Natriuretic Peptides Predict Symptom-Free Survival and Postoperative Outcome in Severe Aortic Stenosis
Background— The prognostic value of natriuretic peptides in aortic stenosis (AS) remains unknown.
Methods and Results— B-type natriuretic peptide (BNP), N-terminal BNP (NtBNP), and N-terminal atrial natriuretic peptide (NtANP) were determined in 130 patients with severe AS (mean age, 70±12 years; mean gradient, 64±21 mm Hg; valve area, 0.64±0.15 cm2) who were followed up for 377±150 days. Natriuretic peptides increased with NYHA class and with decreasing ejection fraction (EF). Even asymptomatic patients frequently had elevated neurohormones. Asymptomatic patients who developed symptoms during follow-up had higher BNP and NtBNP levels at entry compared with those remaining asymptomatic (median for NtBNP, 131 pmol/L [interquartile range, 50 to 202 pmol/L] versus 31 pmol/L [range, 19 to 56 pmol/L]; P<0.001). Symptom-free survival at 3, 6, 9, and 12 months for patients with NtBNP <80 versus ≥80 pmol/L was 100%, 88±7%, 88±7%, and 69±13% compared with 92±8%, 58±14%, 35±15%, and 18±15%, respectively (P<0.001). Seventy-nine patients eventually underwent surgery because of symptoms. Considering preoperative neurohormone levels, age, NYHA class, aortic valve area, EF, and presence of coronary artery disease, we found that neurohormones, EF, and NYHA class predicted survival; neurohormones predicted postoperative symptomatic status; and neurohormones and preoperative EF predicted postoperative EF. However, by multivariate analysis, NtBNP was the only independent predictor of outcome.
Conclusions— In severe AS, natriuretic peptides provide important prognostic information beyond clinical and echocardiographic evaluation. NtBNP independently predicts symptom-free survival, and preoperative NtBNP independently predicts postoperative outcome with regard to survival, symptomatic status, and left ventricular function. Thus, neurohormones may gain particular importance for timing of surgery in asymptomatic severe AS.
Received October 17, 2003; revision received February 9, 2004; accepted February 10, 2004.
Although there is general agreement that symptomatic aortic stenosis (AS) requires urgent valve replacement,1–4 the timing of surgery in asymptomatic severe AS remains a matter of controversy.5,6 The observation of a relatively benign outcome in these patients7,8 led to the recommendation of delaying surgery until symptoms occur.3,5 Nevertheless, a number of concerns remain with regard to this “wait for symptoms” strategy. Patients frequently do not immediately report newly developed symptoms; patients eventually symptomatic are at significant risk while waiting for surgery; and operative risk increases with symptom severity.6 On the other hand, operative risk and prosthetic valve–related long-term morbidity and mortality do not justify elective surgery on every asymptomatic patient with severe AS.9,10 Thus, predictors of outcome that help to identify those patients who benefit from early elective surgery are needed.
Natriuretic peptides have been shown to aid in the recognition of symptoms of congestive heart failure and discrimination between cardiac and noncardiac dyspnea.11,12 They have also been reported to be independent predictors of outcome in congestive heart failure,13–16 primary pulmonary hypertension,17 acute myocardial infarction, and pulmonary embolism.18,19 In AS, recent studies have demonstrated that plasma levels of natriuretic peptides are related to disease severity20–23 and to symptomatic status.24,25 However, their prognostic value and their potential role for timing of surgery, particularly in asymptomatic patients with severe AS, remains unknown.24 Therefore, we tested 2 hypotheses: (1) Plasma levels of natriuretic peptides predict the symptom-free survival in patients with severe AS, and (2) preoperative plasma levels of natriuretic peptides predict postoperative outcome with regard to mortality and morbidity in AS.
A total of 130 consecutive patients referred to our valvular heart disease clinic with severe AS (peak velocity >4 m/s and/or aortic valve area <1.0 cm2) were enrolled in this prospective study. Exclusion criteria included plasma creatinine level >2.5 mg/dL, myocardial infarction within 6 months, more than mild mitral valve disease or aortic regurgitation, and severe pulmonary disease.
The symptomatic status was assessed by experienced cardiologists blinded to the results of neurohormone measurements. The extent of dyspnea was classified with the NYHA classification. Because many patients were expected to be elderly and bicycle exercise testing was not feasible for many of them, this test was not included in the study protocol. However, as is the practice in our clinic, patients in whom classification remained uncertain from history were taken to the staircase and walked upstairs under observation. Patients classified as asymptomatic had to be free of shortness of breath, angina, dizziness, and syncope with exertion.
