This article has Open Peer Review reports available.
Usefulness of N-terminal pro-B-type natriuretic peptide in patients admitted to the intensive care unit: a multicenter prospective observational study
- Chin Kook Rhee1,
- So Yeon Lim2,
- Shin Ok Koh3,
- Won-Il Choi4,
- Young-Joo Lee5,
- Gyu Rak Chon6,
- Je Hyeong Kim7,
- Jae Yeol Kim8,
- Jaemin Lim9,
- Sunghoon Park10,
- Ho Cheol Kim11,
- Jin Hwa Lee12,
- Ji Hyun Lee13,
- Jisook Park14,
- Younsuck Koh15,
- Gee Young Suh2,
- Seok Chan Kim1Email author and
- Validation of Simplified acute physiology score3 in Korean Intensive care unit (VSKI) study group & the Korean study group on respiratory failure (KOSREF)
© Rhee et al.; licensee BioMed Central Ltd. 2014
Received: 11 August 2013
Accepted: 25 February 2014
Published: 10 March 2014
The role of N-terminal pro-B-type natriuretic peptide (NT-pro-BNP) as a prognostic factor in patients admitted to the intensive care unit (ICU) is not yet fully established. We aimed to determine whether NT-pro-BNP is predictive of ICU mortality in a multicenter cohort of critically ill patients.
A total of 1440 patients admitted to 22 ICUs (medical, 14; surgical, six; multidisciplinary, two) in 15 tertiary or university-affiliated hospitals between July 2010 and January 2011 were assessed. Patient data, including NT-pro-BNP levels and Simplified Acute Physiology Score (SAPS) 3 scores, were recorded prospectively in a web-based database.
The median age was 64 years (range, 53–73 years), and 906 (62.9%) patients were male. The median NT-pro-BNP level was 341 pg/mL (104–1,637 pg/mL), and the median SAPS 3 score was 57 (range, 47–69). The ICU mortality rate was 18.9%, and hospital mortality was 24.5%. Hospital survivors showed significantly lower NT-pro-BNP values than nonsurvivors (245 pg/mL [range, 82–1,053 pg/mL] vs. 875 pg/mL [241–5,000 pg/mL], respectively; p < 0.001). In prediction of hospital mortality, the area under the curve (AUC) for NT-pro-BNP was 0.67 (95% confidence interval [CI], 0.64–0.70) and SAPS 3 score was 0.83 (95% CI, 0.81–0.85). AUC increment by adding NT-pro-BNP is minimal and likely no different to SAPS 3 alone.
The NT-pro-BNP level was more elevated in nonsurvivors in a multicenter cohort of critically ill patients. However, there was little additional prognostic power when adding NT-pro-BNP to SAPS 3 score.
KeywordsN-terminal pro-B-type natriuretic peptide Intensive care unit Critical care Prognosis
B-type natriuretic peptide (BNP) is released from cardiac ventricles in response to increased wall tension . Measurement of BNP is useful in establishing the diagnosis of heart failure . N-terminal pro-B-type natriuretic peptide (NT-pro-BNP), a precursor of BNP, provides prognostic information superior to that obtained from BNP in patients with myocardial infarction . NT-pro-BNP has been shown to be a good prognostic marker in patients with cardiac disease .
NT-pro-BNP levels are elevated not only in patients with cardiac disease but also in critically ill patients. Okkonen et al.  showed that the NT-pro-BNP level on patient admission is commonly elevated in patients with acute respiratory failure. Wang et al.  reported that elevated NT-pro-BNP levels may prove to be a powerful predictor of mortality in septic patients. NT-pro-BNP level may be a prognostic marker in critically ill patients [7–10]. However, few large-scale multicenter studies have assessed NT-pro-BNP as a prognostic factor. Also, little is known regarding the correlation between NT-pro-BNP and clinical parameters in critically ill patients. Moreover, it is not clear in which conditions NT-pro-BNP is elevated. Thus, by using a large cohort we aimed to assess the prognostic value of NT-pro-BNP level in patients admitted to the intensive care unit (ICU) and analyzed the relationship between the level of NT-pro-BNP and clinical parameters.
