Limitation of Excessive Extracellular Matrix Turnover May Contribute to Survival Benefit of Spironolactone Therapy in Patients With Congestive Heart Failure
Insights From the Randomized Aldactone Evaluation Study (RALES)
Background—In congestive heart failure (CHF), extracellular matrix turnover is a major determinant of cardiac remodeling. It has been suggested that spironolactone may decrease cardiac fibrosis. We investigated the interactions between serum markers of cardiac fibrosis and the effect of spironolactone on outcome in patients with CHF.
Methods and Results—A sample of 261 patients from the Randomized Aldactone Evaluation Study (RALES) were randomized to placebo or spironolactone (12.5 to 50 mg daily). Serum procollagen type I carboxy-terminal peptide, procollagen type I amino-terminal peptide, and procollagen type III amino-terminal peptide (PIIINP) were assessed at baseline and at 6 months. Baseline PIIINP >3.85 μg/L was associated with an increased risk of death (relative risk [RR] 2.36, 95% CI 1.34 to 4.18) and of death+hospitalization (RR 1.83, 95% CI 1.18 to 2.83). At 6 months, markers decreased in the spironolactone group but remained unchanged in the placebo group. The spironolactone effect on outcome was significant only in patients with above-median baseline levels of markers. RR (95% CI) values for death among patients receiving spironolactone were 0.44 (0.26 to 0.75) and 1.11 (0.66 to 1.88) in subgroups of PIIINP levels above and below the median, respectively. Similarly, RR (95% CI) values for death+hospitalization among patients receiving spironolactone were 0.45 (0.29 to 0.71) and 0.85 (0.55 to 1.33), respectively.
Conclusions—In patients with CHF, high baseline serum levels of markers of cardiac fibrosis synthesis are significantly associated with poor outcome and decrease during spironolactone therapy. The benefit from spironolactone was associated with higher levels of collagen synthesis markers. These results suggest that limitation of the excessive extracellular matrix turnover may be one of the various extrarenal mechanisms contributing to the beneficial effect of spironolactone in patients with CHF.
The Randomized Aldactone Evaluation Study (RALES), a placebo-controlled randomized trial, has demonstrated that the blockade of aldosterone receptors by spironolactone (25 mg daily), in addition to standard therapy including ACE inhibitors, significantly reduces the risk of both morbidity and death among patients with severe congestive heart failure (CHF) due to left ventricular systolic dysfunction.1 The mechanism of these beneficial effects is uncertain. Spironolactone may be effective because it opposes the effects of aldosterone on sodium retention, loss of magnesium and potassium, sympathetic activation, parasympathetic inhibition, baroreceptor function, vascular damage, and/or arterial compliance.2 3 4 Spironolactone may also oppose the effect of aldosterone in promoting cardiac fibrosis.5 6 7 Extracellular matrix (ECM) turnover is an essential process found in cardiac remodeling in hypertensive cardiac hypertrophy, in dilated cardiomyopathy, and after myocardial infarction.8 Cardiac fibrosis is a major determinant of diastolic function and pumping capacity, and it may provide the structural substrate for arrhythmogenicity,8 thus potentially contributing to the progression of heart failure and to sudden death. Measurement of cardiac collagen turnover by use of serological markers is a useful tool for monitoring cardiac tissue repair and fibrosis9 in experimental models10 and in clinical conditions.11 12 13 14 15 16
The purpose of the present study was to examine, in a subgroup of patients with CHF randomized in the RALES trial, the value of these serological markers for cardiac ECM turnover as prognostic predictors of death and hospitalization, as well as the effects of spironolactone on these markers and the interaction between spironolactone-induced changes of these markers with morbidity and mortality.
Patients and Study Design
A sample of 261 patients randomized in French centers into the RALES trial1 participated in this substudy. RALES included patients with severe chronic heart failure (history of New York Heart Association [NYHA] class IV within 6 months and NYHA class III or IV at randomization, left ventricular ejection fraction <35%). Patients were randomized to placebo or spironolactone (12.5 to 50 mg daily) added to conventional therapy.
