Assessing the diagnostic test accuracy of natriuretic peptides and ECG in the diagnosis of left ventricular systolic dysfunction: a systematic review and meta-analysis.

BACKGROUND In 2003 the National Institute of Clinical Excellence published guidelines recommending the use of brain natriuretic peptide (BNP) and the electrocardiogram (ECG) as part of the diagnostic work up of individuals with heart failure. However, the guideline did not address whether one test was superior to the other or whether performing both tests was superior to performing single tests. AIM To investigate the relative test accuracy of the ECG, BNP, N terminal-pro brain natriuretic peptide (NT-proBNP) and combinations of two or more tests in the diagnosis of left ventricular systolic dysfunction (LVSD) in the primary care setting. DESIGN OF STUDY Cohort studies making within-subject comparisons of intervention diagnostic test(s) with reference standard results. METHOD Standard systematic review methodology was followed. RESULTS Thirty-two primary studies met the review inclusion criteria. Studies were of variable quality and highly clinically heterogeneous, therefore restricting the use of meta-analysis. Within these limitations BNP, NT-proBNP and the ECG all had similar test sensitivity (>80% in the majority of studies). Specificity of the three tests was not as good. Three studies directly comparing BNP and the ECG found no difference in sensitivity and limited support for improved specificity of BNP. Two studies found no difference in sensitivity and limited evidence for an improvement in specificity for the combination of the ECG and BNP compared to single tests. CONCLUSION On the basis of existing evidence, the ECG, BNP and NT-proBNP are useful in excluding a diagnosis of LVSD (good sensitivity). However, use of abnormal test results to select individuals for echocardiography may overwhelm services. There is currently no evidence to justify the use of one test over another or the use of tests in combination. The additional cost of BNP is not self-evidently justified by improved test accuracy. Further research is needed to directly compare the diagnostic performance of these tests in homogeneous, representative primary care populations.