Author + information
- Received August 22, 2007
- Revision received October 18, 2007
- Accepted November 28, 2007
- Published online June 17, 2008.
- Mototsugu Nishii, PhD, MD⁎ (, )
- Takayuki Inomata, PhD, MD,
- Hitoshi Takehana, PhD, MD,
- Takashi Naruke, MD,
- Tomoyoshi Yanagisawa, MD,
- Masahiko Moriguchi, PhD, MD,
- Sadao Takeda, PhD, MD and
- Tohru Izumi, PhD, MD
- ↵⁎Reprints requests and correspondence:
Dr. Mototsugu Nishii, Department of Cardio-angiology, Kitasato University School of Medicine, 1-15-1 Kitasato, Sagamihara, 228-8555 Kanagawa, Japan.
Objectives We investigated the clinical utility of B-type natriuretic peptide (BNP) assay in stable outpatients with nonischemic dilated cardiomyopathy (NICM) after decompensated heart failure (HF).
Background Patients with NICM admitted for decompensated HF frequently experience sudden death or redecompensation after hospital discharge. The prognostic value of BNP during hospitalization has been demonstrated. However, clinical utility of BNP in stable outpatient setting has been poorly investigated.
Methods Eighty-three NICM outpatients who were clinically stable in New York Heart Association functional class 1 to 2 for 6 months after discharge for decompensated HF were enrolled, and then followed for an additional 18 months. The main end point was first readmission for decompensated HF or death. B-type natriuretic peptide levels were measured at 3-month intervals from discharge to enrollment, and echocardiographic dimensions at discharge and enrollment.
Results Mean discharge BNP level was 210 ± 148 pg/ml. Twenty-eight patients were readmitted for decompensated HF or suddenly died at a median time of 11 months from the time of discharge. Among various variables including BNP measurements, clinical parameters and echocardiographic dimensions, a 6-month post-discharge BNP of >190 pg/ml was most closely associated with combined event in the Cox proportional hazards model (hazard ratio 2.29; 95% confidence interval 1.42 to 3.56; p = 0.0005), and had the best discriminatory power (area under the receiver operating characteristic curve 0.91, sensitivity 96%; specificity 76%).
Conclusions Even in stable low-risk outpatients with NICM at 6 months after hospital discharge for decompensated HF, BNP assessment predicts a long-term risk of redecompensation.
Decompensated heart failure (HF) is a critical complication of nonischemic dilated cardiomyopathy (NICM). Outcome in those patients hospitalized for decompensated HF is often poor, and may include readmission for subsequent HF or sudden death (1–3). A simple biomarker as a prognostic predictor, thus, has been sought.
B-type natriuretic peptide (BNP) is secreted from the overloaded left ventricle (LV), and the circulating levels adversely increase in accordance with the degree of LV wall stretch (4,5). Plasma BNP levels have proven utility in many settings, including improving the diagnostic evaluation of acute dyspnea (6,7) as well as the prognostic evaluation of HF patients on both hospital admission and discharge (8–10). However, the clinical utility of BNP assay in stable outpatients after decompensated HF has been poorly investigated. We hypothesized that the prognostic utility of BNP assay would extend to this outpatient setting, albeit with lower cutoff points for high- and low-risk statuses.
Here, we examined the relation of BNP assay with long-term outcome in stable outpatients with NICM after decompensated HF.
Informed consent was obtained from all patients for participation in the study in accordance with the protocol, which was approved by the committee on human investigation at our institution. Consecutive outpatients with NICM who were clinically stable in New York Heart Association (NYHA) functional class 1 to 2 for at least 6 months after hospital discharge for decompensated HF were enrolled.
Admission for decompensated HF was determined by adverse cardiac symptoms (NYHA functional class 3 to 4), physical findings (rales and/or S3 gallop), and evidence of pulmonary congestion or pleural effusion on chest X-ray. Nonischemic dilated cardiomyopathy was diagnosed according to cardiac information during the hospital course including intact coronary arteries on coronary angiography and LV systolic dysfunction (left ventricular ejection fraction [LVEF] of <40%, LV diffuse wall motion abnormality) with a dilated nonhypertrophic LV (left ventricular end-diastolic dimension [LVDd] of >5.5 cm, posterior wall and interventricular septum end-diastolic thickness of ≤1.2 cm) on echocardiography. Echocardiography and cardiac catheterization showed moderate regurgitation of mitral and/or tricuspid valve with the marked dilation of the ventricular cavities in almost all patients, but did not show findings that indicate the existence of primary valvular disease, such as prolapse, structural destruction, extreme calcification, or stenosis.