Blood samples were drawn at baseline and 6 months after randomization. Serum samples were stored at −20°C until assay. All baseline and 6-month follow-up samples were analyzed at once. Technicians were blinded to the order of the sample. No changes were observed in the samples analyzed twice. Using a radioimmunoassay from commercially available kits (Orion Diagnostica), we measured 3 ECM serum markers: procollagen type III amino-terminal peptide (PIIINP), procollagen type I carboxy-terminal peptide (PICP), and procollagen type I amino-terminal peptide (PINP). Determination of 3 degradation markers (total membrane metalloproteinase 1 [MMP1], total tissue inhibitor of metalloproteinase 1 [TIMP1], and the MMP1/TIMP1 complex) was performed by ELISA kits (Amersham). The serum levels of free MMP1 and free TIMP1 were calculated after subtracting the values of the MMP1/TIMP1 complex from the values of total MMP1 and total TIMP1, respectively. Interassay and intra-assay variations ranged from 3% to 13% for all variables. The sensitivity (lowest concentration different from zero) was 0.2 ng/mL for PIIINP, 2.0 ng/mL for PINP, 1.2 ng/mL for PICP, 1.7 ng/mL for MMP1, 1.2 ng/mL for TIMP1, and 1.5 ng/mL for the MMP1/TIMP1 complex.17 18 19 20
Baseline variables collected at the time of randomization consisted of the following: sociodemographic (age, sex, and race), clinical (etiology, NYHA class, weight, heart rate, and blood pressure), chronic degenerative comorbidity (bone or joint disease, such as rheumatoid arthritis, arthropathy, osteoporosis, and multiple myeloma; other fibrotic disease, such as various cancers, renal failure, pulmonary fibrosis, and liver cirrhosis; and diabetes), laboratory (serum sodium, serum potassium, and serum creatinine), hemodynamic (left ventricular ejection fraction), and concomitant medication data (diuretics, ACE inhibitors, β-blockers, digitalis, aspirin, potassium supplement, calcium blockers, and corticoids) (Table 1⇓). The survival follow-up period extended from the day of randomization to the end-point date of August 24, 1998. Mean follow-up was 24 months.
Statistical analysis consisted of (1) a descriptive analysis (values are expressed as mean±SD), (2) a study of association between baseline characteristics and baseline ECM turnover serum markers values, (3) a study of serum marker changes from baseline to 6 months (ANOVA with repeated-measures models), (4) an evaluation of baseline serum markers as predictors of survival and CHF hospitalization–free survival,1 and (5) testing the relationship between baseline ECM turnover serum markers levels and survival benefit from spironolactone.
For intergroup comparisons, we used ad hoc methods (univariate analysis, such as Pearson χ2 test, Mann-Whitney test, ANOVA, or correlation; multivariate analysis, such as the multiple linear regression model or the logistic regression model). Survival rates and CHF hospitalization–free survival rates were estimated by use of Kaplan-Meier analyses. All variables listed above were tested with respect to their relation to survival and CHF hospitalization–free survival by use of a univariate Mantel-Cox analysis. Variables significantly related to survival or CHF hospitalization–free survival were entered into a Cox multivariate model, adjusted for other prognostic variables. Associations between markers and survival or CHF hospitalization–free survival for each randomization group were analyzed in 3 different ways, ie, by using marker values as a continuous variable, as a categorical variable with a cutoff value as the median of distribution, and after a receiver operating characteristic–like analysis.
All analyses were performed by use of BMDP© software version 7.0 (1993).
Of the 261 patients included in the present study, 133 were in the placebo group, and 129 were in the spironolactone group. Mean age, sex, etiology, left ventricular ejection fraction, and cardiovascular drug therapy at baseline were similar in both groups (Table 1⇑). Diabetes tended to be more frequent in the placebo group; however, this was not statistically significant (P=0.054).
