Research

Prognostic value of brain natriuretic peptide in acute pulmonary embolism

Guillaume Coutance¹ , Olivier Le Page² , Ted Lo¹ and Martial Hamon^1,3

¹Service des Maladies du Coeur et des Vaisseaux, UF Soins Intensifs Cardiologiques, Centre Hospitalier Universitaire de Caen, Avenue Côte de Nacre, 14033 Caen Cedex, Normandy, France

²Service de Chirurgie Cardiaque, Centre Hospitalier Universitaire de Caen, Avenue Côte de Nacre, 14033 Caen Cedex, Normandy, France

³Inserm 744, Institut Pasteur de Lille, 1 rue du Professeur Calmette, 59019 Lille cedex, France

author email corresponding author email

Critical Care 2008, 12:R109doi:10.1186/cc6996

Published:	22 August 2008

Abstract

Introduction

The relationship between brain natriuretic peptide (BNP) increase in acute pulmonary embolism (PE) and the increase in mortality and morbidity has frequently been suggested in small studies but its global prognostic performance remains largely undefined. We performed a systematic review and meta-analysis of data to examine the prognostic value of elevated BNP for short-term all-cause mortality and serious adverse events.

Methods

The authors reviewed PubMed, BioMed Central, and the Cochrane database and conducted a manual review of article bibliographies. Using a prespecified search strategy, we included a study if it used BNP or N-terminal pro-BNP biomarkers as a diagnostic test in patients with documented PE and if it reported death, the primary endpoint of the meta-analysis, in relation to BNP testing. Studies were excluded if they were performed in patients without certitude of PE or in a subset of patients with cardiogenic shock. Twelve relevant studies involving a total of 868 patients with acute PE at baseline were included in the meta-analysis using a random-effects model.

Results

Elevated BNP levels were significantly associated with short-term all-cause mortality (odds ratio [OR] 6.57, 95% confidence interval [CI] 3.11 to 13.91), with death resulting from PE (OR 6.10, 95% CI 2.58 to 14.25), and with serious adverse events (OR 7.47, 95% CI 4.20 to 13.15). The corresponding positive and negative predictive values for death were 14% (95% CI 11% to 18%) and 99% (95% CI 97% to 100%), respectively.

Conclusion

This meta-analysis indicates that, while elevated BNP levels can help to identify patients with acute PE at high risk of death and adverse outcome events, the high negative predictive value of normal BNP levels is certainly more useful for clinicians to select patients with a likely uneventful follow-up.