INTRODUCTION

Heart failure is a common cause of hospitalization and death in the developed world. Improving the diagnosis and care of patients with this syndrome could have a great impact on its associated mortality and health costs. Against this background, knowing the plasma concentrations of certain natriuretic peptides might be very useful in the diagnosis of heart failure.1

The endocrine function of the heart was confirmed over 20 years ago with the discovery of atrial natriuretic peptide (ANP). This eventually led to the description of a family of peptides involved in cardiovascular homeostasis (the natriuretic peptide family) whose members are of similar structure but different genetic origin. These peptides promote natriuresis and diuresis, act as vasodilators and have antimitogenic effects on cardiovascular tissues.2 Two members of this family, ANP and brain natriuretic peptide (BNP), are secreted by the heart mainly in response to the stretching of the myocardium induced by volume overload or in response to hypertrophy.3 The endocrinologically active C-terminal peptides--ANP and BNP--and their N-terminal prohormone fragments can be found in the plasma.

Natriuretic peptides are also useful as independent markers in risk stratification. Increased blood levels of these peptides have been recorded in congestive heart disease and acute myocardial infarction (AMI), suggesting they play a role in the pathophysiology of these conditions.4-6

High blood pressure affects more than half of elderly people, and prevalence continues to increase with age.7 High blood pressure is associated with an increased risk of developing congestive heart disease, and patients with hypertrophy of the left ventricle or chronic heart failure have a much greater risk of dying.8 High blood pressure also increases the risk of suffering a cerebrovascular accident, coronary artery disease, terminal renal disease and death.7 The majority of people know it is important to monitor their blood pressure, but few do so. The morbidity and mortality associated with the problem therefore remain high. The recommendations of the Seventh Report of the Joint National Committee for the Prevention, Detection, Evaluation and Treatment of High Blood Pressure, and of the WHO-International Society of Hypertension,9 provide the most recent guidelines to family doctors for the control of high blood pressure, as well as information on its prevention, detection, assessment and treatment. Similar information is provided regarding its associated morbidity.

Increased plasma concentrations of N-terminal brain natriuretic peptide (NT-proBNP) are associated with increased wall stress, and high blood pressure is a common cause of the latter in the left ventricle (LV).10 The present study examines plasma concentrations of NT-proBNP in patients with and without a diagnosis of heart failure, in an attempt to evaluate the influence of high blood pressure. The study population was selected from patients with dyspnea, given their high probability of suffering heart failure, in order to evaluate the use of NT-proBNP in the diagnosis of this syndrome in the general population.

PATIENTS AND METHODS

Population Studied

Patients were recruited from the PANES11 study database (Prevalencia de Angina de Pecho en España--The Prevalence of Angina in Spain study). Those selected--all with dyspnea--came from the Valencia Community (eastern Spain: Provinces of Castellón, Valencia, and Alicante). A total of 999 patients in the database lived in this region, of whom 432 declared suffering some degree of dyspnea (after receiving a detailed verbal explanation of the symptoms). All 432 patients were originally referred to their respective hospitals (a total of 10 were involved in the study), of these 215 agreed to enter the study. All underwent blood extraction, an echo-Doppler study, and completed a specific questionnaire that asked whether they had ever been diagnosed with high blood pressure9 and what type of medication they had been prescribed. The blood pressure of every patient was also taken (for use in later calculations); those who met the diagnostic criteria for high blood pressure were considered hypertensive.

The study protocol was approved by the appropriate revision boards or the ethics committees of each center. The study was performed in accordance with the clinical practice guidelines and ethical standards regarding experiments on human beings established by the Declaration of Helsinki. All patients gave their written, informed consent to take part. At the end of the study, complete data were available for 202 patients.

