© 2005 European Society of Cardiology
Pre-discharge B-type natriuretic peptide predicts early recurrence of decompensated heart failure in patients admitted to a general medical unit
a II Unit of Internal Medicine, Azienda Ospedaliera Universitaria Careggi Florence, Italy
b II Unit of Nuclear Medicine, Azienda Ospedaliera Universitaria Careggi Florence, Italy
c Department of Cardiovascular Medicine, Azienda Ospedaliera Universitaria Careggi Florence, Italy
* Corresponding author. Via A. Faccioli No. 41, 50145 Florence, Italy. Tel.: +39 55 4277411; Fax: +39 55 4277313. E-mail address: valerioverdiani{at}virgilio.it
| Abstract |
|---|
|
|
|---|
Background: B-type natriuretic peptide (BNP) represents a promising predictor of early (30 days) re-admission in patients with heart failure (HF) admitted to cardiology units. Whether BNP retains its predictive value in unselected patients admitted to general medical wards is unknown.
Methods: We determined BNP levels on admission and pre-discharge in 100 consecutive patients (71 male, mean age 78±10 years) admitted to a general medical unit due to decompensated HF. Follow-up after discharge was 30 days.
Results: Of the 100 patients, 86 had
1 comorbid conditions. Median BNP was 739 pg/ml on admission (25th–75th percentile 355–1333 pg/ml, respectively), and 414 pg/ml pre-discharge (25th–75th percentile 220–696 pg/ml). Seventeen patients were re-admitted or died within 30 days. Patients with pre-discharge BNP values >75th percentile (696 pg/ml) had greater risk of re-hospitalisation, as compared to values
696 pg/ml (56% vs. 4%, respectively; p<0.001). Negative predictive value for this cut-off was 96%. BNP values >75th percentile were associated with a 15.0 independent relative hazard (RH) of early re-admission or death (95% CI 4.2–53.8; p<0.0001). The other independent predictor was a NYHA class
III at discharge (RH 2.9; 95% CI 1.1–9.3; p<0.05).
Conclusion: In a general medical unit, pre-discharge BNP levels were a strong independent predictor of early re-admission or death due to HF, irrespective of substantial comorbidity and advanced age.
Key Words: Heart failure Natriuretic peptides Prognosis Hospital re-admission
Received July 27, 2004; Revised October 29, 2004; Accepted December 20, 2004
| 1. Introduction |
|---|
|
|
|---|
Re-hospitalisations represent a major problem for patients with heart failure (HF) [1–5] and have been reported to occur in 30–58% within 3–6 months after discharge [6–9]. Recurrent admissions are particularly frequent in elderly patients [10], and in patients with comorbidity [11]. Recent studies have focused on the issue of early re-hospitalisation (within 30 days from discharge), which occurs in nearly one-third of patients admitted with HF [12]. The identification of patients at high risk of early re-admission is important, as individual follow-up programs may reduce their likelihood of early decompensation [13].
B-type natriuretic peptide (BNP) is a cardiac neurohormone secreted from the cardiac ventricles as a response to ventricular volume expansion and pressure overload [14,15]. BNP is an established predictor of outcome in patients with HF [16,17]. Moreover, pre-discharge BNP levels appear promising in the identification of patients at risk of early re-admission or death in selected cohorts hospitalised in Cardiology units [18–20]. However, whether BNP levels retain their predictive value in patients admitted to general medical wards, characterized by multiple comorbidity and advanced age [21], has only previously been assessed in a single, relatively small study [22].
Therefore, in the present study, we chose to evaluate the relevance of pre-discharge BNP levels with regard to the subsequent risk of re-admission or death among patients with HF admitted to a general medical unit in a large regional hospital.
| 2. Methods |
|---|
|
|
|---|
2.1. Patients
Careggi hospital is a community-based multi-speciality hospital, serving the Florence metropolitan area. The study cohort included 100 patients consecutively admitted to one of four general medical units at the hospital due to decompensated HF. Criteria for inclusion were exacerbation of previously documented HF or new onset of HF using standard Framingham criteria [23]. All patients were NYHA class III or IV on admission. The only criterion for exclusion was the presence of terminal non-cardiac illness that could influence short-term prognosis, which occurred in 4 patients with advanced cancer.
