International Journal of Cardiology 177 (2014) 673–675
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Letter to the Editor
Association of serum bicarbonate with long-term outcomes in patients hospitalized with heart failure☆,☆☆ Michael E. Nassif, Eric Novak, Michael W. Rich ⁎ Department of Medicine, Division of Cardiology, Washington University School of Medicine, St. Louis, MO, United States
a r t i c l e
i n f o
Article history: Received 19 September 2014 Accepted 27 September 2014 Available online 5 October 2014 Keywords: Biomarkers Heart failure Mortality Readmissions Serum bicarbonate
Acute decompensated heart failure (ADHF) requiring hospitalization is associated with significant morbidity and mortality, with 30-day readmission rates of 20–25% and 6-month mortality rates of 15–20% [1,2]. Acid–base disorders are common in patients with ADHF due to reduced effective circulatory volume, impaired tissue perfusion, renal insufficiency, activation of the renin–angiotensin–aldosterone-system (RAAS) and diuretic therapy [3–5]. Serum bicarbonate (HCO3) levels, which reflect acid–base homeostasis, are routinely measured at the time of admission for ADHF and are evaluated serially throughout the hospital stay. We conducted a retrospective analysis to determine the relationships between HCO3 levels and changes in HCO3 levels with long-term outcomes, including death and readmissions, in a cohort of patients hospitalized with ADHF. Patients discharged alive from January 2007 through December 2008 with a primary diagnosis of heart failure (HF) were eligible for inclusion. Patients with estimated glomerular filtration rates (eGFR) b30 cc/min/1.73 m2 by the Cockcroft–Gault method were excluded. Serum HCO3 levels were recorded from the basic metabolic profile obtained at admission and prior to discharge. Subjects were stratified into terciles based on the mean of the two HCO3 values. The primary outcome was time to death following hospital discharge, and secondary
☆ Funding: Washington University Department of Internal Medicine Mentors in Medicine grant. ☆☆ Relationships with industry: None. ⁎ Corresponding author at: Washington University School of Medicine, 660 S. Euclid Ave., Campus Box 8086, St. Louis, MO 63110. Tel.: +1 314 454 8146; fax: +1 314 362 2512. E-mail address: [email protected] (M.W. Rich).
http://dx.doi.org/10.1016/j.ijcard.2014.09.166 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
outcomes included time until first HF readmission within 180 days and the composite of death or HF readmission within 180 days. Baseline characteristics according to HCO3 terciles are presented in Table 1. During a mean follow-up of 39.9 ± 23.1 months, there were 214 deaths (53.5%) and 122 patients (30.5%) were readmitted with HF within 180 days. In the primary analysis, there were no significant associations between mean HCO3 tercile and the primary or secondary endpoints (data not shown). However, there was a strong association between absolute change in HCO3 levels from admission to discharge and long-term mortality (Fig. 1; p b 0.001). In a Cox proportional hazards model adjusting for 17 pre-specified variables known to be associated with prognosis in patients hospitalized with ADHF, including renal function, ejection fraction (EF), and diuretic dose, absolute change in HCO3 tercile was an independent predictor of increased mortality, with an adjusted hazard ratio (AHR) between the highest and lowest terciles of 1.65, 95% CI 1.15–2.35, p = 0.006. No association was found between terciles of absolute change in HCO3 levels and HF readmissions at 180 days, but for the combined endpoint of time to first HF readmission or death at 180 days, there was a significant difference between the highest and lowest terciles (AHR 1.91, 95% CI 1.27–2.91, p = 0.002). For this endpoint there was also a trend comparing the middle and lowest terciles, but this did not reach significance (AHR 1.47, 95% CI 0.99–2.20, p = 0.06). The association between absolute change in HCO3 and mortality was not driven by direction of change, but rather by magnitude of change. In both models, further adjustment for change in eGFR, change in blood urea nitrogen during hospitalization, and length of stay did not alter the findings. B-type natriuretic peptide (BNP) was available for 84% of the cohort. When admission BNP was added to the Cox model, the AHR for the combined endpoint of time to death or HF readmission at 180 days comparing the highest and lowest terciles was slightly attenuated but remained significant (AHR 1.70, 95% CI 1.08–2.68, p = 0.02). Absolute difference in HCO3 by terciles provided a 1.6% improvement in the discrimination slope for the outcome death (IDI = 1.6, p = 0.04) and 2.5% improvement for the composite outcome of death or HF readmission at 180 days (IDI = 2.5, p = 0.007). The net reclassification improvement (NRI) was 24.7% when absolute change in HCO3 was added to the 180-day mortality model (p = 0.03), and 14.2% for the composite of death and readmission at 180 days (p = 0.02). Modest improvements in model c-statistics were not significant. To our knowledge, this is the first study to examine the association between serum HCO3 and clinical outcomes in patients hospitalized with ADHF. Although serum HCO3
674
M.E. Nassif et al. / International Journal of Cardiology 177 (2014) 673–675
Table 1 Baseline characteristics.
