| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
*Medical Intensive Care Unit and the Departments of Medicine and Cardiology, Soroka University Medical Center, Ben Gurion University of the Negev, Beer-Sheva, Israel
Address correspondence and reprint requests to Yaniv Almog, MD, Medical Intensive Care Unit, Soroka University Medical Center, P.O.B. 151, Beer-Sheva, 84101, Israel. Address e-mail to almogya{at}bgu.ac.il.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Over the last decade, several studies have indicated that cardiac dysfunction is a frequent and important factor in determining outcome of critically ill patients (5). Few studies have addressed the value of natriuretic peptides in predicting outcome of critically ill patients. These studies were small, measured atrial natriuretic peptide (ANP) or BNP and were confounded by a large proportion of patients with CHF (6–9). Serum half-lives of the NT peptides (NT-pro ANP and NT-proBNP) are considerably longer than those of other natriuretic peptides, resulting in plasma levels that are at least 5 to 15 times larger. Thus, they are less influenced by the conditions under which the sample is drawn and may therefore be more appropriate in critical care settings (10). Moreover, although BNP and NT-pro-BNP have equal predictive values for heart failure, the latter is probably more sensitive in predicting death (11). The aim of this prospective study was to determine if NT-pro-BNP measured at intensive care unit (ICU) admission is an independent marker of myocardial dysfunction that correlates with mortality in a heterogeneous group of critically ill patients.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Multiple demographic, clinical, and laboratory data were recorded daily by an investigator who did not participate in routine patient care and was blinded to the NT-pro-BNP level. The APACHE II score was calculated within 6 h of ICU admission based on the worst values up to that point. Serum levels of cardiac troponin T (cTnT) were measured within 6 h of admission. Blood was drawn for NT-pro-BNP determination within 6 h of admission, and serum was frozen at –20°C. It has been previously shown that NT-pro-BNP remains stable under these conditions for as long as 1 yr (13). After completion of patient recruitment, serum samples were thawed, and NT-pro-BNP levels were determined on the same day using a commercial assay (Roche, Dyn Diagnostics, Indianapolis, IN). The analytical range extended from 20 to 35,000 pg/mL. The interassay coefficient of variation was 3.2% at 175 pg/mL, 2.9% at 355 pg/mL, and 2.6% at 1068 pg/mL (14). cTnT levels were determined using a commercial micro-particle enzyme immunoassay (AxSYM Troponin-T; Abbott Laboratories, Abbott Park, IL). The assay characteristics were as follows: detection limits of 0.3 ng/mL, analytical range of 0–50 ng/mL, and assay coefficient of variation range of 5.2%–7.8%. The cTnT cutoff is 0.03 ng/mL with a coefficient of variation of 10%. Thus, a cTnT serum level >0.03 ng/mL was considered abnormal.
Echocardiography was performed within 24 h using a hand-carried cardiac ultrasound by a noninvasive cardiologist blinded to clinical conditions and NT-pro-BNP levels. LV systolic function was assessed by visual estimation of the LV ejection fraction (LVEF) and scored from normal to severely reduced as follows: normal (LVEF 
50%), mildly reduced (LVEF 40% to < 50%), moderately reduced (LVEF 30% to < 40%), and severely reduced (LVEF < 30%). The portable hand-carried cardiac ultrasound (OptiGo; Philips Medical Systems, Andover, MA) is a small battery-powered device that provides two-dimensional imaging and color Doppler. The two-dimensional images were obtained from parasternal long and short axis, apical 4, 3, and 2 chambers, and subcostal view. This method is reliable and reproducible in assessing LV dysfunction (15).
