BACKGROUND: Brain natriuretic peptide (BNP) and N-terminal (NT)-proBNP have recently emerged as promising parameters for risk assessment in acute pulmonary embolism (PE). However, their positive predictive value is low, and the prognostic implications of NT-proBNP or troponin elevation alone are questionable. METHODS AND RESULTS: To determine whether the combination of NT-proBNP testing with echocardiography may identify both low-risk and high-risk patients with PE, we examined 124 consecutive patients with proved PE. All underwent echocardiography on admission to detect right ventricular dysfunction. NT-proBNP and troponin concentrations were measured in one core laboratory. The primary end point was death or major in-hospital complications. The cutoff level of 1000 pg/mL had a high negative predictive value (95% for a complicated course, 100% for death), but NT-proBNP > or =1000 pg/mL did not independently predict an adverse outcome. Combination of NT-proBNP testing with echocardiography identified 3 major risk groups. A positive echocardiogram was associated with a 12-fold elevation in complication risk compared with patients with low NT-proBNP (P=0.002), whereas NT-proBNP elevation without right ventricular dysfunction on echocardiography only slightly increased the risk of an adverse outcome (P=0.17). The combination of cardiac troponin testing with echocardiography yielded similar complication rates in the lowest-risk group and a similar magnitude of risk elevation for the highest-risk patients, but it also increased the number of intermediate-risk groups. CONCLUSIONS: Our results support a simple risk stratification algorithm for patients with PE, with the use of NT-proBNP or troponin testing as an initial step that should be followed by echocardiography if elevated levels of the biomarker are found.