Discharge criteria were clinically compensated status as follows: NYHA functional class <3, no sign of rales or S3 gallop, stable blood pressure (systolic blood pressure >90 mm Hg), and improvement in renal failure (serum creatinine level <1.5 mg/dl) as well as in chest X-ray findings and HF treatment including beta-blockers, angiotensin-converting enzyme inhibitor (ACEI), and angiotensin receptor blocker (ARB), as recommended by the international guideline (11).
To eliminate the possibility that acute ischemia could precipitate HF decompensation, patients with coronary artery disease were excluded.
Various parameters including BNP level from discharge to enrollment were correlated with outcome. The attending physician was blinded to the BNP results until the end of study.
Outpatient investigations including physical examination, electrocardiogram recording, chest X-ray, and blood sample measurements were conducted every 3 months after discharge. Entry patients were monitored for 18 months after enrollment in the same institutes.
The major end point was sudden death or first readmission for decompensated HF, defined as hospitalization for decompensated HF. The cause of death was determined according to autopsy in all those who died.
Plasma BNP and echocardiographic measurements
Plasma was immediately separated from the blood element by centrifugation at 4°C for the measurement of BNP at discharge, 3 and 6 months after discharge using a specific immunoradiometric assay for human BNP (Shionoria, Osaka, Japan). The minimum detectable quantity of human BNP is 2 pg/ml. The intra-assay and interassay coefficients of variation were 5.2% and 6.1%, respectively.
In accordance with the recommendations of the American Society of Echocardiography, echocardiographic examination was also performed at discharge and 6 months with a Hewlett Packard Sonos 5500 machine (Andover, Massachusetts). Echocardiography was performed by experienced ultrasonographer and repeated by the same operator wherever possible. Left ventricular end-diastolic dimension at the level of the mitral valve leaflet tips and left atrial diastolic dimension (LADd) at the beginning of the QRS complex on the electrocardiography were measured by M-mode echocardiography. Left ventricular ejection fraction was estimated by Simpson's method on 2-dimensional echocardiographs.
All values are expressed as mean ± standard deviation. Differences between event-free patients and combined event patients at baseline were tested by using the Student t test for continuous variables and the chi-square test with continuity equation for categoric variables. Percentage change in BNP level from discharge to subsequent assay in each patient was expressed as follows: (BNP level at 3 or 6 months − BNP level at discharge) × 100/BNP level at discharge. The predictive values of clinical, echocardiographic, and biochemical variables for combined event of all-cause death and readmission for decompensated HF were examined with a Cox proportional hazards model. To compare the predictive values of the parameters, receiver-operating characteristics and their area under the curve were constructed. The beat prognostic cutoff value for a combined event was defined as that which had the best compromise between sensitivity and specificity for predicting readmission or sudden death. Failure curves were generated by the Kaplan-Meier method, and the log-rank test was used to compare the incidence rate of a combined event or readmission alone among the BNP ranges.
The BNP levels at discharge, 3 and 6 months and echocardiographic parameters at discharge and 6 months in readmitted patients or event-free patients were compared by a multivariate approach to repeated measures using the general linear model, to allow correction for the correlation of repeated observations over time. Probability was significant at the value of p < 0.05 level. Statistical analysis was performed with JMP 6.0 software for Windows (SAS Institute Inc., Cary, North Carolina).
Baseline patient characteristics
None of the 85 consecutive NICM patients discharged with a clinically compensated status was readmitted for decompensated HF within the first 6 months. However, only 2 of those suddenly died of a cardiac event. Finally, the remaining 83 outpatients who were clinically stable for 6 months after hospital discharge were enrolled in this study.