Baseline Levels of ECM Turnover Markers
In the whole population sample of 261 patients, baseline levels (mean±SD) were 5.0±2.5 μg/L for PIIINP, 138.8±88.2 μg/L for PICP, 45.6±37.2 μg/L for PINP, 1.6±1.5 μg/L for free MMP1, and 600.9±516.8 μg/L for free TIMP1. There was no difference in mean baseline levels between the placebo and the spironolactone groups (Table 2⇓). Patients with chronic degenerative disease (bone disease, other inflammations, and diabetes) had significantly higher baseline levels of ≥1 collagen serum marker. Multivariate analysis showed that neither sex nor baseline NYHA class nor left ventricular ejection fraction (Figure 1⇓) affected baseline ECM turnover markers. For example, baseline PIIINP was 4.9 μg/L in NYHA class III patients and 5.1 μg/L in class IV patients (P=0.71). These values were 4.7 and 5.1 μg/L, respectively, in patients with a baseline furosemide dose below and above median (60 mg) (P=0.21). However, ischemic heart disease (5.3 versus 4.7 μg/L for ischemic versus nonischemic, P=0.02) and baseline treatment with digoxin (4.7 versus 5.4 μg/L for on digoxin versus off digoxin, P=0.001) were independently associated with baseline PIIINP levels.
Baseline to 6-Month Changes
In 111 patients, a blood sample could not be obtained at 6 months (36 deaths and 75 missing samples). In 151 patients, serum ECM markers could be assessed both at baseline and 6 months after randomization (36 deaths, 75 missing samples). Their baseline characteristics (including ECM turnover marker levels) did not differ from those of the initial group of patients. From baseline to 6 months, PINP and PIIINP decreased only in the spironolactone group. We analyzed changes in markers in subgroups with high (above-median) and low (below-median) levels. When PICP, PINP, and PIIINP baseline levels were above median, they did not change significantly over 6 months in the placebo group, but they decreased significantly in the spironolactone group (Figure 2⇓). Free MMP1 and free TIMP1 levels did not change over the 6-month follow-up period in either treatment group.
Overall, survival and CHF hospitalization–free survival were poor (1-year rates were 74.6% and 57.3%, respectively) and were higher in the spironolactone group than in the placebo group (76.6% and 63.3% versus 72.7% and 51.5%, respectively). Because there was a significant interaction among baseline ECM turnover markers, treatment groups, and outcome, analysis of the prognostic significance of ECM turnover markers was performed separately in the placebo and the spironolactone groups.
Baseline PIIINP, with a cutoff value (receiver operating characteristic analysis) of 3.85 μg/L, had a significant independent negative correlation with survival and CHF hospitalization–free survival in the placebo group. Patients with baseline PIIINP >3.85 μg/L, compared with patients with PIIINP <3.85 μg/L, had a relative risk of death of 2.36 (95% CI 1.34 to 4.18, P=0.003) and a relative risk of death and/or CHF hospitalization of 1.83 (95% CI 1.18 to 2.83, P=0.007). In the 81 patients (61.8%) with baseline PIIINP levels >3.85 μg/L, survival and CHF hospitalization–free survival were 69.1% and 44.3%, respectively, compared with 79.6% and 63.4%, respectively, in the remaining 50 patients with levels <3.85 μg/L (Figure 3⇓). In contrast, in the spironolactone group, baseline PIIINP levels were not associated with survival or with CHF hospitalization–free survival. When the median value is considered as a cutoff point, the results were similar (Figure 4⇓).
Reduction of Risk of Death and of Risk of Death or CHF Hospitalization With Spironolactone According to Baseline Levels of ECM Turnover Markers
In the RALES main study, the reduction of the risk of death among patients in the spironolactone group was 30% overall. This result was similar in all prespecified and retrospective analyses performed according to age, left ventricular ejection fraction, cause of heart failure, median creatinine, median potassium, use of digitalis, ACE inhibitors, and β-blockers, and sex.1 In the present substudy, analyzing the data according to baseline ECM turnover markers revealed that the survival benefit was most predominant in subgroups with PICP, PINP, and PIIINP baseline levels above median (Figures 4⇑ and 5⇓). Analyses of the effect on death and/or hospitalization revealed similar results (Table 3⇓).