The main cardiological diseases diagnosed were ischemic heart disease (n=31), valve disease (n=7), dilated cardiomyopathy (n=3), and congenital heart disease (n=1). Seventy two patients were diagnosed with hypertension and were treated (58% with angiotensin converting enzyme inhibitors, 42% with calcium antagonists, 32% with diuretics, and 11% with beta-blockers). Of these patients, 63.7% received monotherapy, 33.4% received 2 medications, and 2.9% received 3. In addition, the functional class of all patients was determined according to the criteria of the New York Heart Association (NYHA) (Table 1). Thirty one patients were found to have diabetes mellitus, and 37 were diagnosed with other diseases. Of the 202 patients who completed the study, 73 were obese (body mass index >30 kg/m2).

Analysis of NT-proBNP Levels

Blood samples (in EDTA) were extracted by venipuncture after patients had lain in a supine position for at least 30 min. Samples were separated into their component fractions by centrifugation in Eppendorf tubes before transport to the laboratory for analysis. The concentration of NT-proBNP in EDTA-plasma was determined in duplicate (and blind) by in vitro immunoenzymatic analysis using an ELISA kit (Roche Diagnostics);12 the results are expressed in pg/mL (measurement range 0-2069 pg/mL).

Echocardiographic Study

This part of the study was performed using a number of standard echocardiography systems (those available at the 10 hospitals taking part). All were equipped with 2.5 MHz transducers. The images and the Doppler traces were recorded on video tape for centralized analysis (blind to the results of other tests).

All two-dimensional images, the Doppler spectrum and color Doppler, were analyzed using a computerized system (EcoDat; Software Medicina S.A.). Mean values for each Doppler variable were calculated for 4 cardiac cycles in each patient.

The mitral valve flow propagation velocity (Vp) was determined as previously described,13 using the value of 45 cm/s as a cut-off point for the diagnosis of diastolic dysfunction. The A and E waves of the transmitral flow-rate were measured by pulsed Doppler , and the E/A ratio was calculated. The area-length method was used to calculate the ejection fraction (EF) (100x([end diastolic volume-end systolic volume]/end diastolic volume). The cut-off point for systolic dysfunction was taken as 50%.14 Left ventricular mass was determined as previously described.15

The intraobserver and interobserver variability for Vp was 7.2±8.0% and 8.0±8.0% respectively (the absolute difference divided by the mean value of all measurements).

Statistical Analysis

Results are expressed as means ± standard deviation (SD), except those for NT-proBNP (non-symmetrical distribution) which are expressed as the median plus the range. Two groups of patients were distinguished: those with and those without a diagnosis of high blood pressure. The Student t test for independent samples was used to compare continuous, normally distributed variables. The Fisher exact test was used to compare qualitative independent variables. The Mann-Whitney U test was used to compare the NT-proBNP values of the hypertensive and non-hypertensive patients.

When patients with systolic--and later, diastolic--dysfunction were excluded, the Mann-Whitney U test was again used to compare the NT-proBNP levels of the hypertensive and non-hypertensive patients.

All statistical calculations were performed using the Statistical Package for the Social Sciences (SPSS/PC) v10.1, (SPSS Inc. Chicago, Illinois). Significance was set at P<.05.

RESULTS

For the population as a whole, the mean values for NT-proBNP, Vp, EF and the LV mass index were 88 (0-2586) pg/mL, 60±19 cm/s, 63±8%, and 112±47 g/m2 respectively.

The distribution of patients with a diagnosis of some heart problem across the hypertensive and non-hypertensive groups was: ischemic heart disease--hypertensive 48%, non-hypertensive 52%--, valve disease --hypertensive 43%, non-hypertensive 57%--, dilated cardiomyopathy--hypertensive 33%, non-hypertensive 67%--; 1 patient with high blood pressure was diagnosed with congenital heart disease. According to the clinical and echocardiographic studies, 50% of the patients had dyspnea of non-cardiac origin.

Patients with a diagnosis of high blood pressure (n=72, Vp=58±19 cm/s, EF=64±9%, LV mass index=127±45 g/m2) were found to have higher NT-proBNP levels than those with normal blood pressure (n=130, Vp=61±9 cm/s, EF=63±8%, LV mass index=105±47 g/m2): 123 [0-2184] pg/mL compared to 77 [0-2586] pg/mL; P<.01; Table 1).

