Eplerenone in systemic right ventricle: Double blind randomized clinical trial. The evedes study☆,☆☆
Abstract
Background: There is no proven pharmacological strategy for the treatment of the failing systemic right ventricle (SRV) but myocardial fibrosis may play a role in its pathophysiology.
Methods: We designed a double-blind, placebo-controlled clinical trial to assess the effects of eplerenone 50 mg during 12 months on cardiac magnetic resonance parameters (SRV mass and ejection fraction) and neurohormonal and collagen turnover biomarker (CTB) levels.
Results: Twenty six patients with atrial switch repair for transposition of the great arteries were randomized to eplerenone (n = 14) or placebo (n = 12) and 14 healthy volunteers served as controls for comparison of baseline neurohormones and CTB levels. The study population showed a good baseline profile in terms of SRV mass (57.4 ± 17 g/m2) and ejection fraction (54.9 ± 7.5%). However, levels of N-terminal pro-brain natriuretic peptide (NT-proBNP), C terminal propeptide of type I procollagen (CICP) and C-terminal Telopeptide of type I Collagen (ICTP) were significantly elevated when compared to healthy controls. After one year of treatment, a trend toward reduction of CICP, N-terminal pro-Matrix Metalloproteinase 1 (NT-proMMP1), Tissue Inhibitor of Metalloprotein- ases 1 (TIMP1) and galectin 3 levels and a lower increase in ICTP in patients under eplerenone was observed. The reduction of SRV mass and the improvement of SRV function with eplerenone were not conclusive.
Conclusions: Patients with SRV treated with eplerenone showed an improvement of an altered baseline CTB profile suggesting that reduction of myocardial fibrosis might be a therapeutic target in these patients.
1. Introduction
Although the current surgical procedure of choice for the repair of the transposition of the great arteries (TGA) is the arterial switch, there is a considerable number of adult patients followed up at the con- genital cardiac clinics that underwent the atrial switch procedure (AtS) during childhood. Despite the overall good survival of these patients into adulthood, the majority leading a normal life [1–4], heart failure has been identified as a major problem in the long term [5,6] since the morpholog- ic right ventricle is on the systemic position. However, when it comes to the management of the failing systemic right ventricle (SRV), questions arise. The promising results offered by β-blockers in some studies are far from conclusive [5,7–9] and the use of angiotensin-converting en- zyme inhibitors (ACEi) and angiotensin II receptor blockers has brought divergent results [10–15].
Myocardial fibrosis has been identified in patients with SRV by cardiac magnetic resonance (CMR) and it seems to correlate with poorer ventricular function and worse prognosis [16,17]. Aldosterone, one of the most widely studied neurohormones in systemic left ventricu- lar remodeling and failure is known to stimulate the mineralocorticoid receptor in cardiomyocytes, endothelial cells and fibroblasts induc- ing myocardial and vascular fibrosis.
This inflammatory response is a direct humoral effect, rather than secondary to hemodynamic changes [18–21]. Experimental studies have demonstrated that, regardless the dose, treatment with aldosterone antagonists can reduce myocardial fibrosis [19–23] and even reduce ventricular hypertrophy in hypertensive patients presumably due to reduction of fibrotic tissue [24]. Our study was therefore designed to assess the effects of eplerenone, a selective aldosterone antagonist, on myocardial mass and fibrosis of systemic right ventricles.
2. Methods
2.1. Study population
This study started as a collaborative study between the Adult Congenital Cardiac Units of Vall d’Hebron University Hospital and Santa Creu i Sant Pau University Hospital. By the end of the study both units merged into a single tertiary referral which is Adult Congenital Heart Disease Unit for the Autonomous Community of Catalonia. The protocol was ap- proved by the Ethics Review Board of both centers and written informed consent was obtained from each patient before enrollment. This investigation conforms to the princi- ples outlined in the Declaration of Helsinki. The authors of this manuscript have certified that they comply with the Principles of Ethical Publishing in the International Journal of Cardiology.
The study population comprised of all adult (≥18 years of age) patients diagnosed with TGA that underwent AtS repair during childhood and are actively followed up at our clinics. The patients were excluded if they were taking aldosterone antagonists for the preceding 6 months, had creatinine levels N 1.5 mg/dL or potassium levels N 5.0 mmol/L at baseline blood test, had hypersensitivity to the study medication, were pregnant or breast feeding, were unable to undergo CMR (i.e., claustrophobia, pacemaker or implantable cardioverter defibrillator leads), were on heart transplant list, had a life expectancy b1 year or refused to sign informed consent. Conditions known to alter collagen turnover, such as liver cirrhosis, renal impairment, pulmonary fibrosis, metabolic bone disease, connective tissue diseases, malignancy, chronic inflammatory disease, ongoing treatment with corticosteroids or recent surgery, were excluded in participants.