The primary end point was 30-day mortality. Continuous variables were expressed as mean ± sd. Median values and interquartile ranges (IQR) are presented for non-normally distributed continuous variables. Normality of the study variables was tested with a 1-sample Kolmogorov-Smirnov test to indicate the appropriateness of parametric testing. The unpaired Student's t-test was performed for comparative analysis of normally distributed variables, and the Mann-Whitney test was used for nonparametric analysis. Correlations between continuous variables were assessed by the Pearson test and Spearman test for nonparametric variables. Categorical variables were expressed as percentages, and comparisons between groups were made using the 
2 test. To define an optimal decision threshold of NT-pro-BNP level for 30-day mortality prediction, a receiver operator characteristic plot analysis was performed. The Youden index (namely, sensitivity + specificity –1) was used to determine the optimal NT-pro-BNP threshold. Multivariate logistic regression analysis was used to test whether this threshold of NT-pro-BNP level was an independent predictor of mortality. The final model included two variables: NT-pro-BNP levels and APACHE II score. Cumulative survival curves were constructed using the Kaplan-Meier method and compared with the log-rank test. To further examine the relation between levels of NT-pro-BNP and mortality, patients were divided into quartiles of NT-pro-BNP levels. Results were considered significant at P < 0.05. SPSS 13.0 software was used for statistical analysis.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Table 1 describes the baseline characteristics of the study population. The median age was 60 yr (range, 18–82 yr; IQR, 48–74), and 62.8% of the patients were men. The median number of chronic conditions was 2 (IQR, 1–4). Infection was the primary reason for ICU admission in 48 of 78 patients (61.5%), followed by respiratory failure in 17 (21.8%), neurological conditions in 5 (6.4%), metabolic disturbances in 5 (6.4%), gastrointestinal bleeding in 2 (2.6%), and drug overdose in 1 patient (1.3%). Fifty-three patients (68.0%) developed respiratory failure requiring mechanical ventilation at admission, and an additional five were tracheally intubated later during their ICU course. The median length of ICU stay was 4 days (IQR, 2–8 days).
|
The median level of NT-pro-BNP on admission was 1700 pg/mL (range, 23–35,000 pg/mL; IQR, 400–9000 pg/mL). There was no correlation between age and NT-pro-BNP levels (P = 0.48), nor was there any gender effect on NT-pro-BNP levels (men, 1800 pg/mL; IQR 400–10,000, versus women, 1700 pg/mL; IQR, 700–7400; P = 0.99). Patients who required mechanical ventilation had higher NT-pro-BNP levels than those who did not (2200 pg/mL; IQR, 700–10,000, versus 700 pg/mL; IQR, 400–2100; P < 0.001). Patients admitted to the ICU because of an infection had higher levels of NT-pro-BNP compared with the rest of the cohort (5100 pg/mL; IQR, 800–13,100, versus 1000 pg/mL; IQR, 400–1800; P = 0.001). None of the patients in either group had ischemic electrocardiograph changes during their ICU stay.
Figure 1 describes the receiver operator characteristic curve for NT-pro-BNP levels as a predictor of 30-day mortality. The area under the curve was 0.75 (95% confidence interval [CI], 0.62–0.88), and the optimal threshold for NT-pro-BNP was 1900 pg/mL. At this cutoff, the sensitivity for prediction of 30-day mortality was 77.3% and the specificity was 66%, whereas the accuracy was 69% (Youden index = 0.43). At this cutoff, the positive likelihood ratio was 2.4 (95% CI, 1.5–3.7).
|
Table 2 depicts characteristics of the patients stratified according to NT-pro-BNP level less than and more than the threshold of 1900 pg/mL. Kaplan-Meier survival plots (Fig. 2) for the entire population stratified according to NT-pro-BNP levels more than or less than 1900 pg/mL show that the difference in mortality rates becomes evident 5 days after admission.
|
|
Analyzing the characteristics of patients admitted with severe sepsis (n = 38; 48.7%) compared to those without reveals that patients admitted with severe sepsis tended to be younger (54.0 ± 20.2 versus 66.2 ± 13.2; P = 0.002) and had more failing systems (median, 3; IQR, 2–4, versus median, 2; IQR, 1–3; P = 0.006). The prevalence of moderate to severe LV dysfunction was similar between the two groups, and there was no significant difference in mortality between patients with and without severe sepsis (21.1% versus 17.5%; P = 0.46). Figure 3 describes the distribution of NT-pro-BNP levels between survivors versus nonsurvivors (Fig. 3A) and between those with or without severe sepsis on admission (Fig. 3B). The median NT-pro-BNP levels among patients admitted with severe sepsis were 6000 pg/mL (IQR, 850–16,000).
|
Analyzing the risk of mortality relative to quartiles of NT-pro-BNP levels reveals that 30-day mortality progressively increases with increasing levels of NT-pro-BNP. Thus, in the upper quartile of NT-pro-BNP levels, 11 of 20 patients died, reflecting a mortality rate of 55.0% (95% CI, 33.4%–76.6%) compared with 2 of 19 (10.5%; 95% CI, 0%–24.6%) in the lower quartile (P = 0.005).
