOS hospital . ΔBNP is able to predict the LOS ICU and LOS hospital with satisfactory sensitivity and specificity. ΔBNP is closely correlated to the amount of fluid balance during resuscitation period. However, this could only be considered as a hypothesis-generating pilot study due to its small sample size and the observational nature.
What problem does this paper attempt to address?