2.2. Study protocol
The study was approved by the local Ethics Committee and all patients gave informed consent to participate. The initial blood sample for BNP titration was collected within 2 h in patients admitted during office hours, and on the following morning in all other patients (median was 9 h). The final blood sample was collected at 7 a.m. on the morning of discharge, which usually occurred around 1 p.m. Detailed assessment of individual comorbid conditions was recorded. All patients received standard treatment for HF [24], and were discharged following at least 2 days of clinical stability, and according to validated stability criteria [25]. Physicians were blinded to the BNP measurements, and therefore no clinical decisions were based upon these results. Patients were followed-up at 30 days for hospital re-admission or death, by phone and survey of the hospital registers.
2.3. Blood sampling and assay of BNP
Blood samples were taken after 30 min at supine rest. Plasma BNP levels were measured using peripheral venous blood samples by means of a highly sensitive immunoradiometric assay (SHIONORIA BNP assay kit; Shionogi, Osaka, Japan). Samples for determinations of plasma BNP concentrations were transferred to chilled glass tubes containing disodium ethylenediamine tetraacetic acid (1 mg/ml) and aprotinin (500 U/ml) and centrifuged immediately at 4 °C. This assay system uses two monoclonal antibodies against human BNP-32, one recognizing a carboxyterminal sequence and the other recognizing the ring structure of human BNP; the assay measures human BNP by sandwiching it between the two antibodies without extraction of plasma. This technique is different from that used in previous similar studies (i.e., Triage B-Type natriuretic peptide—Biosite Diagnostic Inc., San Diego, California, US) [19,20]; however, a comparative analysis of the two methods at our laboratory provided comparable BNP values in 35 cardiac patients with various degrees of heart failure (p=N.S.; Dr. F. Pirolo, unpublished observation).
2.4. Statistics
Comparisons of group means were made using Student's t-test for independent samples. Chi-square or Fisher's Exact Test were used to compare non-continuous variables expressed as proportions. Relative hazard and 95% confidence intervals were calculated using univariate and multivariate Cox proportional hazard regression models. Relative hazards were first computed using raw BNP values and then verified with log-transformed BNP values, as the distribution of BNP values was positively skewed. Multivariate analyses were performed using a stepwise forward regression model, with an entry probability for each variable set at 0.05. Survival curves were constructed according to the Kaplan–Meier method. A receiver operating characteristic (ROC) curve was employed for identification of the optimal discharge BNP threshold value. All P values are two-sided and considered significant when <0.05.
| 3. Results |
|---|
|
|
|---|
3.1. Baseline clinical features and outcome
Mean age of the 100 study patients was 78±10 years; 85% were >70 years; 71% were male (Table 1). The most frequent cause of HF was ischemic heart disease (62%). Most patients (86%) had at least 1 comorbid condition, and 43% had
2 (Table 1). Mean hospital stay was 8.5±4 days. All patients fulfilled validated clinical stability criteria at the time of discharge [25].
|
Following discharge, 17 patients (17%) were re-admitted within 30 days, all due to exacerbation of HF. There were no planned re-admissions. Two patients died shortly after re-admission; none died at home.
3.2. Plasma BNP levels
On admission, median BNP was 739 pg/ml (25th–75th percentile values were 355 and 1333 pg/ml, respectively); range was 98–2000 pg/ml. Pre-discharge, median BNP was 414 pg/ml (25th–75th percentile 220–696 pg/ml); range was 16–1539. During hospital stay, BNP values decreased in 82 patients and increased in 18; mean variation was –31% (–23% among patients with early re-admission, and –33% among patients without) (Table 1). Overall, admission and pre-discharge BNP values were higher in patients with early re-admission, compared to those without (Table 1).