The final multivariate logistic regression analysis revealed that an NT-pro-BNP level more than the threshold of 1900 pg/mL is the only independent factor predicting 30-day mortality (odds ratio [OR], 5.6; 95% CI, 1.74–18.11). The APACHE II score at admission was not significantly associated with mortality (OR, 1.05 per point increment; 95% CI, 0.97–1.14).
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Characteristically, both BNP and NT-pro-BNP levels reported in various cardiac pathologies, such as IHD and CHF, are in the range of hundreds of picograms per milliliter (1,2,4,16). Few previous studies reported NT-pro-BNP levels similar to those observed in our study. Large concentrations of NT-pro-BNP (median 8877 pg/mL; range, 2940–29,372 pg/mL) were shown to be related to noninfectious severe systemic inflammatory response syndrome associated with cardiovascular dysfunction after coronary artery surgery (17). Recently, marked increases of NT-pro-BNP (range, 2163–35,000 pg/mL) were reported in a small cohort of 6 septic shock patients (8). In a prospective observational study evaluating 39 patients with septic shock requiring mechanical ventilation, the IQR of NT-pro-BNP in the nonsurvivors was 11,735, compared to 49,320 pg/mL in the survivors (11).
LV wall tension is believed to be the primary mechanism regulating BNP secretion (2). Other hemodynamic factors that may contribute to NT-pro-BNP secretion include LV diastolic dysfunction and right ventricular overload and dysfunction. These are all known causes of NT-pro-BNP release and frequently occur in critically ill patients (6,8,11).
Several additional mechanisms may be proposed to account for these high values. Serum creatinine is a factor affecting BNP levels (14). Thus, renal failure may have contributed to the levels observed in our study, particularly in the group with NT-pro-BNP levels more than 1900 pg/mL, where renal failure was more frequent. Selective upregulation of cardiac BNP occurs at the transcriptional and translational levels by proinflammatory cytokines and by conditioned medium derived from mixed lymphocyte reactions via p38 mean arterial blood pressure kinase (3). Moreover, transcriptional activation of the BNP gene in rat cardiac monocytes by lipopolysaccharide has been described, as was interleukin-1 beta upregulation of the human BNP promoter in cardiac monocytes (18,19). Critical illness, in general, and sepsis, in particular, are associated with an intense inflammatory syndrome characterized by markedly increased circulating proinflammatory cytokines (20). We included patients with APACHE II scores more than 12 to avoid those with mild disease who are typically admitted for brief periods. Consequently, patients with higher acuity and complexity were enrolled. Thus, the high levels of NT-pro-BNP observed in our cohort may be attributed to inflammation-induced, cytokine-mediated BNP upregulation and synthesis.
Over the last decade, several studies have indicated that cardiac dysfunction is a frequent and important factor in determining outcome of critically ill patients (21). The pathophysiology of myocardial injury in critically ill patients is complex and multifactorial. It is estimated that as many as 15%–30% of ICU admissions are complicated by some degree of myocardial injury, and as many as 40% of patients with severe sepsis may have cardiac dysfunction. Such cardiac dysfunction during sepsis has been associated with poor prognosis (22). Moreover, as many as 85% of patients with severe sepsis may have increased cTnT, a sensitive surrogate of myocardial injury (22).
Our data are consistent with these reports that 30.6% of patients with NT-pro-BNP levels more than 1900 pg/mL had moderate to severe LV dysfunction, compared with only 9.5% of those with lower levels, a statistically significant difference; 33.3% of patients with NT-pro-BNP levels more than 1900 pg/mL also had abnormal cTnT levels. We used bedside echocardiography to assess LV systolic function after hemodynamic stabilization (50.0% were receiving vasoactive therapy at the time of evaluation). Because this method is dependent upon loading conditions that may be highly variable in these patients, the addition of a biologic marker to the evaluation of myocardial performance may be of supplementary value. Despite growing enthusiasm regarding the diagnostic and therapeutic monitoring utility of BNP in CHF and dyspnea of undetermined etiology (1,16), a recent study assessing the utility of BNP for the evaluation of ICU shock suggested that it cannot be used as a surrogate for pulmonary artery catheterization (23). Future research will be required to determine the precise role of NT-pro-BNP measurement in the evaluation and management of myocardial dysfunction during critical illness and severe sepsis.