Age (years) Male Caucasian Body mass index Vital signs Heart rate Systolic blood pressure Diastolic blood pressure Medical Hx Cornary artery disease Atrial fibrillation Diabets mellitus Hypertension COPD CHF etiology NICM Ischemic cardiomyopathy HFPEF Ejection fraction Lab on admission Hemoglobin (g/dl) Na (mmol/L) K (mmol/L) CrCl (ml/min) BUN (mg/dl) BNP Meds on DC Beta Blocker ACE-I/ARB Spironolactone Nitrates Hydralazine Loop diureticsa
Mean HCO3 ≤ 26 (N = 143)
26 b Mean HCO3 ≤ 28 (N = 124)
Mean HCO3 N 28 (N = 133)
62.3 ± 16.28 84 (59%) 50 (35%) 29.5 ± 8.17
63.8 ± 14.58 67 (54%) 57 (48%) 30.8 ± 8.93
67.6 ± 15.46 63 (47%) 57 (44%) 32.9 ± 10.73
0.014 0.16 0.13 0.011
83.5 ± 15.98 136.2 ± 24.26 80.9 ± 16.02
82.9 ± 17.11 133.1 ± 23.37 75.9 ± 16.90
81.6 ± 18.10 136.4 ± 24.08 75.6 ± 15.59
0.65 0.46 0.010
58 (41%) 46 (32%) 51 (36%) 118 (83%) 17 (12%)
57 (46%) 34 (28%) 46 (37%) 82 (67%) 20 (16%)
59 (44%) 50 (38%) 61 (46%) 103 (77%) 37 (28%)
0.66 0.24 0.21 0.008 0.003
80 (56%) 41 (29%) 22 (15%) 33.3 ± 15.67
68 (55%) 33 (27%) 23 (19%) 34.4 ± 16.34
65 (49%) 36 (27%) 32 (24%) 39.8 ± 16.16
0.47
0.002
12.1 ± 2.16 139.3 ± 4.01 4.4 ± 0.58 75.1 ± 35.68 25.6 ± 14.41 1057 ± 1023
12.2 ± 1.95 139.9 ± 3.16 4.1 ± 0.52 83.2 ± 40.71 24.1 ± 13.39 882 ± 792
11.9 ± 2.04 139.8 ± 3.70 4.1 ± 0.58 86.1 ± 65.32 24.9 ± 12.96 737 ± 814
0.65 0.41 b0.001 0.15 0.69 0.02
111 (78%) 107 (75%) 44 (31%) 39 (27%) 39 (27%) 52.9 ± 49.6
101 (81%) 96 (77%) 38 (31%) 26 (21%) 25 (20%) 57.9 ± 45.5
92 (69%) 93 (70%) 32 (24%) 27 (20%) 30 (23%) 81.8 ± 70.4
0.06 0.39 0.38 0.33 0.36 b0.001
ACE-I/ARB = Angiotensin converting enzyme inhibitor/angiotensin receptor blocker. BNP = B-type natriuretic peptide. BUN = Blood urea nitrogen. COPD = Chronic obstructive pulmonary disease. CrCl = Creatinine clearance. HFPEF = Heart failure with preserved ejection fraction. NICM = Non ischemic cardiomyopathy. a Loop diuretic dosage is reported in furosemide equivalent units.