3.3. Predictive value of pre-discharge BNP
Pre-discharge BNP values were predictive for early re-admission in our cohort (relative hazard – RH – per 50 pg/ml increase 1.10; 95% confidence interval – CI – 1.06–1.13, p<0.0001). The same level of significance was obtained using log-transformed pre-discharge BNP values (RH 6.8; 95% CI 3.2–14.5, p<0.0001). Pre-discharge BNP levels >75th percentile (696 pg/ml) were associated with a 14-fold increase in risk of re-admission (14/25, or 56%, representing the positive predictive value), as compared to levels
696 pg/ml (3/75, or 4%; p<0.001). Negative predictive value was 96% (Fig. 1). ROC curve analysis identified a BNP value of 789 pg/ml as the best cut-off for the identification of patients at risk of early re-admission, with a positive predictive value of 74%, and a negative predictive value of 96% (Fig. 2 and Table 2).
|
|
|
At multivariate analysis, a BNP level exceeding the 75th percentile (696 pg/ml) was associated with an independent RH of 15.0 for early re-admission or death (95% CI 4.2–53.8; p<0.0001). The only other independent predictor was a NYHA class III–IV at discharge (Table 3).
|
3.4. Predictive value of BNP levels on admission and in-hospital variations
Admission BNP values showed a significant association with subsequent risk of early re-admission in our cohort (RH per 50 pg/ml increase 1.02; 95% CI 1.01–1.03, p=0.0001; RH for log-transformed admission BNP 9.6, 95% CI 3.0–30.7, p=0.0001). However, the association with the end-point was not as strong as that of pre-discharge BNP value: at multivariate survival analysis, a BNP value on admission exceeding the 75th percentile (1333 pg/ml) had a RH of 5.1 (95% CI 1.8–14.0, p=0.0017).
Conversely, individual in-hospital BNP variations (admission minus pre-discharge values) were not associated with the subsequent risk of early re-admission (RH per 50 pg/ml decrease 0.99; 95% CI 0.94–1.04, p=0.74).
| 4. Discussion |
|---|
|
|
|---|
The main result of the present study is that pre-discharge BNP levels represent a powerful predictor of re-admission within 30 days in patients hospitalised with decompensated HF. Among 100 consecutive patients, we observed a 17% re-admission rate at 30 days (with subsequent death in two patients), which is comparable to previous reports by other groups [20,26]. Patients discharged with a pre-discharge BNP value exceeding the 75th percentile of the whole study group (696 pg/ml) showed a 15.0 relative hazard of early re-admission, as compared to patients at or below this threshold. A pre-discharge BNP >696 pg/ml had a 56% positive and 96% negative predictive value with regard to early re-admission. Positive predictive value increased to 74% for the optimal threshold BNP level of 789 pg/ml, as calculated by ROC curve analysis.
Our findings are in close agreement with two prior studies addressing the issue of pre-discharge BNP and early re-admission in HF patients. In a pilot study based on 72 male patients with HF, Cheng et al. found that elevated BNP concentrations at discharge strongly predicted a subsequent re-admission within 30 days [19]. Likewise, Logeart et al. recently showed pre-discharge BNP levels to be significantly associated with re-admission due to HF or death at 30 days and 6 months, and indicated a BNP value of 350 pg/ml as the optimal threshold value obtained by ROC curve analysis [20].
As compared to the latter study, however, our patients were significantly older (mean age 78±10 years vs. 71±13, p<0.0001), and had a substantial array of non-cardiac comorbid conditions. Specifically, 86% of our study patients had at least 1 comorbidity, and 43% had two or more. Not surprisingly, average discharge BNP values were higher in our cohort, compared to those reported by the quoted study [20]. These distinctive features of our study cohort reflect the patient population routinely admitted to general medical units as opposed to the more selective cardiology units [21], and represents "the real world" of patients with HF [27].
The issue of comorbidity in patients with HF is of relevance for at least two reasons: (a) risk of early re-hospitalisation is greater among HF patients with associated non-cardiac disease, with particular regard to chronic renal failure, chronic obstructive pulmonary disease, diabetes mellitus, hypertension, anxiety and depression [28]; and (b) several of the aforementioned conditions, as well as age, may increase serum BNP levels and thus alter their predictive accuracy [29–32].
