B-type natriuretic peptide (BNP) and N-terminal pro-BNP measurements are used for the diagnosis of congestive heart failure (HF). However, the diagnostic value of these tests is unknown under septic conditions. We compared patients with severe sepsis or septic shock and patients with acute HF to unravel the influence of the underlying diagnosis on BNP and N-terminal pro-BNP levels. Prospective, clinical study. Academic medical intensive care unit (ICU). A total of 249 consecutive patients were screened for the diagnosis of sepsis or HF. Sepsis was defined according to published guidelines. HF was diagnosed in the presence of an underlying heart disease and congestive HF, pulmonary edema, or cardiogenic shock. BNP and N-terminal pro-BNP were measured from blood samples that were drawn daily for routine analysis. We identified 24 patients with severe sepsis or septic shock and 51 patients with acute HF. At admission, the median (range) BNP and N-terminal pro-BNP levels were 572 (13-1,300) and 6,526 (198-70,000) ng/L in patients with sepsis and 581 (6-1,300) and 4,300 (126-70,000) ng/L in patients with HF. The natriuretic peptide levels increased during the ICU stay, but the differences between the groups were not significant. Nine patients with sepsis and eight patients with HF were monitored with a pulmonary artery catheter. Mean (sd) pulmonary artery occlusion pressure were 16 (4.2) and 22 (5.3) mm Hg (p = .02), and cardiac indexes were 4.6 (2.8) and 2.2 (0.6) L/min/m (p = .03) in patients with sepsis and HF, respectively. Despite these clear hemodynamic differences BNP and N-terminal pro-BNP levels were not statistically different between the two groups. In patients with severe sepsis or septic shock, BNP and N-terminal pro-BNP values are highly elevated and, despite significant hemodynamic differences, comparable with those found in acute HF patients. It remains to be determined how elevations of natriuretic peptide levels are linked to inflammation and sepsis-associated myocardial dysfunction.