Fig. 1. Time until death absolute difference in HCO3 (discharge HCO3 − admission HCO3).
p-Value
M.E. Nassif et al. / International Journal of Cardiology 177 (2014) 673–675
levels measured at admission and discharge were not correlated with mortality or readmissions during follow-up, the absolute change in HCO3 levels from admission to discharge was an independent predictor of long-term mortality and the composite outcome of death or HF readmission within 6 months of discharge. Compared to patients with an absolute change in HCO3 level of ≤ 2 mmol/L, those with an absolute change of N5 mmol/L had a hazard ratio of 1.65 for death and 1.91 for death or HF readmission within 180 days. When added to the multivariate 180-day mortality model, absolute change in HCO3 correctly reclassified 24.7% of patients. These findings suggest that a substantial change in HCO3 level during the course of hospitalization for ADHF is a marker for more tenuous hemodynamic and/or neurohormonal status. The mechanisms for the observed association between HCO3 levels and clinical outcomes are unknown. Patients with greater change in HCO3 levels may have more severe HF. Although we adjusted for EF, BNP, and blood pressure on admission, we did not have data on the New York Heart Association class or other measures of functional status. Change in HCO3 levels could also reflect excessive diuresis with impaired renal perfusion. We adjusted for diuretic dosage, baseline renal function, and changes in renal function during hospitalization. Acid–base balance is principally regulated by the kidneys, the renin– angiotensin–aldosterone system (RAAS), and the lungs (via alterations in ventilation). Chronically elevated aldosterone levels are associated with metabolic alkalosis, and high-dose diuretics acutely increase serum aldosterone levels. Daily furosemide dose also correlates with plasma aldosterone (r = 0.77) [6]. We speculate that patients with greater changes in HCO3 levels had correspondingly greater activation of the RAAS. Since the magnitude of RAAS activation correlates with HF severity and prognosis, we suspect that this mechanism likely contributed to our observations [6–9]. Although confirmation is needed, our findings suggest that monitoring change in HCO3 levels during hospitalization for ADHF could facilitate identification of patients at increased risk for readmission or death. Thus, absolute change in HCO3 during admission may represent
675
a novel biomarker for predicting subsequent outcomes. Given the routine availability of HCO3 levels in patients admitted with ADHF and its very low acquisition cost, additional investigation of its prognostic utility is warranted.
Conflict of interest The authors report no relationships that could be construed as a conflict of interest.
References [1] Lee DS, Austin PC, Rouleau JL, Liu PP, Naimark D, Tu JV. Predicting mortality among patients hospitalized for heart failure: derivation and validation of a clinical model. JAMA 2003;290:2581–7. [2] Levy WC, Mozaffarian D, Linker DT, Sutradhar SC, Anker SD, Cropp AB, et al. The Seattle Heart Failure Model: prediction of survival in heart failure. Circulation 2006; 113:1424–33. [3] Francis GS, Benedict C, Johnstone DE, Kirlin PC, Nicklas J, Liang CS, et al. Comparison of neuroendocrine activation in patients with left ventricular dysfunction with and without congestive heart failure. A substudy of the Studies of Left Ventricular Dysfunction (SOLVD). Circulation 1990;82:1724–9. [4] Milionis HJ, Alexandrides GE, Liberopoulos EN, Bairaktari ET, Goudevenos J, Elisaf MS. Hypomagnesemia and concurrent acid–base and electrolyte abnormalities in patients with congestive heart failure. Eur J Heart Fail 2002;4:167–73. [5] Packer M. Pathophysiology of chronic heart failure. Lancet 1992;340:88–92. [6] Eshaghian S, Horwich TB, Fonarow GC. Relation of loop diuretic dose to mortality in advanced heart failure. Am J Cardiol 2006;97:1759–64. [7] Knight RK, Miall PA, Hawkins LA, Dacombe J, Edwards CR, Hamer J. Relation of plasma aldosterone concentration to diuretic treatment in patients with severe heart disease. Br Heart J 1979;42:316–25. [8] Swedberg K, Eneroth P, Kjekshus J, Wilhelmsen L. Hormones regulating cardiovascular function in patients with severe congestive heart failure and their relation to mortality. CONSENSUS Trial Study Group. Circulation 1990;82:1730–6. [9] Girerd N, Pang PS, Swedberg K, Fought A, Kwasny MJ, Subacius H, et al. Serum aldosterone is associated with mortality and re-hospitalization in patients with reduced ejection fraction hospitalized for acute heart failure: analysis from the EVEREST trial. Eur J Heart Fail 2013;15:1228–35.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Letter to the Editor
Association of serum bicarbonate with long-term outcomes in patients hospitalized with heart failure☆,☆☆ Michael E. Nassif, Eric Novak, Michael W. Rich ⁎ Department of Medicine, Division of Cardiology, Washington University School of Medicine, St. Louis, MO, United States