We used data from the “Validation of simplified acute physiology score 3 in Korean ICUs” (VSKI) study cohort. VSKI was a prospective multicenter cohort study that aimed to validate the simplified acute physiology score (SAPS) 3 in Korean ICU patients that was performed by the Korean Study group on Respiratory Failure (KOSREF) between July 1st, 2010 and January 31st, 2011. VSKI included 22 ICUs (medical, 14; surgical, six; multidisciplinary, two) in 15 tertiary or university-affiliated hospitals. The study was approved by the institutional review board of the Seoul St. Mary’s Hospital, Samsung Medical Center, Severance Hospital, Keimyung University Dongsan Hospital, Aju University Hospital, Konkuk University Chungju Hospital, Korea University Ansan Hospital, Armed Forces Capital Hospital, Chung-Ang University Hospital, Gangneung Asan Hospital, Hallym University Sacred Heart Hospital, Gyeongsang National University Hospital, Ewha Womans University Hospital, Bundang CHA Hospital and Asan Medical Center. The requirement for informed consent was waived because of the observational nature of the study.
All patients admitted to the 22 ICUs during the study period were eligible for the present study. Patients whose NT-pro-BNP levels were measured on ICU admission were included in the study. We excluded patients who were younger than 17 years. Patients who were transferred from other participating ICUs were also excluded. For patients with two or more admissions to the ICU during the same hospital stay, only the data from the first admission were used. Patients whose ICU or hospital mortality was uncertain were excluded in the analysis.
Patient data were recorded prospectively in a web-based database. We obtained data on demographic characteristics (age, gender, body weight, and height), underlying disease, SAPS 3, sequential organ failure assessment (SOFA) score, severe sepsis, or septic shock on ICU admission, acute lung injury (ALI) or acute respiratory distress syndrome (ARDS) on ICU admission, admission category (medical or surgical), admission diagnosis, organ support (mechanical ventilation, renal replacement therapy, and use of vasopressors), length of ICU stay and hospital stay, and mortality on ICU and hospital discharge. The medical history of each patient was reviewed thoroughly, and the initial vital signs on ICU admission were recorded. Laboratory data, including complete blood cell count, chemistry, and arterial blood gas analysis were collected within 24 h of ICU admission. Illness severity was assessed using the SAPS 3 score. Blood NT-pro-BNP level was determined together with all SAPS 3 variables.
We compared the level of NT-pro-BNP on admission in patients with hospital survivors and nonsurvivors. Then, we compared the mortality rate according to the quintile of NT-pro-BNP. We also compared the level of NT-pro-BNP according to the reason of ICU admission and analyzed correlation between NT-pro-BNP and clinical parameters. We calculated prognostic power of NT-pro-BNP to predict mortality and compared with that of SAPS 3 score. Finally, we analyzed if there was additive benefit when combining NT-pro-BNP with SAPS 3 score.
Because most of the data were not normally distributed, all results are presented as medians and interquartile ranges (IQRs), or as numbers (percentages) of patients. Differences between groups were assessed using chi-squared test or Fisher’s exact test for categorical variables, as appropriate. For continuous variables, differences between two groups were assessed by the Mann–Whitney U-test and differences among four groups were assessed by the Kruskal–Wallis test. Correlations between continuous variables were assessed by Spearman’s test. Receiver operating characteristic (ROC) curves were generated for NT-pro-BNP and SAPS 3. SAPS 3 and NT-pro-BNP were evaluated for their association with hospital survival by logistic regression analysis. NT-pro-BNP was transformed into a natural log (ln) scale because of the wide range of levels. All tests were two sided, and P values less than 0.05 were considered to indicate statistical significance. All statistical analyses were performed using PASW Statistics, version 17 (SPSS, Chicago, IL, USA).
Baseline characteristics of patients (n = 1440)
Median (IQR) or No. (%)
Age in years (range)
Status at ICU admission
Severe sepsis or septic shock
ALI or ARDS
Reason for ICU admission
Basic & observational**
SAPS 3 score
Level of NT-pro-BNP (pg/mL) according to the underlying disease or clinical situation
Severe sepsis or septic shock (-)
Severe sepsis or septic shock (+)
ALI or ARDS (-)
ALI or ARDS (+)
Vasoactive drug (-)
Vasoactive drug (+)
278 (93 – 1,164)
Correlations between NT-pro-BNP and clinical parameters
Correlation coefficient (p)
Highest total bilirubin
Lowest HCO3 -
Lowest PF ratio
The area under the ROC curve in prediction of hospital mortality
SAPS3 + lnNT-pro-BNP (continuous variable)
SAPS3 + lnNT-pro-BNP (categorical variable)
Comparison of clinical characteristics according to the quartile of NT-pro-BNP level
Number of patients
Status at ICU admission
Severe sepsis or septic shock
ALI or ARDS
Reason for ICU admission
Basic & observational**
SAPS 3 score
Logistic regression for prediction of hospital mortality (SAPS 3 alone)
SAPS 3 score
Logistic regression for prediction of hospital mortality (SAPS 3 + NT-pro-BNP)
SAPS 3 score
In the present study, we showed that the serum level of NT-pro-BNP on admission has weak prognostic power and little additive prognostic power when combined with SAPS 3 score. Although previous studies [7–10] have shown that the NT-pro-BNP level can be a prognostic factor in critically ill patients, the present study represents the largest scale study.