Entry patient characteristics are shown in Table 1. They were from age 25 to 84 years (mean age 56 ± 20 years) and clinically stable in NYHA functional class 1 to 2 at least until enrollment. Mean BNP levels were 210 ± 148 pg/ml at discharge, 182 ± 158 pg/ml at 3 months, and 191 ± 163 pg/ml at 6 months. On echocardiography at discharge, all patients had a dilated LV (mean LVDd 6.4 ± 0.8 cm) and low LV systolic function (mean LVEF 31 ± 8%), while mean LVDd and LVEF at entry of 6.2 ± 1.0 cm and 37 ± 10%, respectively. Other conditions identified in patients included atrial fibrillation (36%), ventricular tachycardia (19%), hypertension (11%), or diabetes (9%). Six patients were implanted with a permanent pacemaker for sick sinus syndrome (n = 2), atrial fibrillation (n = 2), or intraventricular dyssyncrony (n = 2), and an implantable cardioverter-defibrillator was inserted in 2 patients during hospitalization, although 15 patients who had a low LVEF of <36% with nonsustained or sustained ventricular tachycardia were advised of the need for prophylactic implantable cardioverter-defibrillator implantation to reduce the risk of sudden cardiac death (12). No patients required a mechanical circulatory assist device for decompensated congestive HF. Beta-blocker together with ACEI or ARB was initiated, and titrated in a clinically compensated status. However, withdrawal of the beta-blocker was performed in 8 patients early after hospital discharge, because of symptomatic hypotension or marked bradycardia (resting heart rate of <40 beats/min). Among 75 patients with continuous beta-blocker use, mean dosages of carvedilol and metoprolol during follow-up were 14 ± 6 mg/day and 91 ± 40 mg/day, respectively. Finally, medical treatment at enrollment consisted of digoxin (18%), vasodilators (9%), beta-blockers (90%), ACEI (84%), ARB (20%), diuretics (33%), and amiodarone (15%).
Twenty-three patients were readmitted for decompensated HF, and 5 died suddenly of cardiac events (n = 3) or noncardiac events (n = 2), with a mean or median time from discharge of 11.5 ± 2.9 or 11 months, range 7 to 18 months.
Comparison of clinical characteristics between subgroups
Clinical characteristics are compared between event-free patients and combined event patients in Table 1. Among medications at enrollment, beta-blocker use and dosage of beta-blocker significantly differed between the 2 groups. Among disease history, atrial fibrillation, ventricular tachycardia, and hypertension showed significant differences. Age, gender, and serum creatinine level at enrollment had no significant difference. For BNP measurement, plasma BNP levels at 3 and 6 months showed significant differences between the 2 groups, whereas level at discharge did not. Among echocardiographic parameters, only LVEF and LADd differed.
Cox proportional hazards regression model for predictors of readmission for HF
Associations among BNP measurements as well as clinical and echocardiographic variables for a combined event of death or readmission for decompensated HF were examined using a Cox proportional hazards regression model (Table 2). Among clinical variables, no predictive variables were identified on univariate analysis. For echocardiography, LADd at 6 months was associated with a combined event, notwithstanding the poor predictability of LVEF. For BNP, plasma BNP level at 6 months and percentage change in level between discharge and 6 months were predictive, whereas levels at discharge and 3 months were not. Area under the curve was higher for BNP at 6 months (0.91) than that for the percentage change in BNP between discharge and 6 months (0.83) and LADd at 6 months (0.86).
In multivariate analyses, which were conducted to include significant variables on univariate analysis, only 6-month post-discharge BNP remained a significant predictor of a combined event, with the best cutoff value of 190 pg/ml at this time having 96% sensitivity and 76% specificity (Fig. 1). Six-month post-discharge BNP levels above this cutoff level strongly related to a combined event (hazard ratio [HR] 2.3; 95% confidence interval [CI] 1.4 to 3.6; p = 0.0005).
Graded relation between 6-month post-discharge BNP level and readmission for HF
Kaplan-Meier curves constructed according to 6-month post-discharge BNP values (Fig. 2A) showed that the risk of a combined event increased in a stepwise fashion across increasing ranges of 6-month post-discharge BNP, namely at <190 pg/ml, 190 to 380 pg/ml, and >380 pg/ml. Patients with a 6-month post-discharge BNP of <190 pg/ml had a better outcome than those at 190 to 380 pg/ml (HR 5.0; 95% CI 2.5 to 9.9) or >380 pg/ml (HR 11.5; 95% CI 4.5 to 29.2). Further, Kaplan-Meier curves for incidence of readmission alone (Fig. 2B) showed the same pattern (HR 4.9 or 13.9; 95% CI 2.5 to 9.8 or 4.5 to 30.1, respectively).