a r t i c l e
i n f o
Article history: Received 19 September 2014 Accepted 27 September 2014 Available online 5 October 2014 Keywords: Biomarkers Heart failure Mortality Readmissions Serum bicarbonate
Acute decompensated heart failure (ADHF) requiring hospitalization is associated with significant morbidity and mortality, with 30-day readmission rates of 20–25% and 6-month mortality rates of 15–20% [1,2]. Acid–base disorders are common in patients with ADHF due to reduced effective circulatory volume, impaired tissue perfusion, renal insufficiency, activation of the renin–angiotensin–aldosterone-system (RAAS) and diuretic therapy [3–5]. Serum bicarbonate (HCO3) levels, which reflect acid–base homeostasis, are routinely measured at the time of admission for ADHF and are evaluated serially throughout the hospital stay. We conducted a retrospective analysis to determine the relationships between HCO3 levels and changes in HCO3 levels with long-term outcomes, including death and readmissions, in a cohort of patients hospitalized with ADHF. Patients discharged alive from January 2007 through December 2008 with a primary diagnosis of heart failure (HF) were eligible for inclusion. Patients with estimated glomerular filtration rates (eGFR) b30 cc/min/1.73 m2 by the Cockcroft–Gault method were excluded. Serum HCO3 levels were recorded from the basic metabolic profile obtained at admission and prior to discharge. Subjects were stratified into terciles based on the mean of the two HCO3 values. The primary outcome was time to death following hospital discharge, and secondary
☆ Funding: Washington University Department of Internal Medicine Mentors in Medicine grant. ☆☆ Relationships with industry: None. ⁎ Corresponding author at: Washington University School of Medicine, 660 S. Euclid Ave., Campus Box 8086, St. Louis, MO 63110. Tel.: +1 314 454 8146; fax: +1 314 362 2512. E-mail address: [email protected] (M.W. Rich).
http://dx.doi.org/10.1016/j.ijcard.2014.09.166 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
outcomes included time until first HF readmission within 180 days and the composite of death or HF readmission within 180 days. Baseline characteristics according to HCO3 terciles are presented in Table 1. During a mean follow-up of 39.9 ± 23.1 months, there were 214 deaths (53.5%) and 122 patients (30.5%) were readmitted with HF within 180 days. In the primary analysis, there were no significant associations between mean HCO3 tercile and the primary or secondary endpoints (data not shown). However, there was a strong association between absolute change in HCO3 levels from admission to discharge and long-term mortality (Fig. 1; p b 0.001). In a Cox proportional hazards model adjusting for 17 pre-specified variables known to be associated with prognosis in patients hospitalized with ADHF, including renal function, ejection fraction (EF), and diuretic dose, absolute change in HCO3 tercile was an independent predictor of increased mortality, with an adjusted hazard ratio (AHR) between the highest and lowest terciles of 1.65, 95% CI 1.15–2.35, p = 0.006. No association was found between terciles of absolute change in HCO3 levels and HF readmissions at 180 days, but for the combined endpoint of time to first HF readmission or death at 180 days, there was a significant difference between the highest and lowest terciles (AHR 1.91, 95% CI 1.27–2.91, p = 0.002). For this endpoint there was also a trend comparing the middle and lowest terciles, but this did not reach significance (AHR 1.47, 95% CI 0.99–2.20, p = 0.06). The association between absolute change in HCO3 and mortality was not driven by direction of change, but rather by magnitude of change. In both models, further adjustment for change in eGFR, change in blood urea nitrogen during hospitalization, and length of stay did not alter the findings. B-type natriuretic peptide (BNP) was available for 84% of the cohort. When admission BNP was added to the Cox model, the AHR for the combined endpoint of time to death or HF readmission at 180 days comparing the highest and lowest terciles was slightly attenuated but remained significant (AHR 1.70, 95% CI 1.08–2.68, p = 0.02). Absolute difference in HCO3 by terciles provided a 1.6% improvement in the discrimination slope for the outcome death (IDI = 1.6, p = 0.04) and 2.5% improvement for the composite outcome of death or HF readmission at 180 days (IDI = 2.5, p = 0.007). The net reclassification improvement (NRI) was 24.7% when absolute change in HCO3 was added to the 180-day mortality model (p = 0.03), and 14.2% for the composite of death and readmission at 180 days (p = 0.02). Modest improvements in model c-statistics were not significant. To our knowledge, this is the first study to examine the association between serum HCO3 and clinical outcomes in patients hospitalized with ADHF. Although serum HCO3
674
M.E. Nassif et al. / International Journal of Cardiology 177 (2014) 673–675
Table 1 Baseline characteristics.