We found that high serum levels of several markers for cardiac fibrosis in patients with CHF receiving conventional therapy, including ACE inhibitors, were associated with high mortality and hospitalization rates. In patients randomized to placebo, markers continued to increase or remained unchanged after a 6-month follow-up. On the contrary, adding spironolactone (25 mg daily) significantly decreased the levels of these serum markers during the same follow-up period. Most important, the spironolactone-related morbidity and mortality benefit was most predominant in subgroups with the highest baseline levels of PICP, PINP, and PIIINP.
There is now accumulating evidence to show that serum levels of procollagen peptide fragments and of metalloproteinases can be used as markers for cardiac collagen turnover. PIIINP is the most frequently studied marker. Collagen scar formation after acute myocardial infarction causing left ventricular dysfunction could be quantified by measurements of serum PIIINP concentrations.11 PIIINP levels are also raised chronically in patients with hypertensive left ventricular hypertrophy as well as in patients with dilated cardiomyopathy. The average baseline levels of PIIINP in our patients were similar to average baseline levels reported in patients after acute myocardial infarction (5.08±0.36 μg/L)11 and in patients with CHF due to dilated cardiomyopathy (6.1±0.4 μg/L)14 receiving conventional therapy, including ACE inhibitors.
The validation of the studied serum markers as indicators of cardiac ECM turnover has been reported in experimental models.19 21 22 23 24 Because intact PINP, PICP, and PIIINP are liberated during collagen biosynthesis, it is possible to use them as markers of this process.9 25 26 Recently, Querejeta et al16 showed a strong correlation between myocardial collagen content and the serum concentration of PICP in hypertensive patients. Changes in PIIINP have been shown to be induced by acute myocardial infarction in humans13 and may reflect both synthesis and degradation of collagen.27 Elevated baseline levels of synthesis markers in our CHF patients and, most important, in the ischemic CHF subgroup for PIIINP are consistent with the concept of fibrosis being a “dynamic tissue,” as reviewed by Sun and Weber.28 According to this concept, collagen synthesis is an ongoing process involving a persistent population of metabolically active myofibroblasts nourished by a neovasculature. This is in contrast to previous concepts of fibrous tissue as acellular inert tissue. Indeed, active myofibroblasts may be found in the infarcted human heart as late as 17 years after acute myocardial infarction,29 and type I collagen mRNA expression could still be increased in the infarcted rat heart as late as 90 days after myocardial infarction.30
Prognostic Significance of Serum Levels of ECM Turnover
In patients with acute myocardial infarction, PICP, PINP, and PIIINP serum levels were shown to be correlated with infarct size, left ventricular dysfunction, and the presence of coronary artery occlusion.11 In another study, PIIINP levels on admission and for the few days after acute myocardial infarction were found to be higher in patients with poor outcome.13 In patients with CHF from idiopathic or ischemic dilated cardiomyopathy, serum PIIINP levels were independent predictors of mortality.14 Our results confirm the prognostic value of PIIINP serum levels in a large group of patients with CHF receiving conventional therapy, including an ACE inhibitor. All 3 markers were independently associated with an increased risk of death. Furthermore, our results are the first to describe a correlation between these markers and the risk of hospitalization for heart failure. Interestingly, patients with ischemic heart disease had higher levels of PIIINP than did patients with CHF of a nonischemic cause. This may be a further indication of the prognostic value of PIIINP, inasmuch as mortality is usually higher in patients with CHF of an ischemic origin.31 On the other hand, we found that patients chronically treated with digitalis had a lower level of PIIINP. This is consistent with the finding that digoxin reduces death due to the progression of heart failure.32
Factors involved in collagen degradation (MMP1 and TIMP1) were not associated with the risk of death. It may be that ECM turnover in CHF is more directed to a higher synthesis rate, as assessed by the concentrations of circulating procollagen peptides, than to an altered collagen degradation. Also, other degradation products may be more relevant and need further investigation. The rate-limiting step in the extracellular degradation of collagen is MMP1, which accounts for the degradation of up to 40% of the newly synthesized collagen in different tissues.33 The net level of MMP1 activity is dependent on the relative concentrations of active enzyme and a family of naturally occurring tissue inhibitors of metalloproteinases, namely, TIMP1.34 To our knowledge, there is no clinical report of free MMP1 and free TIMP1 assessment in patients with heart failure or acute myocardial infarction.