BNP levels and echocardiographic dimensions from discharge to 6 months post-discharge
Event-free patients showed a sustained decrease in BNP level during the 6 months after hospital discharge (mean ± standard error of the mean [SEM] of BNP level, 188 ± 21 pg/ml at discharge, 141 ± 21 pg/ml at 3 months, 108 ± 15 pg/ml at 6 months; discharge vs. 3 months, p = 0.0167; 3 months vs. 6 months, p = 0.0094). In contrast, levels in patients readmitted for decompensated HF did not decrease during follow-up (256 ± 28 pg/ml at discharge, 267 ± 31 pg/ml at 3 months, 349 ± 28 pg/ml at 6 months; discharge vs. 3 months, p = 0.79; 3 months vs. 6 months, p < 0.0001) (Fig. 3A). Regarding changes in echocardiographic dimensions, readmitted patients showed no decrease between discharge and 6 months in LVDd or LADd, or any improvement in LVEF (mean ± SEM at discharge vs. 6 months 6.5 ± 0.1 cm vs. 6.7 ± 0.2 cm, p = 0.07; 4.8 ± 0.2 cm vs. 5.4 ± 0.2 cm, p = 0.0092; 31 ± 1% vs. 29 ± 2%, p = 0.38, respectively), in contrast to event-free patients (6.5 ± 0.1 cm vs. 6.0 ± 0.1 cm, p < 0.0001; 4.3 ± 0.1 cm vs. 4.1 ± 0.1 cm, p = 0.0064; 31 ± 1% vs. 39 ± 1%, p < 0.0001, respectively) (Fig. 3B).
Our data show that BNP levels provide important long-term prognostic information even in stable low-risk outpatients after decompensated HF. The increased-risk threshold in healthy population screening is extremely low, on the order of 80 pg/ml as shown in the Framingham data (13). In acute HF or acute coronary syndrome, when BNP levels are elevated, the cutoff point for increased risk increases to about 800 pg/ml as shown by the ADHERE (Acute Decompensated Heart Failure National Registry) data (14). As levels decline over time, however, the cutoff point on the risk assessment curve at hospital discharge falls to about 500 pg/ml (9,10). Our data extend this observation in time, to show a cutoff point of about 200 pg/ml at 6 months after discharge (Fig. 1).
Entry patients in our study represent a low-risk group. They were clinically stable in NYHA functional class 1 to 2 for at least 6 months after discharge for decompensated HF, although Krumholz et al. (1) indicated that almost one-half of all patients admitted for decompensated HF are readmitted within 6 months after discharge. Circulatory stabilization on hospital discharge and socioeconomic status contribute to this favorable clinical course (9,10,15). The 28% of patients in our study had a discharge BNP level of <100 pg/ml, the optimal level in the exclusion of decompensated HF compared with only 10% in previous study (16), and our population had an extremely low median BNP level of about 200 pg/ml at discharge (Table 1). In addition, our patients received sufficient therapy for decompensated HF without socioeconomic restriction. With regard to the association between female gender and mortality, limited data exist in patients with dilated cardiomyopathy (17), and the very small number of women in our population (Table 1) may also reflect that we selected a low-risk population.
Even this low-risk group had a long-term risk of readmission for decompensated HF or sudden death, despite HF treatment including beta-blocker and ACEI, as recommended by the international guideline (11). This discrepancy can be explained by low maintenance dosages of and withdrawal of beta-blockers (18,19). Dosages of beta-blockers in our population were extremely lower compared with those reached during recent multicenter trials (20,21), and patients who underwent the withdrawal of beta-blockers were readmitted (Table 1). In addition, some patients might be nonresponders to chronic beta-blocker therapy (22). It is also possible that disease history of arrhythmias might also affect outcome (23,24), although we could not confirm its prognostic role (Table 2).
Our data also show that 6-month post-discharge BNP is closely associated with long-term outcome in this observational series, but BNP at 3 months as well as at discharge is not predictive of it (Table 2). In particular, the risk of readmission for decompensated HF adversely increased along with increasing ranges of this level, although we were unable to confirm an association with sudden death, owing to the small number of patients who died (Fig. 2). High plasma levels of BNP have been identified as major predictors of progressive LV remodeling after acute myocardial infarction (25). In our study, readmitted patients, in contrast to event-free patients, showed an increase in BNP levels despite clinical stabilization and had no improvement in echocardiographic parameters during the 6 months follow-up (Fig. 3). Increases in 6-month post-discharge BNP before clinical evidence of decompensation in part reflect a poor reverse LV remodeling. Beta-blockers have been shown to reverse LV remodeling (26) and to improve clinical outcome in patients with HF (20,21). Even in clinically stable low-risk group, higher-risk outpatients with a relatively high level of BNP, thus, may require an optimization of HF treatment including beta-blocker, implying the potential role of BNP as a therapeutic guide even in a stable outpatient setting.