In spite of these potential concerns, we found pre-discharge BNP values to maintain a strong predictive value for early re-admission due to HF, independent and irrespective of substantial comorbidity and advanced age. These findings thus justify the clinical use of pre-discharge BNP assessment to establish patient stability and likelihood for HF recurrence in unselected patient populations admitted to general medical units. However, in agreement with Cheng et al. [19], the threshold value which we found clinically meaningful in this setting was substantially higher than that reported in a more selected patient cohort [20]. This fact needs to be taken into account, since lower BNP thresholds would inevitably lead to an excessive rate of false positives in cohorts such as ours.
The implications of our study findings are twofold: on one hand, very high BNP values on the planned day of discharge may identify patients with insufficient clinical balance, despite compliance with generally accepted stability criteria [25], and thus warrant prolongation of the hospital stay. Alternatively, patients with increased pre-discharge BNP may be included in close individual domiciliary or outpatient follow-up programs [13], although the efficacy of such strategy in reducing their likelihood of early re-admission remains to be established.
4.1. Admission vs. pre-discharge BNP levels
In the present cohort, elevated BNP levels on admission were also associated with increased risk of early re-admission, although such association was weaker than for pre-discharge BNP. This finding is in agreement with the data by Cheng et al. [19], although studies with longer follow-up do not confirm an association of admission BNP levels with outcome at 6 and 12 months [22,33]. One of the reasons for this discrepancy could be that admission BNP values are only predictive in the short term. Furthermore, a substantial similarity of admission and pre-discharge BNP levels in our sickest patients, reflecting limited capacity for clinical improvement, probably accounted for their comparable predictive value in the present study.
Finally, absolute individual BNP changes during hospitalisation were not associated with early re-admission or death. This finding is in disagreement with the work by Cheng et al. [19], and again may be related to the limited variation in BNP levels in a very elderly patient population with multiple comorbidities.
Thus, our findings support the exclusive use of pre-discharge BNP levels for assessment of short-term outcome in HF patients, whereas the utility of assessing BNP on admission or in-hospital BNP variations appears questionable [33].
| 5. Conclusions |
|---|
|
|
|---|
In a general medical unit, BNP levels at discharge were a strong independent predictor of early re-admission or death due to HF, irrespective of substantial comorbidity and advanced age. The optimal threshold BNP value, however, was substantially higher in this setting compared to the more selected patient populations admitted to cardiology units. These findings may be of relevance in planning discharge strategies for patients at high risk of HF recurrence.
| References |
|---|
|
|
|---|
- Sharpe N., Doughty R. Epidemiology of heart failure and ventricular dysfunction. Lancet (1998) 352(Suppl. I):3–7.[CrossRef][Web of Science][Medline]
- McMurray J.J., Stewart S. Epidemiology, aetiology, and prognosis of heart failure. Heart (2000) 83:596–602.
[Free Full Text] - O'Connel J.B., Bristow M. Economic impact of heart failure in the United States: time for a different approach. J. Heart Lung Transplant. (1993) 13:S107–S112.[Web of Science]
- Ghali J.K., Cooper R., Ford E. Trends in hospitalization for heart failure in the United States, 1973–1986: evidence for increasing population prevalence. Arch. Intern. Med. (1990) 150:769–773.
[Abstract/Free Full Text] - Bellotti P., Badano L.P., Acquarone N., Griffo R., Lo Pinto G., Maggioni A.P., et al. Specialty-related differences in the epidemiology, clinical profile, management and outcome of patients hospitalized for heart failure: the OSCUR study. Outcome dello Scompenso Cardiaco in relazione all'Utilizzo delle Risorse. Eur. Heart J. (2001) 22:596–604.
[Abstract/Free Full Text] - Kahn K.L., Rogers W.H., Rubenstein L.V., Sherwood M.J., Reinisch E.J., Keeler E.B., et al. Measuring quality of care with explicit process criteria before and after implementation of the DRG-based prospective payment system. JAMA (1990) 264:1969–1973.
[Abstract/Free Full Text] - Ashton C.M., Kuykendall D.H., Johnson M.L., Wray N.P., Wu L. The association between the quality of inpatient care and early readmission. Ann. Intern. Med. (1995) 122:415–421.