Effects of Spironolactone on Serum Levels of Collagen Markers
Aldosterone has been shown to promote cardiac fibrosis in various experimental models.7 35 36 37 The temporal cellular response and appearance of myocardial fibrosis associated with chronic elevation of angiotensin II and/or aldosterone differ, indicating that separate pathogenic mechanisms are operative with these effector hormones of the renin-angiotensin-aldosterone system.6 In several reports, it has been demonstrated that spironolactone may oppose the effect of aldosterone in promoting cardiac fibrosis.5 6 7 38 In a small clinical study, a high dose of spironolactone (50 to 100 mg/d) produced a significant decrease of PIIINP serum levels in 21 patients with CHF.39 Our results are the first to report a significant effect of a low dose of spironolactone (26 mg/d on average) on several serum markers of collagen and to relate this effect to survival benefit in a large group of patients. This finding may be interpreted as a result of the limitation by spironolactone of collagen synthesis in the failing heart.
The regulation of the key enzyme of collagen degradation, MMP1, which has been identified in the heart, is largely unknown. MMP1 activity in a cultured cardiac fibroblast preparation was not influenced by aldosterone.40 Because collagen accumulation in the myocardium represents the balance between collagen synthesis and degradation, aldosterone appears to lead to a net accumulation of collagen, which is inhibited by spironolactone. The effects of spironolactone on collagen synthesis were most predominant in patients with the highest above-median baseline levels of PIIINP.
Relationship Between Effects of Spironolactone on Collagen Markers and on Patients’ Outcomes
In the present study, retrospective subgroup analysis showed that the beneficial effects of spironolactone on survival and hospital admission were almost primarily clustered within the subgroup of patients with the highest pretreatment levels of PICP, PINP, or PIIINP. All other prespecified and retrospective subgroup analyses in the main RALES1 failed to characterize any demographic, clinical, or laboratory variable that would influence the clinical benefit of spironolactone therapy. The demonstrated prognostic importance of serum levels of the collagen markers in the present study and the significant decrease of these levels in the spironolactone treated group of patients strongly suggest that limitation of the aldosterone-stimulated collagen synthesis may be one of the various extrarenal mechanisms contributing to the clinical benefit of spironolactone in the RALES trial.
Limitation of the Present Study
Some limitations of the present study should be acknowledged. This substudy was performed in a subset of patients from the RALES trial. Although the major characteristics of our patients were comparable to those of the overall RALES patient population, extrapolation of our results to all patients in the trial should be made with caution.
Blood samples at the 6-month follow-up could not be obtained from all patients in the substudy because of missing samples and early deaths. However, the changes of the serum levels of collagen markers over 6 months were analyzed by paired data analysis, ie, analysis only in patients with available baseline and 6-month data.
The interaction analysis between baseline markers levels, treatment group, and outcome was a retrospective analysis in nonprespecified subgroups. Therefore, one cannot exclude that our result may be a chance finding. Nevertheless, we believe that the coherence of the results of analyses of the prognostic values of serum markers and their respective congruent change over time in the spironolactone group and the placebo group give credit to the main result of the interaction analysis.
