Brain natriuretic peptide to predict successful liberation from mechanical ventilation in critically ill patients: a systematic review and meta-analysis

Abstract

Background: Predicting successful liberation from mechanical ventilation (MV) in critically ill patients is challenging. Brain natriuretic peptide (BNP) has been proposed to help guide decision-making for readiness to liberate from MV following a spontaneous breathing trial (SBT).

Methods: We performed a systematic review and meta-analysis of randomized and prospective observational studies that measured BNP levels at the time of SBT in patients receiving MV. The primary endpoint was successful liberation from MV (absence of reintubation or non-invasive ventilation at 48 h). Statistical analyses included bi-variate and Moses-Littenberg models and DerSimonian-Laird pooling of areas under ROC curve (AUROC).

Results: A total of 731 articles were screened. Eighteen adult and 2 pediatric studies were fulfilled pre-specified eligibility. The measure of the relative variation of BNP during SBT (ΔBNP%) after exclusion of SBT failure by clinical criteria in adults yielded a sensitivity and specificity of 0.889 [0.831-0.929] and 0.828 [0.730-0.896] for successful liberation from MV, respectively, with a pooled AUROC of 0.92 [0.88-0.97]. The pooled AUROC for any method of analysis for absolute variation of BNP (ΔBNP), pre-SBT BNP, and post-SBT BNP were 0.89 [0.83-0.95], 0.77 [0.63-0.91], and 0.85 [0.80-0.90], respectively.

Conclusion: The relative change in BNP during a SBT has potential value as an incremental tool after successful SBT to predict successful liberation from MV in adults. There is insufficient data to support the use of BNP in children or as an alternate test to clinical indices of SBT, or the use of ΔBNP, BNP-pre, and BNP-post as an alternate or incremental test.

Trial registration: PROSPERO CRD42018087474 (6 February 2018).

Keywords: Critical care; Intensive care; Natriuretic peptide (brain); Respiration (artificial); Ventilator weaning.

Fig. 1

Flow diagram for study assessment. This diagram illustrates the flow of study selection for the systematic review and meta-analysis

Fig. 2

Flow diagram of group allocation for inclusion versus exclusion of SBT failure in the liberation failure analysis. SBT: Spontaneous breathing trial; successful liberation: absence of non-invasive ventilation or reintubation at 48h; unsuccessful liberation: failure of SBT or non-invasive ventilation or reintubation at 48h. This flow diagram details how the different stages at which unsuccessful liberation from mechanical ventilation were handled in regard to statistical analysis. Group 1 included SBT failure within the unsuccessful liberation umbrella for analysis of liberation failure. Group 2 excluded SBT failure from the umbrella of unsuccessful liberation for the purpose of analysos of liberatio failure

Fig. 3

Bi-variate estimate of sensitivity and specificity for measures of either ∆BNP or ∆BNP% in studies that included SBT failure (group 1) and studies that excluded SBT failure (group 2) from liberation failure analysis. TP: True positive; TN: True negative; FP: False positive; FN: False negative. This diagram illustrates the bi-variate analysis performed on 5 studies and provides an estimate of sensitivity and specificity for studies regardless of the method of BNP measurement (DBNP or DBNP%) and regardless of inclusion or exclusion of SBT failure from the liberation failure analysis (group 1 or 2). This was obtained through analysis of the TP, TN, FP and FN obtained in studies that provided full description of their population and outcomes

Fig. 4

Pooled diagnostic AUC for either ∆BNP or ∆BNP% methods of measurements, in studies that included SBT failure (group 1) and studies that excluded SBT failure (group 2) from liberation failure analysis. These figures illustrate the individual AUC obtained from the studies for each method of BNP measurement separately (DBNP and DBNP%). Both studies that included SBT failure (group 1) and excluded SBT failure (group 2) in liberation failure analysis were included to increase statistical power. The pooled AUC for DBNP% showed a high AUC with low heterogeneity, while the pooled AUC for DBNP showed a high AUC with moderate heterogeneity

Fig. 5

Pooled diagnostic AUC either ∆BNP or ∆BNP% methods of measurements in studies that excluded SBT failure (group 2) from liberation failure analysis. These figures illustrate the individual AUC obtained from the studies for each method of DBNP and DBNP% measurements separately. Only studies that excluded SBT failure (group 2) in liberation failure analysis were included to provide more precise data of a specific clinical subgroup. The pooled AUC for DBNP% showed a high AUC with low heterogeneity. The pooled AUC for DBNP showed a high AUC and high heterogeneity but is limited by the inclusion of only 2 studies in the analysis

Fig. 6

Pooled diagnostic AUC of either BNP-pre or BNP-post methods of measurement in studies that included SBT failure (group 1) and studies that excluded SBT failure (group 2) from liberation failure analysis. These figures illustrate the individual AUC obtained from the studies for each method of individual measures of BNP separately (BNP-pre and BNP-post). Both studies that included SBT failure (group 1) and excluded SBT failure (group 2) in liberation failure analysis were included to increase statistical power. The pooled AUC for BNP-pre showed a moderatehigh AUC with high heterogeneity. The pooled AUC for BNP-post showed a high AUC and low heterogeneity