Clinical Question: Does an elevated level of brain natriuretic peptide (BNP) predict an adverse outcome in patients with pulmonary embolism (PE)?

Setting: Inpatient (any location)

Study Design: Cohort (prospective)

Synopsis: Large PEs can cause right ventricular dysfunction, which is, in turn, associated with an increase in the levels of BNP and other cardiac enzymes. This small study examines the ability of BNP levels to serve as an aid in the prognosis of patients with PE.

Consecutive patients with acute PE admitted to a tertiary care center had their BNP levels measured on admission and were followed prospectively. Adverse outcomes were defined as death; cardiopulmonary resuscitation; mechanical ventilation; or the need to use pressors, thrombolysis, catheter fragmentation, or surgical embolectomy. We are not told whether the researchers who judged the outcomes were blinded to the BNP levels of the participants, but we must assume they were not.

Of 73 patients in the study, 20 had an adverse outcome. A BNP level of less than 90 pg per mL was 85 percent sensitive and 75 percent specific for the prediction of a benign course (negative predictive value, 93 percent; positive predictive value, 57 percent). Three patients with symptoms for fewer than 12 hours had a false-negative BNP, and five patients with chronic cardiac or lung disease had a false-positive BNP.

The authors argue that they would get better results with a cutoff level of 50 pg per mL, but this requires prospective confirmation. Another option would be to wait 12 hours before drawing the BNP. To its credit, BNP was a stronger independent predictor of adverse events than was troponin-T.

Bottom Line: This study provides preliminary evidence that measurement of BNP is a useful prognostic tool in patients with PE; 93 percent of patients with a BNP value of less than 90 pg per mL had a benign course. The small size of this study, failure to blind observers, and questions about the best time to draw blood and the best cutoff level mean we need more information before making this a routine part of our practice. BNP may eventually have a role, if validated, in identifying patients for outpatient therapy. (Level of Evidence: 4)