2.2. Study protocol
We conducted a collaborative bicenter, double-blind, placebo-controlled clinical trial. At baseline, the patients underwent physical examination, electrocardiogram, transthoracic echocardiogram, maximal symptom-limited cardiopulmonary exercise testing (CPX), CMR and laboratory testing including neurohormones and collagen turnover biomarkers (CTB) (see online Supplementary data for examinations of protocol details). The patients were then randomly assigned to receive eplerenone (25 mg daily) or placebo for two weeks after which the dose of eplerenone was increased to a maximum of 50 mg per day provided that no side effects or hyperpotassemia was detected. Randomization was centrally performed by the Pharmacy Department of Vall d’Hebron Hospital in blocks of 10 (5 + 5) and stratified by baseline treatment with ACEi. Information regarding assigned arm was provided to investigators through a phone call to the Pharmacy Department.
Follow-up visits were scheduled at two and four weeks, three months and every three months thereafter until completion of the study. The renal function and serum potassium levels were measured one week after the initiation of the study treatment, at all scheduled clinical visits and after any change of dose. At completion of the protocol, all the examina- tions performed at baseline were repeated. If at any time during the study process the serum potassium concentration was N 5.5 mmol/L, the study medication was halved or temporarily discontinued until the serum potassium concentration fell below 5.5 mmol/L. In addition, blood samples were also obtained from a control cohort of age- and sex- matched healthy volunteers (n = 14) for comparison of baseline levels of neurohormones and CTB.
2.3. Statistical analysis
The primary endpoint was the difference between both groups in the change from baseline in SRV mass as measured by CMR. Secondary endpoints were the difference be- tween groups in change in SRV ejection fraction and gadolinium enhancement mass as measured by CMR and change in neurohormones and CTB levels.
Based on the study of Hornung et al. [25] we assumed a mean baseline right ventricular mass of 97 g/m2 (SD 25 g/m2). Assuming an absolute reduction of right ventricular mass of 15 g/m2 as clinically relevant, 39 patients per arm would be necessary to detect such a dif- ference between groups as statistically significant considering a two-side hypothesis (type I error 0.05) with a statistical power of 80% (type II error 0.2). Considering 40% of eligible patients with at least one exclusion criterion (including refusal to participate) and 10% of therapy discontinuations, the estimated sample of eligible patients was 117 patients.
The quantitative variables are presented as means and standard deviations. For variables in which normal distribution could not be demonstrated, median and interquartile range (Q1Q3) are provided. Discrete variables are described as absolute counts and percentages. Differences between study groups were compared using the chi-square test, Mann–Whitney test, Fisher’s exact test or Pearson test where appropriate. Differences between baseline and final study visits within study groups concerning quantitative variables were tested by McNemar’s test. The differences between groups in the mean changes between baseline and final study visits were compared using Mantel–Haenszel test. All analyses were performed by intention to treat.
Differences in baseline levels of neurohormones and collagen turnover biomarkers between the global sample and 14 healthy controls were calculated by Mann–Whitney test. Correlation analysis among CMR parameters, neurohormones and collagen turnover biomarkers was performed by Pearson test.
Given the potential influence of ACEi in the collagen turnover and the inhomogeneous distribution of patients receiving this treatment between both groups (despite stratification by baseline treatment with ACEi treatment), we conducted a subgroup analysis excluding the patients under such treatment. All the tests were considered significant for a p-value of b0.05. Statistical analysis was performed using V13 SPSS statistical package (Illinois, USA).
3. Results
During 18 months of time-frame recruitment, only 27 out of a total sample size of 110 patients signed the informed consent to participate (Fig. 1), thus leading to discontinuation because of insufficient recruit- ment. In addition, one patient withdrew informed consent before ran- domization, thus 26 patients were randomized to eplerenone (n = 14) or placebo (n = 12). We next present the clinical, cardiac MRI, neurohor- monal, and functional results of these 26 patients.
