Direct comparison of three natriuretic peptides for prediction of short- and long-term mortality in patients with community-acquired pneumonia

BACKGROUND: Early and accurate risk stratification for patients with community-acquired pneumonia (CAP) is an unmet clinical need. METHODS: We enrolled 341 unselected patients presenting to the ED with CAP in whom blinded measurements of N-terminal pro-B-type natriuretic peptide (NT-proBNP), midregional pro-atrial natriuretic peptide (MR-proANP), and B-type natriuretic peptide (BNP) were performed. The potential of these natriuretic peptides to predict short- (30-day) and long-term mortality was compared with the pneumonia severity index (PSI) and CURB-65 (confusion, urea plasma level, respiratory rate, BP, age over 65 years). The median follow-up was 942 days. RESULTS: NT-proBNP, MR-proANP, and BNP levels at presentation were higher in short-term (median 4,882 pg/mL vs 1,133 pg/mL; 426 pmol/L vs 178 pmol/L; 436 pg/mL vs 155 pg/mL, all P > .001) and long-term nonsurvivors (3,515 pg/mL vs 548 pg/mL; 283 pmol/L vs 136 pmol/L; 318 pg/mL vs 103 pg/mL, all P > .001) as compared with survivors. Receiver operating characteristics analysis to quantify the prognostic accuracy showed comparable areas under the curve for the three natriuretic peptides to PSI for short-term (PSI 0.76, 95% CI, 0.71-0.81; NT-proBNP 0.73, 95% CI, 0.67-0.77; MR-proANP 0.72, 95% CI, 0.67-0.77; BNP 0.68, 95% CI, 0.63-0.73) and long-term (PSI 0.72, 95% CI, 0.66-0.77; NT-proBNP 0.75, 95% CI, 0.70-0.80; MR-proANP 0.73, 95% CI, 0.67-0.77, BNP 0.70, 95% CI, 0.65-0.75) mortality. In multivariable Cox-regression analysis, NT-proBNP remained an independent mortality predictor (hazard ratio 1.004, 95% CI, 1.00-1.01, P = .02 for short-term; hazard ratio 1.004, 95% CI, 1.00-1.01, P = .001 for long-term, increase of 300 pg/mL). A categorical approach combining PSI point values and NT-pro-BNP levels adequately identified patients at low, medium, and high short- and long-term mortality risk. CONCLUSIONS: Natriuretic peptides are simple and powerful predictors of short- and long-term mortality for patients with CAP. Their prognostic accuracy is comparable to PSI.