For the first time, we observe that serum levels of markers of cardiac collagen synthesis were significantly associated with poor outcome in patients with CHF and could be decreased by aldosterone receptor blockade with spironolactone. The morbidity and mortality benefit from spironolactone is predominant in patients with the highest levels of markers. These results, which require further prospective confirmation, suggest that limitation of the aldosterone-related excessive ECM turnover may be one of the various extrarenal mechanisms contributing to the beneficial effect of spironolactone in patients with CHF.
Dr Alfonso Perez is an employee of Searle.
- Received May 18, 2000.
- Revision received July 11, 2000.
- Accepted July 14, 2000.
- Copyright © 2000 by American Heart Association
Pitt B, Zannad F, Remme WJ, et al. The effect of spironolactone on morbidity and mortality in patients with sever heart failure: Randomized Aldactone Evaluation Study Investigators. N Engl J Med. 1999;341:709–717.
Weber KT. Aldosterone and spironolactone in heart failure. N Engl J Med. 1999;341:753–754.
Zannad F. Angiotensin-converting enzyme inhibitor and spironolactone combination therapy: new objectives in congestive heart failure treatment. Am J Cardiol. 1993;71:34A–39A.
Zannad F. Aldosterone and heart failure. Eur Heart J. 1995;16(suppl):98–102.
Brilla CG, Matsubara LS, Weber KT. Anti-aldosterone treatment and the prevention of myocardial fibrosis in primary and secondary hyperaldosteronism. J Mol Cell Cardiol. 1993;25:563–575.
Lijnen P, Petrov V. Antagonism of the renin-angiotensin-aldosterone system and collagen metabolism in cardiac fibroblasts. Methods Find Exp Clin Pharmacol. 1999;21:215–227.
Fullerton MJ, Funder JW. Aldosterone and cardiac fibrosis: in vitro studies. Cardiovasc Res. 2000;1994:1863–1867.
Swynghedauw B. Molecular mechanisms of myocardial remodeling. Physiol Rev. 1999;79:215–262.
Weber KT. Monitoring tissue repair and fibrosis from a distance. Circulation. 1997;96:2488–2492.
Laviades C, Varo N, Fernandez J, et al. Abnormalities of the extracellular degradation of collagen type I in essential hypertension. Circulation. 1998;98:535–540.
Uusimaa P, Risteli J, Niemela M, et al. Collagen scar formation after acute myocardial infarction: relationships to infarct size, left ventricular function, and coronary artery patency. Circulation. 1997;96:2565–2572.
Sato Y, Kataoka K, Matsumori A, et al. Measuring serum aminoterminal type III procollagen peptide, 7S domain of type IV collagen, and cardiac troponin T in patients with idiopathic dilated cardiomyopathy and secondary cardiomyopathy. Heart. 1997;78:505–508.
Host NB, Jensen LT, Bendixen PM, et al. The aminoterminal propeptide of type III procollagen provides new information on prognosis after acute myocardial infarction. Am J Cardiol. 1995;76:869–873.
Klappacher G, Franzen P, Haab D, et al. Measuring extracellular matrix turnover in the serum of patients with idiopathic or ischemic dilated cardiomyopathy and impact on diagnosis and prognosis. Am J Cardiol.. 1995;75:913–918.
Diez J, Laviades C, Mayor G, et al. Increased serum concentration of procollagen peptides in essential hypertension: relation to cardiac alteration. Circulation. 1995;91:1450–1456.
Querejeta R, Varo N, Lopez B, et al. Serum carboxy-terminal propeptide of procollagen type I is a marker of myocardial fibrosis in hypertensive heart disease. Circulation. 2000;201:1729–1735.
Lein M, Nowak L, Jung K, et al. Analytical aspects regarding the measurement of metalloproteinases and their inhibitors in blood. Clin Biochem. 1997;30:491–496.
Melkko J, Kauppila S, Niemi S, et al. Immunoassay for intact amino-terminal propeptide of human type I procollagen. Clin Chem. 1996;42:947–954.
Melkko J, Niemi S, Risteli L, et al. Radioimmunoassay of the carboxyterminal propeptide of human type I procollagen. Clin Chem. 1990;36:1328–1332.