Patient characteristics are given in Table 1. Eighty-seven patients were classified as symptomatic at study entry. Compared with the 43 asymptomatic patients, they did not differ in age and presence of hypertension but were more likely to be female and had higher gradients, smaller valve area, more frequent coronary artery disease, and lower ejection fraction (EF). None of the asymptomatic patients had reduced left ventricular (LV) function. The presence of coronary artery disease was defined as history of myocardial infarction or significant stenosis in a previous coronary angiogram.
The University of Vienna Medical School Ethics Committee approved the study protocol, and all patients gave written informed consent.
Commercially available ultrasound systems (Hewlett Packard 5500 Sonos, Philipps Inc; Accuson Sequoia, Siemens Inc; Vingmed Vivid 7 and Vivid 5, General Electric Inc) were used. All patients underwent a comprehensive examination, including M-mode and 2D echocardiography and continuous-wave, pulsed-wave, and color Doppler. Mean aortic valve gradient was calculated by averaging the instantaneous gradients throughout systole with the on-board quantification package. Aortic valve area was calculated by the continuity equation; EF, with the biplane Simpson method.
Measurement of Neurohormone Levels
Venous blood samples were drawn at each visit from an antecubital vein into chilled EDTA acid Vacutainer test tubes after 30 minutes of supine rest. Samples were placed immediately on ice, and plasma separation was performed at −4°C. Plasma samples were frozen at −70°C until assay. A fluorescence immunoassay was performed for quantification of B-type natriuretic peptide (BNP) levels (Triage BNP Test, Biosite Diagnostics, Inc). For N-terminal BNP 1 to 76 (NtBNP) determination, an electrochemiluminescence immunoassay (ProBNP Elecsys, Roche Diagnostics GmbH) was used. N-terminal atrial natriuretic peptide 1 to 98 (NtANP) levels were measured with an ELISA (Biomedica, GmbH).
Baseline clinical and echocardiographic data and neurohormone levels were assessed at the patient’s first visit (study entry). Surgery was considered to be indicated when patients presented with symptoms or with reduced LV function while asymptomatic. Patients were reevaluated (clinical and echocardiographic data) between 3 and 6 months after valve replacement. Asymptomatic patients were instructed to report the development of symptoms immediately. As long as they remained free of symptoms and LV dysfunction, patients were reevaluated in 6-month intervals. Patients presenting before their regular visit because of symptoms were reevaluated at that time (clinical, echocardiographic, and neurohormone data), referred for surgery, and reevaluated 3 to 6 months thereafter.
Symptom-free survival was carefully documented for all patients entering the study without symptoms to test hypothesis 1. All patients symptomatic at study entry or during follow-up who were therefore referred for surgery formed the study population evaluated to test hypothesis 2.
Continuous variables are expressed as mean±SD or median with interquartile range when appropriate. Differences between patient groups were analyzed with ANOVA and 2-sample t test for continuous variables or χ2 test for categorical variables. A value of P<0.05 was considered significant. Symptom-free survival was analyzed by the Kaplan-Meier method for patient groups with BNP <130 or ≥130 pg/mL, NtBNP <80 or ≥80 pmol/L, and NtANP <5000 or ≥5000 fmol/mL. Cutoff values were selected according to previous studies of other patient populations.26 Differences between groups were analyzed with the log-rank test. Cox regression analysis was performed to identify predictors of postoperative survival. The model was built stepwise; the value for entering and staying in the model was set at P=0.05. For prediction of postoperative ventricular function (EF <0.50 or ≥0.50) and functional status (NYHA class ≤II or >II), stepwise logistic regression analysis was performed.
For all analyses, commercially available statistical package software was used (SPSS version 11.5 and StatView SAS version 5.0.1).
Neurohormones at Study Entry
At baseline, plasma levels of all natriuretic peptides increased with NYHA class (Figure 1A through 1C). Differences between classes I and II did not reach statistical significance, but differences between classes II and III and between classes III and III-IV were significant for all 3 natriuretic peptides (P<0.0001). BNP, NtBNP, and NtANP correlated inversely with EF (r=0.51, P<0.0001; r=0.57, P<0.0001; and r=0.47, P<0.0001, respectively). Even asymptomatic patients frequently had markedly elevated neurohormones (BNP >100 pg/mL in 47%, NtBNP >50 pmol/L in 47%, and NtANP >1945 fmol/mL in 88% of patients). The presence of coronary artery disease did not influence neurohormone levels. Only 5 of the 87 symptomatic patients presented without dyspnea. They had relatively low neurohormone levels (median BNP, 77 pg/mL [interquartile range, 61 to 108 pg/mL]; NtBNP, 54 pmol/L [interquartile range, 33 to 69 pmol/L]; NtANP, 2487 fmol/mL [interquartile range, 1948 to 3037 fmol/mL]).