In the present study, the serum NT-pro-BNP level on admission was significantly higher in ICU nonsurvivors than in survivors. Moreover, the mortality progressively increased with increasing levels of NT-pro-BNP. Our findings are compatible with previous reports. All four studies [7–10] consistently showed similar results.
Compared with previous studies, the prognostic power of NT-pro-BNP was similar or somewhat weaker. Almog et al.  showed that the area under the ROC curve of NT-pro-BNP levels for prediction of mortality was 0.75 (95% CI, 0.62–0.88). Meyer et al.  showed that the area under the ROC curve was 0.70 (95% CI, 0.64–0.77). In the present study, the area under the ROC curve was 0.67 (95% CI, 0.64–0.70). The reason for the lower prognostic power in the present study is unknown. To address this issue, additional validation of the prognostic power of NT-pro-BNP is needed. One possible explanation is that the population enrolled in the two previous studies may have been biased. Thus, the prognostic power of NT-pro-BNP might be overestimated. In the study by Almog and colleagues , the number of patients analyzed was only 78, and all were admitted to the medical ICU in a single center. In the study by Meyer and colleagues , 289 patients in only a single center were evaluated. Although Meyer et al.  included the entire spectrum of medical patients with a critical illness, many had cardiac disease because the ICU in that study was in the Department of Cardiology. In those patients, the NT-pro-BNP level was more likely to have prognostic power. Compared with the previous two studies, a much larger number of number of patients were enrolled in the present study (n = 1440), and they were admitted not only to the medical ICU but also to the surgical ICU at multiple centers. Single measurement of the NT-pro-BNP level on admission can be an attractive prognostic factor if it is sufficiently powerful because other scoring systems, such as the SAPS 3, are somewhat complex and time consuming. Unfortunately, our data suggest that the NT-pro-BNP level alone is not a powerful prognostic factor in patients admitted to the ICU.
Along with its weak power for prediction of mortality, our data suggest that the NT-pro-BNP level plays little adjunct role when combined with SAPS 3 score. This result is not compatible with previous reports. Meyer et al.  showed that the SAPS 2 score and the NT-pro-BNP level were independently associated with hospital survival in a logistic regression model. Kotanidou et al.  reported that the Acute Physiology and Chronic Health Evaluation (APACHE) II score and NT-pro-BNP level were independent predictors of mortality in multiple logistic regression analysis. The reason for the NT-pro-BNP level having little adjunct role in this study when combined with SAPS 3 is unknown. One possible explanation is that, in the present study, the NT-pro-BNP level was well correlated with many clinical parameters that are components of SAPS 3. Thus, prognosis may be already predicted enough by only SAPS 3 score.
In the present study, when combined with the SAPS 3 score, the AUC was increased little. As shown in Tables 3, many clinical parameters and characteristics were overlapped by the SAPS 3 score and NT-pro-BNP level. Thus, when analyzed using the combination of the SAPS 3 and NT-pro-BNP level, the power of NT-pro-BNP was offset. For example, renal complication is the reason whose level of NT-pro-BNP is highest among the reason of ICU admission (Figure 3). This suggests that the prognostic power of NT-pro-BNP is due partly to the poor prognosis of renal failure patients. However, blood urea nitrogen, creatinine, pH, HCO3 -, and urine output are components of the SAPS 3 score. The NT-pro-BNP level was also well correlated with these parameters, which are important prognostic factors in patients with renal failure. The poor prognosis of renal failure is already sufficiently predicted by the SAPS3 score. Thus, NT-pro-BNP had only a limited effect upon combination with the SAPS 3 score.
Conversely to usual biomarkers, severity score have been proven to be poorly calibrated to predict individual mortality. This may be because of differences in the patient case-mix or changing medical practice over time . For example, SAPS 3 score showed mixed result in the external validation studies [12–17]. In this study, this phenomenon was also observed. Although SAPS 3 score was independent predictor for mortality, the area under the ROC curve was only 0.83. For NT-pro-BNP, there has been also variability for the power to predict mortality in previous studies. Thus, further validation study for NT-pro-BNP is mandatory.