Age (years) Male Caucasian Body mass index Vital signs Heart rate Systolic blood pressure Diastolic blood pressure Medical Hx Cornary artery disease Atrial fibrillation Diabets mellitus Hypertension COPD CHF etiology NICM Ischemic cardiomyopathy HFPEF Ejection fraction Lab on admission Hemoglobin (g/dl) Na (mmol/L) K (mmol/L) CrCl (ml/min) BUN (mg/dl) BNP Meds on DC Beta Blocker ACE-I/ARB Spironolactone Nitrates Hydralazine Loop diureticsa
Mean HCO3 ≤ 26 (N = 143)
26 b Mean HCO3 ≤ 28 (N = 124)
Mean HCO3 N 28 (N = 133)
62.3 ± 16.28 84 (59%) 50 (35%) 29.5 ± 8.17
63.8 ± 14.58 67 (54%) 57 (48%) 30.8 ± 8.93
67.6 ± 15.46 63 (47%) 57 (44%) 32.9 ± 10.73
0.014 0.16 0.13 0.011
83.5 ± 15.98 136.2 ± 24.26 80.9 ± 16.02
82.9 ± 17.11 133.1 ± 23.37 75.9 ± 16.90
81.6 ± 18.10 136.4 ± 24.08 75.6 ± 15.59
0.65 0.46 0.010
58 (41%) 46 (32%) 51 (36%) 118 (83%) 17 (12%)
57 (46%) 34 (28%) 46 (37%) 82 (67%) 20 (16%)
59 (44%) 50 (38%) 61 (46%) 103 (77%) 37 (28%)
0.66 0.24 0.21 0.008 0.003
80 (56%) 41 (29%) 22 (15%) 33.3 ± 15.67
68 (55%) 33 (27%) 23 (19%) 34.4 ± 16.34
65 (49%) 36 (27%) 32 (24%) 39.8 ± 16.16
0.47
0.002
12.1 ± 2.16 139.3 ± 4.01 4.4 ± 0.58 75.1 ± 35.68 25.6 ± 14.41 1057 ± 1023
12.2 ± 1.95 139.9 ± 3.16 4.1 ± 0.52 83.2 ± 40.71 24.1 ± 13.39 882 ± 792
11.9 ± 2.04 139.8 ± 3.70 4.1 ± 0.58 86.1 ± 65.32 24.9 ± 12.96 737 ± 814
0.65 0.41 b0.001 0.15 0.69 0.02
111 (78%) 107 (75%) 44 (31%) 39 (27%) 39 (27%) 52.9 ± 49.6
101 (81%) 96 (77%) 38 (31%) 26 (21%) 25 (20%) 57.9 ± 45.5
92 (69%) 93 (70%) 32 (24%) 27 (20%) 30 (23%) 81.8 ± 70.4
0.06 0.39 0.38 0.33 0.36 b0.001
ACE-I/ARB = Angiotensin converting enzyme inhibitor/angiotensin receptor blocker. BNP = B-type natriuretic peptide. BUN = Blood urea nitrogen. COPD = Chronic obstructive pulmonary disease. CrCl = Creatinine clearance. HFPEF = Heart failure with preserved ejection fraction. NICM = Non ischemic cardiomyopathy. a Loop diuretic dosage is reported in furosemide equivalent units.