[Abstract/Free Full Text] - Brophy J.M., Deslauriers G., Boucher B., Rouleau J.L. The hospital course and short term prognosis of patients presenting to the emergency room with decompensated congestive heart failure. Can. J. Cardiol. (1993) 9:219–224.[Web of Science][Medline]
- Rich M.W., Beckham V., Wittenberg C., Leven C.L., Freedland K.E., Carney R.M. A multidisciplinary intervention to prevent the readmission of elderly patients with congestive heart failure. N. Engl. J. Med. (1995) 333:1190–1195.
[Abstract/Free Full Text] - Krumholz H.M., Parent E.M., Tu N., Vaccarino V., Wang Y., Radford M.J., et al. Readmission after hospitalization for congestive heart failure among Medicare beneficiaries. Arch. Intern. Med. (1997) 157:99–104.
[Abstract/Free Full Text] - Brown A.M., Cleland J.G.F. Influence of concomitant disease on patterns of hospitalization in patients with heart failure discharged from Scottish hospitals in 1995. Eur. Heart J. (1998) 19:1063–1069.
[Abstract/Free Full Text] - Jaarsma T., Halfens R., Huijer Abu-Saad H., Dracup K., Gorgels T., van Ree J., et al. Effect of education and support on self-care and resource utilization in patients with heart failure. Eur. Heart J. (1999) 20:673–682.
[Abstract/Free Full Text] - Grady K.L., Dracup K., Kennedy G., Moser D.K., Piano M., Stevenson L.W., et al. Team management of patients with heart failure. A statement for healthcare professionals from the Cardiovascular Nursing Council of the American Heart Association. Circulation (2000) 102:2443–2456.
[Free Full Text] - Nagagawa O., Ogawa Y., Itoh H., Suga S., Komatsu Y., Kishimoto I. Rapid transcriptional activation and early mRNA turnover of BNP in cardiocyte hypertrophy: evidence for BNP as an "emergency" cardiac hormone against ventricular overload. J. Clin. Invest. (1995) 96:1280–1287.[Web of Science][Medline]
- Dickstein K. Natriuretic peptides in detection of heart failure. Lancet (1998) 35:3–4.
- Mair J., Hammerer-Lercher A., Puschendorf B. The impact of cardiac natriuretic peptide determination on the diagnosis and management of heart failure. Clin. Chem. Lab. Med. (2001) 39:571–588.[CrossRef][Web of Science][Medline]
- Bettencourt P. NT-proBNP: biomarkers for heart failure management. Eur. J. Heart Fail. (2004) 6:359–363.
[Abstract/Free Full Text] - Yu C.M., Sanderson J.E. Plasma brain natriuretic peptide—an independent predictor of cardiovascular mortality in acute heart failure. Eur. J. Heart Fail. (1999) 1:59–65.
[Abstract/Free Full Text] - Cheng V., Kazanagra R., Garcia A., Lenert L., Krishnaswamy P., Gardetto N., et al. A rapid bedside test for B-type peptide predicts treatment outcomes in patients admitted for decompensated heart failure: a pilot study. J. Am. Coll. Cardiol. (2001) 37:386–391.
[Abstract/Free Full Text] - Logeart D., Thabut G., Jourdain P., Chavelas C., Beyne P., Beauvais F., et al. Predischarge B-type natriuretic peptide assay for identifying patients at high risk of re-admission after decompensated heart failure. J. Am. Coll. Cardiol. (2004) 43:635–641.
[Abstract/Free Full Text] - Di Lenarda A., Scherillo M., Maggioni A.P., Acquarone N., Ambrosio G.B., Annicchiarico M., et al. Current presentation and management of heart failure in cardiology and internal medicine hospital units: a tale of two worlds—the TEMISTOCLE study. Am. Heart J. (2003) 146(4):e12.[Medline]
- Bettencourt P., Ferreira S., Azevedo A., Ferreira A. Preliminary data on the potential usefulness of B-type natriuretic peptide levels in predicting outcome after hospital discharge in patients with heart failure. Am. J. Med. (2002) 113:215–219.[CrossRef][Web of Science][Medline]
- Ho K.K., Pinsky J.L., Kannel W.B., Levy D. The epidemiology of heart failure: the Framingham Study. J. Am. Coll. Cardiol. (1993) 22:6A–13A.[Medline]
- The Task Force for the Diagnosis and Treatment of Chronic Heart Failure, European Society of Cardiology. Guidelines for the diagnosis and treatment of chronic heart failure. Eur. Heart J. (2001) 22:1527–1560.