The use of cardiac biomarkers to predict outcome in critically ill patients is a growing field of research. cTns have been extensively studied as prognostic tools in critically ill patients, yielding conflicting results. Whereas some researchers have suggested that cTnT levels correlate with myocardial damage and poor outcome, our group and others did not confirm this association (5,22,24,25). Our data do not support an independent predictive value for cTnT. Several studies (6,9,23) have evaluated the prognostic value of natriuretic peptides (ANP and BNP) in critically ill patients. Although these studies were limited by small sample sizes and some were confounded by a large proportion of patients with CHF, they uniformly suggested a strong independent prognostic value for increased natriuretic peptides. In fact, one of these studies argued that extreme increases in BNP levels more strongly predicted mortality than the APACHE II score (23). Two recent studies evaluated NT-pro-BNP as a prognostic factor and a marker of myocardial dysfunction in patients with septic shock. Both studies suggest that NT-pro-BNP is an early predictor of prognosis and myocardial dysfunction (2,11).
Our data are consistent with these observations, suggesting that admission NT-pro-BNP levels more than 1900 pg/mL are an independent marker of both ICU and 30-day mortality. However, with an accuracy of 69% and the lower border of 95% CI for positive likelihood ratio (1.5), the clinical implications of this association may be limited and should be interpreted with caution.
The prognostic impact of NT-pro-BNP level is further highlighted by analyzing mortality rates according to quartiles of NT-pro-BNP level. This analysis reveals a dose-response trend of increased mortality with increasing NT-pro-BNP levels, although the significance of this finding is restricted by insufficient power. The significance of increase NT-pro-BNP in critically ill patients will have to await further research and validation to determine its precise role as a prognostic tool.
The present study included a relatively small number of patients. Thus, the predictive power of our findings is somewhat limited, as illustrated by the wide 95% CI around the point estimate of NT-pro-BNP area under the curve. However, most studies evaluating natriuretic peptides as prognostic markers in critically ill patients have been of similar size. The significance of post hoc subgroup analysis (i.e., sepsis patients) and our ability to identify other independent determinants of ICU mortality are also limited. As indicated earlier, only systolic function was assessed at a single point in time, and serial measurements of NT-pro-BNP were not performed. In addition, echocardiography was performed within 24 hours of admission, whereas NT-pro-BNP was measured within 6 hours, resulting in time inconsistency between assessments. Thus, additional associations between other components of myocardial function and NT-pro-BNP levels could not be ascertained. Although the time course of NT-pro-BNP and its relation to clinical course, myocardial dysfunction, and outcome may be of interest, the main purpose of our study was to determine whether early NT-pro-BNP increase is clinically relevant. Finally, the large variability in NT-pro-BNP levels may prove to be an important limitation in its future widespread application.
We conclude that, among critically ill patients, NT-pro-BNP values are highly variable. An ICU admission NT-pro-BNP level more than 1900 pg/mL is an independent marker of short-term mortality. Larger studies will validate the optimal threshold and independent predictive power of NT-pro-BNP in the critical care setting.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. Sodeck, H. Domanovits, M. Schillinger, K. Janata, M. Thalmann, M. P. Ehrlich, G. Endler, and A. Laggner Pre-operative N-terminal pro-brain natriuretic peptide predicts outcome in type A aortic dissection. J. Am. Coll. Cardiol., March 18, 2008; 51(11): 1092 - 1097. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Sear and G. Howard-Alpe Preoperative plasma BNP concentrations: do they improve our care of high-risk non-cardiac surgical patients? Br. J. Anaesth., August 1, 2007; 99(2): 151 - 154. [Full Text] [PDF]
|
|
|
|
 |
|
 E. A. Lucas and F. H. Wians Jr. NT-proBNP as a Prognostic Marker in Critically Ill Patients Anesth. Analg., April 1, 2007; 104(4): 998 - 999. [Full Text] [PDF]
|
|
|
|
|
|
Y. O. Almog NT-proBNP as a Prognostic Marker in Critically Ill Patients Anesth. Analg., April 1, 2007; 104(4): 999 - 1000. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