Fig. 1. Time until death absolute difference in HCO3 (discharge HCO3 − admission HCO3).
p-Value
M.E. Nassif et al. / International Journal of Cardiology 177 (2014) 673–675
levels measured at admission and discharge were not correlated with mortality or readmissions during follow-up, the absolute change in HCO3 levels from admission to discharge was an independent predictor of long-term mortality and the composite outcome of death or HF readmission within 6 months of discharge. Compared to patients with an absolute change in HCO3 level of ≤ 2 mmol/L, those with an absolute change of N5 mmol/L had a hazard ratio of 1.65 for death and 1.91 for death or HF readmission within 180 days. When added to the multivariate 180-day mortality model, absolute change in HCO3 correctly reclassified 24.7% of patients. These findings suggest that a substantial change in HCO3 level during the course of hospitalization for ADHF is a marker for more tenuous hemodynamic and/or neurohormonal status. The mechanisms for the observed association between HCO3 levels and clinical outcomes are unknown. Patients with greater change in HCO3 levels may have more severe HF. Although we adjusted for EF, BNP, and blood pressure on admission, we did not have data on the New York Heart Association class or other measures of functional status. Change in HCO3 levels could also reflect excessive diuresis with impaired renal perfusion. We adjusted for diuretic dosage, baseline renal function, and changes in renal function during hospitalization. Acid–base balance is principally regulated by the kidneys, the renin– angiotensin–aldosterone system (RAAS), and the lungs (via alterations in ventilation). Chronically elevated aldosterone levels are associated with metabolic alkalosis, and high-dose diuretics acutely increase serum aldosterone levels. Daily furosemide dose also correlates with plasma aldosterone (r = 0.77) [6]. We speculate that patients with greater changes in HCO3 levels had correspondingly greater activation of the RAAS. Since the magnitude of RAAS activation correlates with HF severity and prognosis, we suspect that this mechanism likely contributed to our observations [6–9]. Although confirmation is needed, our findings suggest that monitoring change in HCO3 levels during hospitalization for ADHF could facilitate identification of patients at increased risk for readmission or death. Thus, absolute change in HCO3 during admission may represent
675
a novel biomarker for predicting subsequent outcomes. Given the routine availability of HCO3 levels in patients admitted with ADHF and its very low acquisition cost, additional investigation of its prognostic utility is warranted.
Conflict of interest The authors report no relationships that could be construed as a conflict of interest.
References [1] Lee DS, Austin PC, Rouleau JL, Liu PP, Naimark D, Tu JV. Predicting mortality among patients hospitalized for heart failure: derivation and validation of a clinical model. JAMA 2003;290:2581–7. [2] Levy WC, Mozaffarian D, Linker DT, Sutradhar SC, Anker SD, Cropp AB, et al. The Seattle Heart Failure Model: prediction of survival in heart failure. Circulation 2006; 113:1424–33. [3] Francis GS, Benedict C, Johnstone DE, Kirlin PC, Nicklas J, Liang CS, et al. Comparison of neuroendocrine activation in patients with left ventricular dysfunction with and without congestive heart failure. A substudy of the Studies of Left Ventricular Dysfunction (SOLVD). Circulation 1990;82:1724–9. [4] Milionis HJ, Alexandrides GE, Liberopoulos EN, Bairaktari ET, Goudevenos J, Elisaf MS. Hypomagnesemia and concurrent acid–base and electrolyte abnormalities in patients with congestive heart failure. Eur J Heart Fail 2002;4:167–73. [5] Packer M. Pathophysiology of chronic heart failure. Lancet 1992;340:88–92. [6] Eshaghian S, Horwich TB, Fonarow GC. Relation of loop diuretic dose to mortality in advanced heart failure. Am J Cardiol 2006;97:1759–64. [7] Knight RK, Miall PA, Hawkins LA, Dacombe J, Edwards CR, Hamer J. Relation of plasma aldosterone concentration to diuretic treatment in patients with severe heart disease. Br Heart J 1979;42:316–25. [8] Swedberg K, Eneroth P, Kjekshus J, Wilhelmsen L. Hormones regulating cardiovascular function in patients with severe congestive heart failure and their relation to mortality. CONSENSUS Trial Study Group. Circulation 1990;82:1730–6. [9] Girerd N, Pang PS, Swedberg K, Fought A, Kwasny MJ, Subacius H, et al. Serum aldosterone is associated with mortality and re-hospitalization in patients with reduced ejection fraction hospitalized for acute heart failure: analysis from the EVEREST trial. Eur J Heart Fail 2013;15:1228–35.