Patients were followed up for 377±150 days. Patient outcome is summarized in Figure 2. Because no patient had or developed LV dysfunction while asymptomatic, all patients who underwent surgery had symptoms.
At study entry, 87 patients were symptomatic and 43 patients were asymptomatic. Of these 43, 14 developed symptoms after 194±104 days. Of the 101 total symptomatic patients, 7 refused surgery despite recommendation, and 12 were not referred for surgery because of significant comorbidity and high age. Of these 19 symptomatic patients not undergoing surgery, 9 eventually died (congestive heart failure, 8; sudden death, 1). One additional patient died while on the waiting list, and 79 eventually underwent surgery (2 patients were still waiting for surgery when the study was terminated and the data were analyzed). There were 6 perioperative deaths (bleeding, 2; sepsis, 2; respiratory complications, 2). Two patients died 50 and 63 days after surgery (cause of death, congestive heart failure in both). Of the surviving patients, 21 were in NYHA class I, 44 were in class II, and 6 were in class III; 63 patients had a normal EF. No patient died while asymptomatic.
Natriuretic Peptides and Development of Symptoms
Patients developing symptoms during follow-up had significantly higher baseline BNP and NtBNP levels (Table 2). Groups were best separated by NtBNP levels. At follow-up, neurohormones increased further in patients who developed symptoms but did not change in those remaining asymptomatic (Table 2). Six initially asymptomatic patients who developed acute congestive heart failure during follow-up had particularly high neurohormones at study entry (BNP, 502±273 pg/mL; NtBNP, 265±159 pmol/L). In contrast, 2 patients who developed angina without dyspnea had low BNP (44 and 48 pg/mL) and NtBNP (16 and 62 pmol/L).
By Kaplan-Meier analysis (Figure 3A through 3C), natriuretic peptides were significant predictors of symptom-free survival, which was 100% at 3 months, 90±7% at 6 months, 90±7% at 9 months, and 66±16% at 12 months for 25 patients with BNP <130 pg/mL compared with 94±5%, 64±12%, 45±14%, and 34±14%, respectively, for 18 patients with BNP ≥130 pg/mL (P<0.05; Figure 3A). Symptom-free survival was 100% at 3 months, 88±7% at 6 months, 88±7% at 9 months, and 69±13% at 12 months for 31 patients with NtBNP <80 pmol/L compared with 92±8%, 58±14%, 35±15%, and 18±15% for 12 patients with NtBNP ≥80 pmol/L (P<0.001; Figure 3B). For 31 patients with NtANP <5000 fmol/mL, symptom-free survival was 97±3% at 3 months, 82±8% at 6 months, 82±8% at 9 months, and 58±13% at 12 months compared with 100%, 68±16%, 36±19%, and 36±19% for 12 patients with NtANP ≥5000 fmol/mL (P<0.05; Figure 3C).
Compared with patients remaining asymptomatic, those who developed symptoms did not differ significantly with regard to age, peak and mean gradients, and presence of coronary artery disease. They had slightly smaller valve areas (0.65±0.13 versus 0.76±0.16 cm2, P<0.05) and slightly lower EF (0.63±.7 versus 0.68±.6, P<0.05). By multivariate analysis, however, only NtBNP (P<0.05) and EF (P<0.05) were independent predictors of remaining free of symptoms during follow-up.
Neurohormones and Postoperative Outcome
By univariate analysis, preoperative BNP (P<0.05), NtBNP (P<0.001), and NtANP (P<0.01) and preoperative NYHA class (P<0.01) and EF (P<0.01) were significant predictors of survival, whereas age, presence of coronary artery disease, and aortic valve area were not. By multivariate analysis, however, NtBNP (P<0.001) remained the only independent predictor of survival (Table 3).
Preoperative BNP (P<0.05) and NtBNP (P<0.01) were significant predictors of postoperative symptomatic status (NYHA class ≤II versus >II) by univariate analysis, but NtBNP (P<0.05) remained the only independent predictor by multivariate analysis (Table 3).
By univariate analysis, preoperative BNP (P<0.05), NtBNP (P<0.01), and preoperative EF (P<0.01) were significant predictors of normal postoperative EF (EF <50% versus ≥50%). Both preoperative EF (P<0.05) and NtBNP (P<0.05) remained significant predictors of normal postoperative EF by multivariate analysis (Table 3).
