BACKGROUND AND OBJECTIVE: Correct and early risk stratification for critically ill pneumonia patients remains an unmet medical need. This study aimed to test whether N-terminal pro B-type natriuretic peptide (NT-proBNP) can serve as a prognostic marker in this setting. METHODS: This prospective study enrolled 216 pneumonia patients admitted to intensive care unit. Plasma NT-proBNP samples were obtained upon admission and primary outcome was all-cause mortality at 30 days. Meanwhile, Acute Physiology and Chronic Health Evaluation (APACHE) II and Infectious Diseases Society of America/American Thoracic Society (IDSA/ATS) 2007 minor criteria were assessed. RESULTS: Overall 30-day mortality was 30%. NT-proBNP levels were significantly higher in nonsurvivors than survivors (11 938 ± 13 121 vs 5658 ± 9240 pg/mL, P = 0.001). Area under receiver operating characteristic curves of NT-proBNP, APACHE II and IDSA/ATS 2007 minor criteria were not significantly different regarding prediction of mortality (0.715, 0.754 vs 0.654, P = 0.085). Adding NT-proBNP to APACHE II significantly increased the area under receiver operating characteristic curve from 0.754 to 0.794 (P = 0.048). Receiver operating characteristic analysis revealed optimal NT-proBNP and APACHE II cut-offs of 2177.5 pg/mL and 25.5, respectively. In multivariate analysis, both NT-proBNP and APACHE II values above cut-offs had a significantly higher probability of death than those below cut-offs. A categorical approach combining NT-proBNP and APACHE II cut-offs provides additional risk stratification over a single marker approach. CONCLUSIONS: For pneumonia patients admitted to intensive care unit, NT-proBNP strongly and independently predicts mortality, and its prognostic accuracy is comparable with APACHE II and IDSA/ATS 2007 minor criteria.
BACKGROUND AND OBJECTIVE: Correct and early risk stratification for critically ill pneumoniapatients remains an unmet medical need. This study aimed to test whether N-terminal pro B-type natriuretic peptide (NT-proBNP) can serve as a prognostic marker in this setting. METHODS: This prospective study enrolled 216 pneumoniapatients admitted to intensive care unit. Plasma NT-proBNP samples were obtained upon admission and primary outcome was all-cause mortality at 30 days. Meanwhile, Acute Physiology and Chronic Health Evaluation (APACHE) II and Infectious Diseases Society of America/American Thoracic Society (IDSA/ATS) 2007 minor criteria were assessed. RESULTS: Overall 30-day mortality was 30%. NT-proBNP levels were significantly higher in nonsurvivors than survivors (11 938 ± 13 121 vs 5658 ± 9240 pg/mL, P = 0.001). Area under receiver operating characteristic curves of NT-proBNP, APACHE II and IDSA/ATS 2007 minor criteria were not significantly different regarding prediction of mortality (0.715, 0.754 vs 0.654, P = 0.085). Adding NT-proBNP to APACHE II significantly increased the area under receiver operating characteristic curve from 0.754 to 0.794 (P = 0.048). Receiver operating characteristic analysis revealed optimal NT-proBNP and APACHE II cut-offs of 2177.5 pg/mL and 25.5, respectively. In multivariate analysis, both NT-proBNP and APACHE II values above cut-offs had a significantly higher probability of death than those below cut-offs. A categorical approach combining NT-proBNP and APACHE II cut-offs provides additional risk stratification over a single marker approach. CONCLUSIONS: For pneumoniapatients admitted to intensive care unit, NT-proBNP strongly and independently predicts mortality, and its prognostic accuracy is comparable with APACHE II and IDSA/ATS 2007 minor criteria.