[Free Full Text] - Stevenson L.W., Sietsema K., Tillisch J.H., Lem V., Walder J., Kobashigawa J.A., et al. Exercise capacity for survivors of cardiac transplantation or sustained medical therapy for stable heart failure. Circulation (1990) 81:78–85.
[Abstract/Free Full Text] - Cline C., Broms K., Willenheimer R., Israelsson B., Erhardt L. Hospitalization and health care costs due to congestive heart failure in the elderly. Am. J. Geriatr. Cardiol. (1996) 5:10–23.[Medline]
- Graves E.J. National hospital discharge survey: annual summary 1991. (1993) Washington, DC: National Center for Health Statistics.
- Braunstein J.B., Anderson G.F., Gerstenblith G., Weller W., Niefeld M., Herbert R., et al. Noncardiac comorbidity increases preventable hospitalizations and mortality among Medicare beneficiaries with chronic heart failure. J. Am. Coll. Cardiol. (2003) 42:1226–1233.
[Abstract/Free Full Text] - Dutka D.P., Olivotto I., Ward S., Oakley C.M., Impallomeni M., Cleland J.G.F. Effects of aging on neuroendocrine activation in subjects and patients in the presence and absence of heart failure with left ventricular systolic dysfunction. Am. J. Cardiol. (1996) 77:1197–1201.[CrossRef][Web of Science][Medline]
- Dutka D.P., Olivotto I., Ward S., Nihoyannopoulos P., al-Subaili M., Oakley C.N., et al. Plasma neuro-endocrine activity in very elderly subjects and patients with and without heart failure. Eur. Heart J. (1995) 16:1223–1230.
[Abstract/Free Full Text] - Nielsen O.W., Kirk V., Bay M., Boesgaard S., Nielsen H. Value of N-terminal pro brain natriuretic peptide in the elderly: data from the prospective Copenhagen Hospital Heart Failure study (CHHF). Eur. J. Heart Fail. (2004) 6:275–279.
[Abstract/Free Full Text] - Akiba T., Tachibana K., Togashi K., Hiroe M., Marumo F. Plasma human brain natriuretic peptide in chronic renal failure. Clin. Nephrol. (1995) 44(Suppl. 1):S61–S64.[Web of Science][Medline]
- O'Brien R.J., Squire I.B., Demme B., Davies J.E., Ng L.L. Pre-discharge, but not admission, levels of NT-proBNP predict adverse prognosis following acute LVF. Eur. J. Heart Fail. (2003) 5:499–506.
[Abstract/Free Full Text]
This article has been cited by other articles:
![]() |
P. Ray, S. Delerme, P. Jourdain, and C. Chenevier-Gobeaux Differential diagnosis of acute dyspnea: the value of B natriuretic peptides in the emergency department QJM, November 1, 2008; 101(11): 831 - 843. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Nishii, T. Inomata, H. Takehana, T. Naruke, T. Yanagisawa, M. Moriguchi, S. Takeda, and T. Izumi Prognostic utility of B-type natriuretic peptide assessment in stable low-risk outpatients with nonischemic cardiomyopathy after decompensated heart failure. J. Am. Coll. Cardiol., June 17, 2008; 51(24): 2329 - 2335. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. Moertl, A. Hammer, M. Huelsmann, R. Pacher, and R. Berger Prognostic value of sequential measurements of amino-terminal prohormone of B-type natriuretic peptide in ambulatory heart failure patients Eur J Heart Fail, April 1, 2008; 10(4): 404 - 411. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. M. Felker, J. W. Petersen, and D. B. Mark Natriuretic peptides in the diagnosis and management of heart failure. Can. Med. Assoc. J., September 12, 2006; 175(6): 611 - 617. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||





