Multimarker panel to rule out acute coronary syndromes in low-risk patients.

Abstract

To test novel markers of acute coronary syndrome (ACS), monocyte chemoattractant protein-1 (MCP), myeloperoxidase (MPO), C-reactive protein (CRP), and brain natriuretic peptide (BNP) in low-risk emergency department (ED) patients who were evaluated for ACS in a chest pain unit (CPU). A convenience sample of 414 patients underwent CPU evaluation, including provocative testing, and were followed prospectively for 45 days for ACS, which was defined as death, myocardial infarction (MI), revascularization, or >60% coronary artery stenosis prompting new medical treatment, adjudicated by three blinded reviewers. Published diagnostic thresholds were used to calculate diagnostic indices for each marker and for the multimarker panel. The prevalence of ACS was 7 in 414 (1.7%; 95% CI = 0.7% to 3.5%). Only MCP demonstrated a negative likelihood ratio [LR(-)] of less than 0.5, with a sensitivity of 85% (95% CI = 42% to 99%), specificity of 72% (95% CI = 67% to 76%), and LR(-) of 0.20 (95% CI = 0.04 to 0.71). For MPO, CRP, and BNP, LR(-) was 0.89 (95% CI = 0.26 to 2.05), 0.79 (95% CI = 0.40 to 1.01), and 0.90 (95% CI = 0.51 to 1.03), respectively. The sensitivity, specificity, and LR(-) of an abnormal multimarker panel were 86% (95% CI = 42% to 100%), 17% (95% CI = 13% to 21%), and 0.84 (95% CI = 0.15 to 3.12), respectively. The prevalence of ACS was very low but was similar to reports from other CPUs. BNP and CRP had high specificities, but had limited sensitivities, whereas MPO had a low specificity. Only MCP had a low LR(-) and should be studied further. The combined multimarker panel had an unexpectedly low sensitivity and specificity, yielding an LR(-) of 0.84, suggesting that the panel would not be an efficient screening test to decrease unnecessary CPU testing.