When patients with left ventricular systolic dysfunction were excluded (EF<50%) (see Table 2 for the echocardiographic traces of patients with EF>50%), and the hypertensive and non-hypertensive patients compared (hypertensive patients: n=65, Vp=59±19 cm/s; non-hypertensive patients n=114, Vp=63±19 cm/s), those with high blood pressure were again found to have higher NT-proBNP levels (119 [0-2184] compared to 72 [0-997] pg/mL in non-hypertensive patients; P<.01). When patients with signs of left ventricular diastolic dysfunction (Vp<45 cm/s) were also excluded from the analysis, the NT-proBNP level of the 44 hypertensive patients remaining was 85 (0-430) pg/mL compared to 66 (0-997) pg/mL in the 86 remaining non-hypertensive patients (P=NS).

The comparison of NT-proBNP levels between hypertensive and non-hypertensive patients with respect to etiological group showed that those with no cardiological diagnosis but with high blood pressure (n=52) had higher plasma NT-proBNP levels than those with normal blood pressure (n=108) (34 [0-568] pg/mL compared to 19 [0-289] pg/mL; P<.01). When patients with systolic dysfunction were excluded, those with high blood pressure (n=48) had higher NT-proBNP levels than did patients with normal blood pressure (n=99) (33 [0-586] pg/mL compared to 19 [0-289] pg/mL; P<.01). When patients with a Vp of <45 cm/s were also excluded, the difference disappeared.

In patients with valve disease (n=7), hypertensive patients were found to have a mean NT-proBNP level of 228 (125-316) pg/m, whereas those with normal blood pressure had 50 (28-84) pg/mL (P<.05). When patients with an EF<50% were excluded, those with high blood pressure were found to have higher NT-proBNP levels (228 [125-316] compared to 49 [28-55] pg/mL in non-hypertensive patients; P<.05). When patients with both systolic and diastolic dysfunctions were removed from the analysis, no significant differences were found between hypertensive and non-hypertensive patients.

In patients with ischemic heart disease (n=31), no significant differences were seen in NT-proBNP levels between those who were hypertensive and those who were not.

The small number of patients with dilated cardiomyopathy (n=3) meant no such comparisons could be made.

DISCUSSION

The natriuretic peptide system becomes most active when there is ventricular dysfunction. Recently it has been confirmed that, of all the neurohormones studied, the cardiac natriuretic peptides are the best predictors of morbidity and mortality in AMI patients (when measured in the subacute phase), and in patients with chronic heart failure.2,16 These peptides have an excellent negative predictive power, particularly in high risk patients. An increase in BNP values is sufficient to justify further tests.2 In addition, the natriuretic peptides are useful in the control of therapy and for monitoring the course of disease in patients with heart failure. They can also be employed in risk stratification with respect to heart failure and myocardial infarction.2,17 High plasma levels of NT-proBNP are also associated with increased wall stress (which is directly related to ventricular pressure) --and high blood pressure is a common cause of increased left ventricular wall stress.10 High blood pressure affects more than half of all elderly people and its prevalence continues to increase with age.7 Since heart failure and hypertension frequently co-exist, this study examined the plasma concentrations of NT-proBNP in a population with dyspnea; those with and without a diagnosis of high blood pressure were compared given the clinical and epidemiological importance of any potential findings.

The results show that plasma NT-proBNP levels are higher in patients with high blood pressure compared to those with no such diagnosis. The distribution of patients with heart disease between the hypertensive and non-hypertensive patient groups was similar. Since systolic dysfunction is associated with increased NT-proBNP levels,18 these patients were excluded in the second phase of analysis, and it was found that hypertensive patients still showed high levels of the peptide. In a third step, those patients with diastolic dysfunction (another potential cause of high NT-proBNP levels--see Lubian et al19) were excluded from the analysis.13 Lubian et al19 found increased NT-proBNP levels in patients with diastolic dysfunction but normal systolic function, but differences between hypertensive and non-hypertensive patients disappeared. This shows that, in the present hypertensive patients, the levels of the peptide increased because of changes in diastolic function. In other words, NT-proBNP levels in patients with dyspnea and high blood pressure probably increase because of diastolic dysfunction.20,21