Risteli J, Niemi S, Trivedi P, et al. Rapid equilibrium radioimmunoassay for the amino-terminal propeptide of human type III procollagen. Clin Chem. 1988;34:715–718.
Diez J, Laviades C. Monitoring fibrillar collagen turnover in hypertensive heart disease. Cardiovasc Res. 1997;35:202–205.
Diez J, Panizo A, Gil MJ, et al. Serum markers of collagen type I metabolism in spontaneously hypertensive rats. Circulation. 1996;93:1026–1032.
Sekita S, Katagiri T, Sasai Y, et al. Studies on collagen in the experimental myocardial infarction. Jpn Circ J. 1985;49:171–178.
Whittaker P, Boughner DR, Kloner RA. Analysis of healing after myocardial infarction using polarized light microscopy. Am J Pathol. 1989;134:879–893.
Risteli J, Risteli L. Analysing connective tissue metabolites in human serum: biochemical, physiological and methodological aspects. J Hepatol. 1995;22(suppl 2):77–81.
Jensen LT, Horslev-Petersen K, Toft P, et al. Serum aminoterminal type III procollagen peptide reflects repair after myocardial infarction. Circulation. 1990;81:52–57.
Peuhkurinen KJ, Risteli L, Melkko J, et al. Thrombolytic therapy with streptokinase stimulates collagen breakdown. Circulation. 1991;83:1969–1975.
Sun Y, Weber KT. Infarct scar: a dynamic tissue. Cardiovasc Res. 2000;46:250–256.
Willems IE, Havenith MG, De Mey JG, et al. The alpha-smooth muscle actin-positive cells in healing human myocardial scars. Am J Pathol. 1994;145:868–875.
Cleutjens JP, Verluyten MJ, Smiths JF, et al. Collagen remodeling after myocardial infarction in the rat heart. Am J Pathol. 1995;147:325–338.
Zannad F, Briancon S, Juilliere Y, et al. Incidence, clinical and etiologic features, and outcomes of advanced chronic heart failure: the EPICAL Study: Epidémiologie de l’Insuffisance Cardiaque Avancée en Lorraine. J Am Coll Cardiol. 1999;34:734–742.
Yusuf S. Digoxin in heart failure: results of the recent Digoxin Investigation Group trial in the context of other treatments for heart failure. Eur Heart J. 1997;18:1685–1688.
Woessner JF Jr. Matrix metalloproteinases and their inhibitors in connective tissue remodeling. FASEB J. 1991;5:2145–2154.
Denhardt DT, Feng B, Edwards DR, et al. Tissue inhibitor of metalloproteinases (TIMP, aka EPA): structure, control of expression and biological functions. Pharmacol Ther. 1993;59:329–341.
Campbell SE, Janicki JS, Weber KT. Temporal differences in fibroblast proliferation and phenotype expression in response to chronic administration of angiotensin II or aldosterone. J Mol Cell Cardiol. 1995;27:1545–1560.
Robert V, Van TN, Cheav SL, et al. Increased cardiac types I and III collagen mRNAs in aldosterone-salt hypertension. Hypertension. 1994;24:30–36.
Young M, Fullerton MJ, Dilley R, et al. Mineralocorticoids, hypertension and cardiac fibrosis. J Clin Invest. 1994;93:2578–2583.
Brilla CG, Weber KT. Mineralocorticoid excess, dietary sodium and myocardial fibrosis. J Lab Clin Med. 1992;120:893–901.
MacFadyen RJ, Barr CS, Struthers AD. Aldosterone blockade reduces vascular collagen turnover, improves heart rate variability and reduces early morning rise in heart rate in heart failure patients. Cardiovasc Res. 1997;35:30–34.
Brilla CG, Maisch B, Weber K. Renin-angiotensin system and myocardial collagen matrix remodeling in hypertensive heart disease: in vivo and in vitro studies on collagen matrix regulation. Clin Invest. 1993;71:S35–S41.