ML162

Plasma Levels of N-Terminal Pro-Brain Natriuretic Peptide in Patients With Coronary Artery Disease and Relation to Clinical Presentation, Angiographic Severity, and Left Ventricular

Comparative assessment of N-terminal pro-brain natri- uretic peptide (NT–pro-BNP) across a wide spectrum of angiographic and clinical coronary artery disease (CAD) in a consecutive series of patients has not been reported. This study examined 879 subjects (684 patients who had angiographically proved CAD and 195 controls who did not have CAD). NT–pro-BNP concentrations were measured before an angiographic procedure that allowed diagnosis of CAD and measurements of left ventricular ejection fraction and end-diastolic blood pressure. Median values (25th and 75th percentiles) of NT–pro-BNP in patients and controls were 474.5 pg/ml (162.3 and 1,542.8) and 117.0 pg/ml (60.1 and 230.6), respectively (p <0.001). In patients who had stable angina, unstable angina, and acute myocardial infarction, NT–pro-BNP concentrations were 327.7 pg/ml (129.2 and 973.2), 660.6 pg/ml (201.2 and 1,563.5), and 1,045.0 pg/ml (323.8 and 2,486.0, p rain-type natriuretic peptide (BNP) is a cardiac hormone that is synthesized predominantly in the left ventricle in response to wall stress1–3 and other stimuli.4 BNP is produced as a prohormone (pro-BNP) that is enzymatically cleaved into the BNP and the N-terminal fragment (NT–pro-BNP) of the prohor- mone.5 Circulating levels of BNP and NT–pro-BNP are increased in patients who have heart failure in proportion to a decrease in left ventricular function.6,7 Previous clinical studies that involved small numbers of patients have also demonstrated an increase in circulating levels of natriuretic peptides in patients who have stable angina,8 unstable angina,8,9 and acute myocardial infarction (AMI).10–12 However, to our knowledge, a comparative assessment of NT–pro-<0.001). NT–pro-BNP concentrations in subgroups with 1-, 2-, and 3-vessel CAD were 385.5 pg/ml (117.2 and 1,266.0), 463.0 pg/ml (135.0 and 1,480.5), and 533.8 pg/ml (221.8 and 1,809.4), respectively (p = 0.005). Multivariable analysis showed that NT–pro-BNP was an independent correlate of the presence of CAD (chi- square 10.8, odds ratio 1.08, 95% confidence interval 1.03 to 1.13 for 100-pg/ml increase in concentration; p <0.001), acute coronary syndromes (chi-square 6.3, odds ratio 1.01, 95% confidence interval 1.00 to 1.02 for 100-pg/ml increase in concentration, p = 0.01) and a strong trend that was independently associated with angiographic severity (chi-square 3.68, p = 0.055). This study shows that NT–pro-BNP concentrations are high across the entire spectrum of CAD and parallel the clinical or angiographic severity of CAD. ©2005 by Ex- cerpta Medica Inc. BNP across a wide spectrum of angiographic and clinical coronary artery disease (CAD) in a consecu- tive series of patients has not been reported. Further, recent studies have demonstrated that BNP or NT– pro-BNP measurements performed within the first days after acute coronary syndromes provide indepen- dent, predictive information on mortality,13–16 thus raising important implications with regard to the use of BNP and NT–pro-BNP as biomarkers for assess- ment of these patients. Thus, we undertook this study to compare NT–pro-BNP levels across a wide spec- trum of clinical and angiographic CAD in a consecu- tive series of patients and to assess the strength of correlation between left ventricular ejection fraction (LVEF) and NT–pro-BNP levels in patient subgroups with different grades of myocardial ischemia. METHODS Patients: This case-control study examined 879 pa- tients who underwent angiographic examination and cor- onary stenting due to significant CAD from August 2001 to February 2002 in the Deutsches Herzzentrum (Mu- nich, Germany). A consecutive series of 684 patients who had angiographically proved significant CAD (385 patients who had stable angina, 108 who had unstable angina, and 191 who had AMI) formed the patient group. The control group consisted of 195 subjects who had normal coronary artery angiograms collected within the same interval as the patient group. All subjects gave written informed consent for examinations and partici- pation in the study. Diagnostic and risk factor definitions: All subjects un- derwent coronary angiography due to chest pain or other symptoms suggestive of CAD. CAD was confirmed by the presence of coronary stenoses with ≥50% lumen obstruction in ≥3 main coronary arteries. Normal coro- nary arteries were defined as arterial wall irregularities of <10% lumen narrowing (at most) without regional left ventricular wall motion abnormalities. The diagnosis of AMI was established by the presence of chest pain lasting >20 minutes associated with electrocardio- graphic changes (ST-segment elevation ≥1 mm in ≥2 extremity electrocardiographic leads, ≥2 mm in ≥2 contiguous precordial leads, or new-onset left bundle branch block) or increased activity of serum creatine kinase to ≥2 times the upper limit of normal. The diag- nosis of unstable angina was based on well-accepted criteria.17 Arterial hypertension was defined as active management with antihypertensive medications or doc- umented systolic blood pressure ≥140 mm Hg and/or diastolic blood pressure ≥90 mm Hg on ≥2 separate occasions. Hypercholesterolemia was defined as a doc- umented total cholesterol concentration ≥240 mg/dl. Current smokers were defined as those currently smok- ing tobacco. Diabetes mellitus was defined as active management with insulin or oral hypoglycemic agents. For those who used only diet to control diabetes, docu- mentation of an abnormal finding with fasting blood glucose testing or glucose tolerance testing was required.18

Biochemical measurements: Blood was collected be- fore angiography in tubes that contained ethylenedia- minetetraacetic acid and promptly centrifuged. Plasma aliquots were stored frozen at —80°C until assayed within batches. Blood counts, serum lipids, and other metabolites were determined immediately after collec- tion according to standard methods.