Timing of Surgery in Severe AS: The Dilemma in Asymptomatic Patients
Outcome of hemodynamically severe but asymptomatic AS has been reported to be rather favorable.7,8 In particular, sudden death appears to be uncommon as long as the patient remains free of symptoms.2,8 Thus, so far, the only class I indication for valve replacement in AS has been the development of symptoms.3,5 Nevertheless, many physicians feel uncomfortable about delaying surgery until symptoms occur for several reasons. First, milder symptoms of dyspnea and fatigue are notoriously difficult to assess, particularly in the elderly. This patient group may also unconsciously avoid exercise because of early symptoms and then deny symptoms when interrogated. Second, patients frequently do not immediately present to their physicians when symptoms develop. Third, symptomatic patients are at significant risk while awaiting surgery. Fourth, severe myocardial hypertrophy and fibrosis may preclude a favorable long-term outcome after eventual valve replacement. Fifth, operative risk increases with symptom severity. Finally, although these concerns would favor elective surgery in severe asymptomatic AS, operative risk, particularly in elderly patients,27,28 and prosthetic valve–related long-term morbidity and mortality9,10 are strong arguments against simply operating on every asymptomatic patient with severe AS. Thus, recent research has focused on the identification of predictors of outcome that help to select those patients who benefit from early elective surgery.
Echocardiographic assessment of the extent of valve calcification and hemodynamic progression2 and exercise testing8,29 have been reported to be helpful in this respect. However, additional variables would definitely be desirable to improve risk stratification.
Natriuretic Peptides in AS
BNP and NtBNP have been reported to correlate with the transvalvular gradient,21,22 LV end-systolic wall stress,20 LV end-diastolic pressure,21 and extent of LV hypertrophy23 in AS, suggesting that their plasma levels increase with disease severity and that they could potentially be used to monitor disease progression. However, disease severity also can easily be assessed and monitored by echocardiography, and it remained unclear from these studies whether neurohormones could indeed improve timing of surgery.
Gerber et al24 have recently reported that increased plasma natriuretic peptide levels reflect symptom onset in AS and that their measurement may complement clinical and echocardiographic evaluation. Considering the difficulties in deciding whether patients should be classified symptomatic or not, primarily in the case of nonspecific findings such as mild dyspnea or fatigue in the elderly, these results appeared promising. Although our results confirm that natriuretic peptide levels increase with NYHA class, they nevertheless differ in an important aspect. Gerber at al found natriuretic peptide levels to be significantly higher in patients with NYHA class II than in those with class I, suggesting that they can be used to discriminate between early symptoms of heart failure and normal exercise tolerance.24 In contrast, we found only a marked increase in natriuretic peptide levels from class II to III and from class III to III-IV, although there was wide overlap between classes I and II. This may be due to the differences in patient population. Gerber et al included patients with transvalvular peak velocity >2.5 m/s; therefore, many of the asymptomatic patients had nonsignificant AS. In contrast, only patients with severe AS were included in the present study. In fact, approximately half the patients with asymptomatic severe AS had significantly elevated BNP and NtBNP, and even 88% had markedly elevated NtANP levels. Thus, natriuretic peptide levels may be less helpful for discriminating between asymptomatic and mildly symptomatic patients than assumed in this previous study.
Natriuretic Peptides as Predictors of Outcome in AS
Even more important, however, is the fact that natriuretic peptide levels turned out to be significant predictors of outcome in the present study. In the group of asymptomatic patients with severe stenosis and normal LV function, they predicted symptom-free survival, whereas other variables such as age, transvalvular gradient, and presence of coronary artery disease did not. Asymptomatic patients with severe AS who had plasma BNP levels <130 pg/L and plasma NtBNP levels <80 pmol/L were unlikely to develop symptoms and require surgery over the following 6 to 9 months (90% and 88% symptom-free survival). On the other hand, those patients with plasma natriuretic peptide levels above these previously defined cutoff values had a high chance (45% and 35%) of becoming symptomatic and requiring valve replacement during this short time period. Thus, natriuretic peptide levels measured every 6 to 9 months may markedly improve our ability to select mildly symptomatic or asymptomatic patients who benefit from early elective surgery.
In addition, in those patients who underwent surgery in the present study, NtBNP independently predicted postoperative outcome with regard to survival, functional class, and LV function. This finding again supports the idea that an increase in neurohormone levels may define a subgroup of asymptomatic patients who benefit from surgery.