Many Doppler indices are used to describe diastolic dysfunction, but the present work took a Vp of 45 cm/s as the cut-off point for diastolic dysfunction (as described by García et al13). Unlike other indices, the Vp is a well-known Doppler variable associated with left ventricular diastolic dysfunction.22 Furthermore, the Vp reflects changes in relaxation, distensibility and rigidity, and its fluctuations are simultaneous with changes in other variables (Table 2). In addition, in the present work, Vp results were found to be very reproducible.

With respect to EF, the 50% level was used as a cut-off point for systolic dysfunction: the literature shows a value of <50% is normally used in populational studies.14 Although the mean EF recorded was acceptable, it should be remembered that only about 50% of patients had dyspnea of cardiac origin, and it was not possible to confirm the declarations of the remaining patients when specific tests (NYHA) were performed. Furthermore, although in clinical practice EF is used as an absolute value (despite the fact that in the diagnosis of left ventricular systolic dysfunction a cut-off point has to be selected), there is no doubt that it is a relative value for each heart and that small reductions in a patient's EF (compared to earlier readings) can be interpreted as systolic dysfunction. In a general population of elderly patients (such as those of the present study), changes in systolic and diastolic function can both occur. This might lead to changes in the severity of symptoms without this being reflected in resting EF readings.23

The equipment used to record the NT-proBNP levels was first generation, and the results showed a wide range from 0-2586 pg/mL. The lower limit of detection may not have been sufficiently sensitive--thus the readings of 0 pg/mL. Nonetheless, these results show that the values are in coherence with the well validated Doppler indices and the statistical analysis supports the conclusions inferred from the results. The values of a range that superimpose one group on another in the general population would not exist when dealing with an individual patient.

For non-hypertensive patients with an EF>50 and a Vp of >45 cm/s, the NT-proBNP values ranged from 0-997 pg/mL--a higher value than might be expected. This could be due to mitral or aortic regurgitation in some patients who still have good ventricular function, but who show increased left ventricular volume.24,25

Ideally, the statistical analysis used would have been two-way ANOVA, in which the effect of the presence or absence of high blood pressure on the appearance of systolic or diastolic dysfunction would have been examined.26 In optimum conditions, the results obtained would have been more powerful. However, since the distribution was not normal, two-way ANOVA was impossible, and the procedure described above was selected.

In conclusion, this work shows that, in a general population, hypertensive subjects have higher NT-proBNP levels than those with normal blood pressure. This difference is maintained when patients with systolic dysfunction are removed from the analysis. However, when those with signs of both systolic and diastolic dysfunction are removed, no significant differences are seen between hypertensive and non-hypertensive patients. The fact that NT-proBNP levels were higher in the hypertensive population with dyspnea (largely due to changes in diastolic function) could introduce confusion that might reduce the specificity of NT-proBNP in the diagnosis of heart failure. These findings should be taken into consideration in clinical and epidemiological studies in which there are patients with heart failure and high blood pressure. The present study could serve as a basis for the use of NT-proBNP levels in the control of treatment and monitoring of patients with high blood pressure.

Full English text available at: www.revespcardiol.org

ABBREVIATIONS
NT-proBNP: N-terminal pro-brain natriuretic peptide.
PANES: the Prevalence of Angina in Spain study (Estudio Prevalencia de Angina de Pecho en España).
LV: left ventricle.
Vp: mitral valve flow propagation velocity.
EF: ejection fraction.
E/A: E to A wave ratio of transmitral flow-rate.

This project was financed by the Fondo de Investigaciones Sanitarias del Instituto de Salud Carlos III, Project FIS 01/0943.

Correspondence: Dr. M. Rivera Otero.
José María Haro, 59, puerta 59. 46022 Valencia. España.
E-mail: rivera_jmi@gva.es