NT–pro-BNP measurements were performed on a Roche Elecsys 1010 automated analyzer (Roche Di- agnostics, Mannheim, Germany). The Roche NT–pro- BNP sandwich electrochemiluminescent assay19–21 uses 2 polyclonal antibodies that recognize epitopes located in the N-terminal part (1 to 76 amino acid residues) of pro-BNP (1 to 108 amino acid residues). The measuring range, defined by the lower detection limit and the maximum of the master curve, provided by the manufacturer is 5 to 35,000 pg/ml (Roche Diagnostics ProBNP technical bulletin for Elecsys systems). Functional sensitivity, i.e., the lowest ana- lyte concentration that can be reproducibly measured with a between-run coefficient of variation of 20%, is <50 pg/ml. Expected NT–pro-BNP concentrations in a healthy population depend on age and gender. NT– pro-BNP in the 95th percentile in women increases from 152 pg/ml (at <55 years of age) to 265 pg/ml (at >65 years of age). In men of the same age, NT–pro- BNP increases from 75.8 pg/ml to 157 pg/ml.20

Laboratory personnel were unaware of the results of clinical diagnosis or coronary artery angiography. Statistical analysis: Data are presented as medians (with 25th and 75th percentiles) or counts and propor- tions (percentages). Data distribution was analyzed by 1-sample Kolmogorov-Smirnov test. Categorical data were compared with chi-square test. Continuous data were compared with Kruskal-Wallis rank-sum test. The strength of the link between the NT–pro-BNP concen- tration and LVEF was tested by Spearman’s rank corre- lation coefficient. Multiple logistic regression analysis (binary or ordinal) was applied to identify correlates of the presence of CAD. Multiple linear regression analysis was used to assess the independent influence of various factors on NT–pro-BNP concentration. When testing for independent correlates of CAD, the following variables were entered into the multivariable model: age, gender, diabetes, arterial hypertension, smoking, hypercholester- olemia, NT–pro-BNP concentration (as a continuous variable), C-reactive protein concentration, LVEF, left ventricular end-diastolic pressure, and creatinine concen- tration. When testing for the influence of various factors on NT–pro-BNP levels (as a continuous variable), CAD was entered into the model instead of NT–pro-BNP. A p value <0.05 was considered statistically significant. RESULTS Characteristics of patients and controls: Distribution of the main cardiovascular risk factors among patients and controls is presented in Table 1. As expected, patients versus control subjects had a more adverse cardiovascular risk profile. Patients were older and there were fewer women compared with control sub- jects. Diabetes, arterial hypertension, and hypercho- lesterolemia were encountered in a larger proportion of patients than of controls. Higher levels of C-reac- tive protein and of creatinine were found in patients compared with control subjects. With regard to pa- rameters of left ventricular function, patients’ left ven- tricular function was decreased as demonstrated by smaller values of LVEF and increased end-diastolic pressure compared with control subjects. On an aver- age basis, NT–pro-BNP concentrations were >4 times higher in patients than in control subjects (474.5 pg/ml,162.3 and 1,542.8, vs 117.0 pg/ml, 60.1 and 230.6, respectively; p <0.001). Multiple binary logistic regression analysis was used to test for independent correlates for the presence of CAD. The model showed that gender (chi-square 50.1, adjusted odds ratio [OR] 0.17, 95% confidence interval [CI] 0.10 to 0.27, p <0.001), LVEF (chi-square 48.0, OR 2.5, 95% CI 1.9 to 3.2 for a 10% decrease, p <0.001), hypercholesterolemia (chi- square 39.1, OR 4.0, 95% CI 2.6 to 6.2, p <0.001),age (chi-square 20.8, OR 1.7, 95% CI 1.4 to 2.1 for a 10-year increase, p <0.001), NT–pro-BNP (chi- square 10.8, OR 1.08, 95% CI 1.03 to 1.13 for 100-pg/ml increase in concentration, p <0.001), and arte- rial hypertension (chi-square 4.4, OR 1.63, 95% CI 1.03 to 2.58, p = 0.04) were independent correlates for the presence of CAD. FIGURE 1. Median values of NT–pro-BNP in control subjects and patients who had 1-, 2-, or 3-vessel CAD. NT-proBNP levels according to angiographic and clinical presentations of CAD: There were 159 