Among several echocardiographic parameters, only LADd at 6 months showed a significant association with outcome (Table 2). The left atrium acts as a reservoir during LV systole (27), and LV diastolic dysfunction results in left atrial overload (28). Enlargement of the LADd may in part reflect adverse LV diastolic dysfunction. However, this variable showed no predictive power in multivariate analysis, including BNP level. But previous reports in patients with HF have described an association between LV diastolic dysfunction and outcome (29,30). Here, the involvement of patients with tachycardia, atrial fibrillation, or permanent pacing meant that we were unable to confirm the predictive power in LV diastolic filling pattern.
Our data offer circumstantial, indirect support for the setting of a target BNP level of <200 pg/ml in the outpatient management of HF, as originally hypothesized by the STARS-BNP (Plasma Brain Natriuretic Peptide-Guided Therapy to Improve Outcome in Heart Failure) multicenter study (31). The number of patients in our study is, however, relatively small, and additional prospective multicenter studies will be necessary both to confirm our observation and to test a therapeutic utility of BNP assessment in stable HF outpatients.
To our knowledge, this study is the first to indicate that even in low-risk outpatients with nonischemic dilated cardiomyopathy who have been asymptomatic for 6 months after hospital discharge for decompensated HF, BNP assessment identifies a long-term risk of readmission for decompensated HF. Our data confirm and extend the clinical utility of BNP assessment in patients with HF.
The authors thank Guy Harris for restyling the manuscript.
This study was supported by a Grant-in-Aid for Scientific Research from the Postgraduate Research Project at Kitasato University and a Grant for Scientific Research from the Ministry of Education, Science and Culture of Japan (No. 20311954).
- Abbreviations and Acronyms
- angiotensin-converting enzyme inhibitor
- angiotensin receptor blocker
- B-type natriuretic peptide
- confidence interval
- heart failure
- hazard ratio
- left atrial diastolic dimension
- left ventricle/ventricular
- left ventricular end-diastolic dimension
- left ventricular ejection fraction
- nonischemic dilated cardiomyopathy
- New York Heart Association
- Received August 22, 2007.
- Revision received October 18, 2007.
- Accepted November 28, 2007.
- American College of Cardiology Foundation
- Cowie M.R.,
- Fox K.F.,
- Wood D.A.,
- et al.
- Yoshimura M.,
- Yasue H.,
- Okamura K.,
- Ogawa H.,
- Jougasaki M.,
- Kukoyama M.
- Yasue H.,
- Yoshimura M.,
- Sumida H.,
- et al.
- Morrison L.K.,
- Harrison A.,
- Krishnaswamy P.,
- Kazanegra R.,
- Clopton P.,
- Maisel A.
- Cheng V.L.,
- Kazanegra R.,
- Garcia A.,
- et al.
- Logeart D.,
- Thabut G.,
- Jourdain P.,
- et al.
- Hunt S.A.,
- Baker D.W.,
- Chin M.H.,
- et al.
- Fonarow G.C.,
- Peacock W.F.,
- Phillips C.O.,
- Givertz M.M.,
- Lopatin M.,
- ADHERE Scientific Advisory Committee and Investigators
- Hogenhuis J.,
- Voors A.A.,
- Jaarsma T.,
- Hillege H.L.,
- Hoes A.W.,
- van Veldhuisen D.J.
- Bourassa M.G.,
- Gurne O.,
- Bangdiwala S.I.,
- et al.
- Waagstein F.,
- Caidahl K.,
- Wallentin I.,
- Bergh C.H.,
- Hjaimarson A.
- Di Lenarda A.,
- Sabbadini G.,
- Salvatore L.,
- et al.,
- the Heart-Muscle Disease Study Group
- Olshausen K.,
- Schafer A.,
- Mehmel H.C.,
- Schwarz F.,
- Senges J.,
- Kubler W.
- Doughty R.N.,
- Whalley G.A.,
- Gamble G.,
- MacMahon S.,
- Sharpe N.,
- on behalf of the Australia-New Zealand Heart Failure Research Collaborative Group
- Hoit B.D.,
- Shao Y.,
- Gabel M.,
- Walsh R.A.
- Giannuzzi P.,
- Temporelli P.L.,
- Bosimini E.,
- et al.
- Jourdain P.,
- Jondeau G.,
- Funck F.,
- et al.