The limitation of this study is potential selection bias. In this study, there were no definite criteria for the measurement of NT-pro-BNP. Basically, this study was multicenter observational study. Thus, in some centers, NT-pro-BNP was measured routinely for all patients, while in other centers, NT-pro-BNP was measured by clinician’s preference. Thus, selection bias for enrollment of patients may exist. However, the level of NT-pro-BNP was variable, which suggests that patients who were unlikely to have cardiac disease were also enrolled. Baseline characteristics also showed that the percentage of ischemic heart disease or heart failure was not high (11.1 and 7.5%, respectively), which supports selection bias for patients with cardiac disease was unlikely to occur.
The present study is valuable in that it was of a multicenter, prospective design. In the present study, multiple ICU settings (medical or surgical or multidisciplinary) in multiple centers were analyzed. Thus, the enrolled population may be representative of real-world ICU patients. To be a good prognostic marker, NT-pro-BNP should be validated in a large-scale multicenter study. Our results provide valuable information regarding the limited role of NT-pro-BNP in general ICU patients. Further investigation of the role of NT-pro-BNP in critically ill patients with various clinical characteristics is needed.
In conclusion, the NT-pro-BNP level was more elevated in nonsurvivors in a multicenter cohort of critically ill patients. However, there was little additional prognostic power when adding NT-pro-BNP to SAPS 3 score. Further investigation of NT-pro-BNP as a prognostic factor in patients admitted to the ICU is needed.
The members of VSKI are as follows. Samsung Medical Center, Sungkyunkwan University School of Medicine (Gee Young Suh, So Yeon Lim, Kyeongman Jeon, Chi Min Park), Asan Medical Center, University of Ulsan College of Medicine (Younsuck Koh, Chae-Man Lim, Sang-Bum Hong, Jin Won Huh, Suk Kyung Hong), Severance Hospital, Yonsei University College of Medicine (Shin Ok Koh, Sungwon Na), Keimyung University, Dongsan Hospital (Won-Il Choi), Aju university college of medicine (Young-Joo Lee, Kwang Joo Park), Seoul St. Mary’s Hospital, Catholic University of Korea (Seok Chan Kim, Chin Kook Rhee, Chan Kwon Park), Chungju hospital, School of medicine of Konkuk university (Gyu Rak Chon), Korea University Ansan Hospital (Je Hyeong Kim), Chung-Ang University College of Medicine (Jae Yeol Kim), Gangneung Asan Hospital, Gangneung, University of Ulsan Medical College of internal medicine (Jaemin Lim), Hallym University Sacred Heart Hospital (Sunghoon Park), Gyeongsang Institute of Health Sciences, Gyeongsang National University (Ho Cheol Kim), Ewha Womans University School of Medicine (Jin Hwa Lee), Bundang CHA hospital CHA university (Ji Hyun Lee), Seoul Women’s University (Jisook Park), Inha University College of Medicine (Jae Hwa Cho).
- Maisel AS, Krishnaswamy P, Nowak RM, McCord J, Hollander JE, Duc P, Omland T, Storrow AB, Abraham WT, Wu AH, Clopton P, Steg PG, Westheim A, Knudsen CW, Perez A, Kazanegra R, Herrmann HC, McCullough PA: Rapid measurement of B-type natriuretic peptide in the emergency diagnosis of heart failure. N Engl J Med. 2002, 347: 161-167. 10.1056/NEJMoa020233.View ArticlePubMedGoogle Scholar
- Baughman KL: B-type natriuretic peptide – a window to the heart. N Engl J Med. 2002, 347: 158-159. 10.1056/NEJMp020057.View ArticlePubMedGoogle Scholar
- Richards AM, Nicholls MG, Yandle TG, Frampton C, Espiner EA, Turner JG, Buttimore RC, Lainchbury JG, Elliott JM, Ikram H, Crozier IG, Smyth DW: Plasma N-terminal pro-brain natriuretic peptide and adrenomedullin: new neurohormonal predictors of left ventricular function and prognosis after myocardial infarction. Circulation. 1921–1929, 1998: 97-Google Scholar