patients who had 1-vessel CAD, 190 who had 2-vessel CAD, and 335 who had 3-vessel CAD. NT–pro-BNP concentra- tions in subgroups with 1-, 2-, and 3-vessel CAD were 385.5 pg/ml (117.2 and 1,266.0), 463.0 pg/ml (135.0 and 1,480.5), and 533.8 pg/ml (221.8 and 1,809.4), respec- tively (Figure 1; p = 0.005). After adjustment in the multiple ordinal logistic regression model (see Methods for variables entered into the model), NT–pro-BNP con- centration showed a strong trend toward an independent association with the angiographic extent of CAD (coded as 0-, 1-, 2-, and 3-vessel disease), although statistical significance was not achieved (chi-square 3.68, p = 0.055). Variables that were independently associated with angiographic extent of CAD were age, gender, diabetes, hypercholesterolemia, and LVEF. Characteristics of patients according to clinical presentation of CAD are presented in Table 2. A progressive and significant increase in NT–pro-BNP concentration was observed in patient subgroups that had stable angina, unstable angina, and AMI (Table 2 and Figure 2). The independent association of NT–pro-BNP and clinical presentation of CAD was tested in the multiva- riable model. For this reason, patients were categorized into groups with acute coronary syn- dromes (unstable angina and AMI, 299 patients) and without acute coronary syndromes (stable angina, 385 pa- tients). The same parameters that were used for multivariable analysis for presence of CAD were entered into the model. The model showed that LVEF (chi-square 11.2, OR 1.2, 95% CI 1.1 to 1.4 for a 10% decrease, p <0.001), smoking (chi-square 7.9, OR 1.8, 95% CI 1.2 to 2.8, p = 0.005), NT–pro- BNP concentration (chi-square 6.3, OR 1.01, 95% CI 1.00 to 1.02 for 100-pg/ml increase in concentration, p = 0.01), hypercholesterolemia (chi- square 4.5, OR 0.7, 95% CI 0.5 to 0.97, p = 0.03), and gender (chi-square 4.2, OR 1.49, 95% CI 1.02 to 2.19, p = 0.04) were independent correlates of acute coronary syndromes. Correlation between NT-proBNP and left ventricular function: We tested the strength of the link between NT–pro-BNP concentration and LVEF in patients who had stable angina, those who had unstable angina, and those who had AMI. An inverse correlation be- tween NT–pro-BNP and LVEF was found in all 3 subgroups of patients who had CAD. The strongest inverse correlation between NT–pro-BNP and LVEF was found in patients who had stable angina (r = —0.54, p <0.001). In contrast, we found the weakest correlation between NT–pro-BNP and LVEF in pa- tients who had AMI (r = —0.30, p <0.001; Figure 3). In patients who had unstable angina, the coefficient of correlation between NT–pro-BNP concentration and LVEF was —0.48 (p <0.001). The correlation be- tween the 2 variables was significantly weaker in patients who had AMI or unstable angina compared with patients who had stable angina (p <0.001 for the 2 correlation coefficient comparisons). Factors that influence NT-proBNP concentration: Multiple linear regression analysis was performed to identify factors that were independently associated with higher concentrations of NT–pro-BNP (see Methods for variables entered into the model). The model showed that C-reactive protein (p <0.001), creatinine level (p <0.001), LVEF (p <0.001), age (p <0.001), CAD (p <0.001), hypercholesterolemia (p <0.001), gender (p = 0.03), and smoking (p = 0.04) were independent correlates of higher NT–pro-BNP concentrations. The multiple squared coefficient of correlation for the model was 0.34. DISCUSSION This study indicates that concentrations of NT–pro- BNP are increased across the entire spectrum of clinical CAD. Further, data from this study demonstrate, in a consecutive series of patients who had CAD, that the level of NT–pro-BNP parallels the severity of myocar- dial ischemia. On a median basis, patients who had stable angina, those who had unstable angina, and those who had AMI showed increases of 2.8-, 5.6-, and 8.9-fold in preserved, even after adjusting the multivariable model for LVEF and other cardiovascular risk factors or clinical variables. A positive association between increased lev- els of NT–pro-BNP and severity of angiographic CAD (number of coronary arteries narrowed) was also ob- served, although the strength of this association was weakened by