Although our study does not allow definite conclusions as to which neurohormone is the best predictor of outcome, NtBNP appeared to be the most powerful with this regard, whereas NtANP may be of less value.
Despite the large total number of patients in the present study, the subgroups of asymptomatic patients with different neurohormone levels remained relatively small. Although the results have already reached convincing statistical strength, further studies of larger patient groups are necessary to define the most powerful neurohormone or combination of neurohormones and their optimum cutoff values for recommendation of surgery.
As pointed out earlier, the assessment of symptoms in AS remains difficult, particularly as far as mild shortness of breath is concerned, in the elderly. Thus, the discrimination between NYHA classes I and II especially is in general limited.
Exercise testing has been shown to be helpful in distinguishing between asymptomatic and symptomatic patients and has been suggested to offer prognostic information.29 Many patients in the present study were of relatively old age. Therefore, we were not able to perform bicycle exercise tests in a sufficient percentage of patients to allow consideration of this test as an addition to clinical assessment of functional status or to compare its predictive value with that of neurohormones.
In patients with severe AS, plasma levels of natriuretic peptides provide important prognostic information beyond clinical and echocardiographic evaluation. NtBNP predicts symptom-free survival in asymptomatic patients, and preoperative NtBNP independently predicts postoperative outcome with regard to survival, postoperative symptomatic status, and LV function. Thus, neurohormones may gain particular importance for the timing of surgery in asymptomatic patients with severe AS.
Bonow RO, Carabello B, de Leon AC Jr, et al. ACC/AHA guidelines for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Management of Patients with Valvular Heart Disease). J Am Coll Cardiol. 1998; 32: 1486–1588.
Carabello BA. Timing of valve replacement in aortic stenosis: moving closer to perfection. Circulation. 1997; 95: 2241–2243.
Iung B, Gohlke-Barwolf C, Tornos P, et al. Recommendations on the management of the asymptomatic patient with valvular heart disease. Eur Heart J. 2002; 23: 1253–1266.
Rosenhek R, Maurer G, Baumgartner H. Should early elective surgery be performed in patients with severe but asymptomatic aortic stenosis? Eur Heart J. 2002; 23: 1417–1213.
Otto CM, Burwash IG, Legget ME, et al. Prospective study of asymptomatic valvular aortic stenosis: clinical, echocardiographic and exercise predictors of outcome. Circulation. 1997; 95: 2262–2270.
Richard AM, Nicholls MG, Yandle TG, et al. Plasma N-terminal pro-brain natriuretic peptide and adrenomedullin: new neurohormonal predictors of left ventricular function and prognosis after myocardial infarction. Circulation. 1998; 97: 1921–1929.
Tsutamoto T, Wada A, Maeda K, et al. Attenuation of compensation of endogenous cardiac natriuretic peptide system in chronic heart failure: prognostic role of plasma brain natriuretic peptide concentration in patients with chronic symptomatic left ventricular dysfunction. Circulation. 1997; 96: 509–516.
Nagaya NN, Nishikimi T, Uematsu M, et al. Plasma brain natriuretic peptide as a prognostic indicator in patients with primary pulmonary hypertension. Circulation. 2000; 102: 865–870.
Richards AM, Nicholls MG, Espiner EA, et al. B-type natriuretic peptides and ejection fraction for prognosis after myocardial infarction. Circulation. 2003; 107: 2786–2792.
Kucher N, Printzen G, Goldhaber SZ. Prognostic role of brain natriuretic peptide in acute pulmonary embolism. Circulation. 2003; 107: 2545–2547.
Gerber IL, Stewart RAH, Legget ME, et al. Increased plasma natriuretic peptides reflect symptom onset in aortic stenosis. Circulation. 2003; 107: 1884–1890.
Bergler-Klein J, Klaar U, Rosenhek R, et al. Prognostic value of natriuretic peptides in asymptomatic and symptomatic severe aortic stenosis. Circulation. 2002; 106 (suppl II): II-640. Abstract.
Berger R, Huelsmann M, Strecker K, et al. B-type natriuretic peptide predicts sudden death in patients with chronic heart failure. Circulation. 2002; 105: 2392–2397.
Kolh P, Lahaye L, Gerard P, et al. Aortic valve replacement in the octogenarians: perioperative outcome and clinical follow-up. Eur J Cardiothorac Surg. 1999; 16: 68–73.
Amato MC, Moffa PJ, Werner KE, et al. Treatment decision in asymptomatic aortic valve stenosis: role of exercise testing. Heart. 2001; 86: 381–386.