- Omland T, Persson A, Ng L, O’Brien R, Karlsson T, Herlitz J, Hartford M, Caidahl K: N-terminal pro-B-type natriuretic peptide and long-term mortality in acute coronary syndromes. Circulation. 2002, 106: 2913-2918. 10.1161/01.CIR.0000041661.63285.AE.View ArticlePubMedGoogle Scholar
- Okkonen M, Varpula M, Linko R, Perttila J, Varpula T, Pettila V: N-terminal-pro-BNP in critically ill patients with acute respiratory failure: a prospective cohort study. Acta Anaesthesiol Scand. 2011, 55: 749-757. 10.1111/j.1399-6576.2011.02439.x.View ArticlePubMedGoogle Scholar
- Wang F, Wu Y, Tang L, Zhu W, Chen F, Xu T, Bo L, Li J, Deng X: Brain natriuretic peptide for prediction of mortality in patients with sepsis: a systematic review and meta-analysis. Crit Care. 2012, 16: R74-10.1186/cc11331.View ArticlePubMedPubMed CentralGoogle Scholar
- Almog Y, Novack V, Megralishvili R, Kobal S, Barski L, King D, Zahger D: Plasma level of N terminal pro-brain natriuretic peptide as a prognostic marker in critically ill patients. Anesth Analg. 1809–1815, 2006: 102-Google Scholar
- Meyer B, Huelsmann M, Wexberg P, Delle Karth G, Berger R, Moertl D, Szekeres T, Pacher R, Heinz G: N-terminal pro-B-type natriuretic peptide is an independent predictor of outcome in an unselected cohort of critically ill patients. Crit Care Med. 2007, 35: 2268-2273. 10.1097/01.CCM.0000284509.23439.5B.View ArticlePubMedGoogle Scholar
- Kotanidou A, Karsaliakos P, Tzanela M, Mavrou I, Kopterides P, Papadomichelakis E, Theodorakopoulou M, Botoula E, Tsangaris I, Lignos M, Ikonomidis I, Ilias I, Armaganidis A, Orfanos SE, Dimopoulou I: Prognostic importance of increased plasma amino-terminal pro-brain natriuretic peptide levels in a large noncardiac, general intensive care unit population. Shock (Augusta, Ga). 2009, 31: 342-347. 10.1097/SHK.0b013e31818635b6.View ArticleGoogle Scholar
- De Geer L, Fredrikson M, Oscarsson A: Amino-terminal pro-brain natriuretic peptide as a predictor of outcome in patients admitted to intensive care. A prospective observational study. Eur J Anaesthesiol. 2012, 29: 275-279. 10.1097/EJA.0b013e32835470a8.View ArticlePubMedGoogle Scholar
- Lim SY, Koh SO, Jeon K, Na S, Lim CM, Choi WI, Lee YJ, Kim SC, Chon GR, Kim JH, Kim JY, Lim J, Rhee CK, Park S, Kim HC, Lee JH, Park J, Koh Y, Suh GY: Validation of SAPS3 admission score and its customization for use in Korean intensive care unit patients: a prospective multicentre study. Respirology (Carlton, Vic). 2013, 18: 989-995. 10.1111/resp.12115.View ArticleGoogle Scholar
- Khwannimit B, Bhurayanontachai R: The performance and customization of SAPS 3 admission score in a Thai medical intensive care unit. Intensive Care Med. 2010, 36: 342-346. 10.1007/s00134-009-1629-7.View ArticlePubMedGoogle Scholar
- Lim SY, Ham CR, Park SY, Kim S, Park MR, Jeon K, Um SW, Chung MP, Kim H, Kwon OJ, Suh GY: Validation of the simplified acute physiology score 3 scoring system in a korean intensive care unit. Yonsei Med J. 2011, 52: 59-64. 10.3349/ymj.2011.52.1.59.View ArticlePubMedGoogle Scholar
- Soares M, Salluh JI: Validation of the SAPS 3 admission prognostic model in patients with cancer in need of intensive care. Intensive Care Med. 1839–1844, 2006: 32-Google Scholar
- Ledoux D, Canivet JL, Preiser JC, Lefrancq J, Damas P: SAPS 3 admission score: an external validation in a general intensive care population. Intensive Care Med. 1873–1877, 2008: 34-Google Scholar
- Metnitz B, Schaden E, Moreno R, Le Gall JR, Bauer P, Metnitz PG: Austrian validation and customization of the SAPS 3 Admission Score. Intensive Care Med. 2009, 35: 616-622. 10.1007/s00134-008-1286-2.View ArticlePubMedGoogle Scholar
- Poole D, Rossi C, Anghileri A, Giardino M, Latronico N, Radrizzani D, Langer M, Bertolini G: External validation of the Simplified Acute Physiology Score (SAPS) 3 in a cohort of 28,357 patients from 147 Italian intensive care units. Intensive Care Med. 1916–1924, 2009: 35-Google Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2253/14/16/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.