adjustment in the multivariable analysis. With regard to patients who had stable or unstable an- gina, although the difference in LVEF was only 2.5% on a median basis (58% vs 55.5%), the increase in NT–pro- BNP concentration was 2 times higher in patients who had unstable angina than in those who had stable angina, thus stressing that myocardial ischemia that has no evi- dence of myocardial necrosis is a potent stimulus for increased NT–pro-BNP. FIGURE 2. Median values of NT–pro-BNP in controls and patients who had stable angina, those who had unstable angina, and those who had AMI. Previous small-scale studies have demonstrated that myocardial ischemia provoked by ex- ercise testing21 or balloon inflation during coronary angioplasty22 in- creases circulating levels of natriuretic peptides. Plasma BNP levels at peak exercise correlated closely with the size and degree of myocardial ischemia as assessed by thallium radionu- clide imaging.5 Cross-sectional studies have demonstrated statistically significant increases in BNP and NT–pro- BNP levels in patients who have un- stable angina and no evidence of myocardial necrosis compared with patients who had stable angina or con- trol subjects.8,9 Increased levels of BNP or NT–pro-BNP have been con- sistently reported in patients who have ST-elevation myocardial infarction10 –12 and non–ST-elevation acute coronary syn- dromes.13–16 Mechanisms of increased production of na- triuretic peptides in myocardial ischemia are unclear. One mechanism may involve ischemia-induced increases in ventricular wall stress that may stimulate ven- tricular secretion of natriuretic peptides through a variety of molecular mechanisms.23,24 Experimental studies in a rat model of AMI have demonstrated ventricular gene expression and increased tissue concentrations of BNP in infarcted and noninfarcted regions.25 Increased BNP gene expression in the ischemic human left ventricle has also been demonstrated.26 A recent study in adult pa- tients who had cyanotic congenital heart disease has demonstrated that hypoxia is a direct stimulus for BNP release from human cardiac myocytes.27 All these stud- ies credit the theory that myocardial ischemia per se induces production and release of BNP and NT–pro- BNP. FIGURE 3. Correlation between natural-log (Ln) NT–pro-BNP and LVEF in patients who had stable angina (left) and those who had AMI (right). Because NT–pro-BNP concen- trations were skewed, logarithmic transformation was applied. Another finding of this study concerns the correlation between left ventricular function and plasma con- centrations of NT–pro-BNP in patients who have dis- tinct entities of clinical CAD. In patients who had stable angina (low-grade myocardial ischemia), the correlation between LVEF and plasma level of NT– pro-BNP was significantly stronger than in patients who had unstable angina or AMI (more severe myo- cardial ischemia). This finding may be of particular importance because it may explain the strength of prognostic information provided by NT–pro-BNP measurements compared with LVEF. It has recently been reported that NT–pro-BNP is a stronger predictor of mortality than LVEF in patients who have acute coronary syndromes.28 Based on our findings, we hy- pothesize that myocardial ischemia, by interfering with left ventricular function and production and re- lease of natriuretic peptide, weakens the correlation between LVEF and plasma concentrations of NT–pro- BNP. On the one hand, myocardial ischemia increases the production/release of NT–pro-BNP independently of left ventricular function; on the other, it may pro- vide or put into operation factors such as myocardial stunning or regional hyperkinesia that as a conse- quence may blunt the predictive accuracy of LVEF. This study has shown in a large series of consec- utive patients who had stable, unstable angina, and AMI that plasma NT–pro-BNP concentrations in- creased progressively with the increase in severity of CAD. Patients who had acute coronary syndromes (unstable angina and AMI) showed a weaker correla- tion between plasma concentrations of NT–pro-BNP and LVEF than did patients who had stable angina. These findings may explain the predictive strength of ML162 NT–pro-BNP measurements in patients who have acute coronary syndromes.