3.1. Clinical
After two weeks of treatment, one patient in the eplerenone group permanently discontinued the study medication due to dizziness not attributable to any concomitant hemodynamic or biochemical disorder. There were no further treatment discontinuations and, therefore, a total of 25 patients completed the study. There were 3 additional patients in the eplerenone group complaining of minor adverse events, namely pal- pitations, pollakiuria and pruritus which did not require treatment modifications. No adverse events were registered in the placebo group. The creatinine concentration remained within normal limits through- out the study in both groups. As for hyperpotassemia, there was one case in the eplerenone group that initially led to provisional reduction of the medication dose and finally obliged to complete the protocol two weeks ahead of schedule. This patient was concomitantly on ACEi and the hyperpotassemia (potassium concentration of 6.1 mmol/L) was resolved after removal of study medication.
There were no modifications in the reported baseline NYHA func- tional class of the majority of patients. Only one patient, in the placebo group, experienced a worsening from class I to class II.Table 1 presents the demographics and baseline characteristics of both groups. Participating patients were predominantly male, had mostly undergone a Senning procedure, were generally in NYHA func- tional class I, and only 5 of them (19%) were taking medications. As expected from a young population, the cardiovascular risk factor burden was low and no cases of diabetes or dyslipidemia were documented. De- spite stratification for treatment with ACEi, after randomization there were 3 patients treated with ACEi in the eplerenone group and 1 patient in the placebo group (21.4% vs 8.3% respectively, p = 0.6).
3.2. Cardiac MRI
Table 2 shows the baseline cardiac MRI measurements and their evolution in both groups during follow-up. Systolic function was mostly preserved since only 7 patients (27%) showed a SRV ejection fraction below 50%. Contrast-enhancement was detected in 7 patients (27%) which showed a non-significant tendency to lower RVEF (50.8 ± 7.6% vs 56.4 ± 7%, p = 0.09), and higher RV mass (65 ± 17 g/m2 vs 54 ± 17 g/m2, p = 0.16) than the rest of the population. Two late gadolinium enhancement patterns were found: 1. Focal enhancement in the RV free wall (4/7) and 2. Focal enhancement in the inferior part of the septum at the RV–LV junction (4/7), 1 patient showing extension to the diaphrag- matic wall.
Both groups were quite well balanced for all parameters except for gadolinium enhancement mass, which was higher in the eplerenone RVEDV = Right ventricular end-diastolic volume, RVEF = Right ventricular ejection fraction, LVEDV = Left ventricular end-diastolic volume, LVEF = Left ventricular ejection fraction.
3.3. Neurohormonal and collagen turnover biomarkers
Table 3 shows the baseline levels of neurohormones and collagen turnover biomarkers of the study group and of 14 age- and sex- matched healthy controls for comparison. Mean levels of all biomarkers were comparable between both treatment groups except for NT- proMMP1 which was significantly elevated in the eplerenone arm when compared to placebo at baseline (p = 0.05).
Mean levels of N-terminal B-type natriuretic peptide (NT-proBNP) were significantly higher whereas mean N-terminal Atrial Natriuretic Peptide (NT-proANP) levels were significantly lower in patients when compared to controls. On the other hand, only 3 patients (all of them taking ACEi) showed increased concentrations of plasma renin activity. As for aldosterone, no differences were found between patients and con- trols. Regarding collagen turnover biomarkers, C-terminal Propeptide of type I Procollagen (CICP) and C-terminal Telopeptide of type I Collagen (ICTP) showed significantly higher mean concentrations at baseline than healthy controls (Table 3) whereas mean galectin 3 and Tissue Inhibitor of MetalloProteinases 1 (TIMP1) levels were significantly lower and no differences were found on N-terminal pro-Matrix Metalloproteinase-1 (NT-proMMP1) concentrations.
No significant correlation was found between neurohormones or CTB and CMR measurements of SRV function, mass or end-diastolic volume. The presence of late gadolinium enhancement did not show a significant association with neurohormone levels or CTB concentration. After one year of treatment, no significant variations attributable to treatment effect were detected in the neurohormone and collagen turn- over profile, although a trend toward a reduction in CICP, NT-proMMP1, TIMP1 and galectin 3 levels and a lower increase in ICTP in patients under eplerenone was observed (Table 4). A subgroup analysis excluding patients with ACEi therapy confirmed the same trend, with NT-proMMP1 reaching statistical significance (−0.18 ± 0.21 vs −0.02 ± 0.12, p = 0.014) (Fig. 3).
3.4. Echocardiography and cardiopulmonary exercise testing
There were no hemodynamically significant residual lesions in the study population. Moderate tricuspid regurgitation (n = 2, 8%), tinny baffle leaks (n = 3, 11.5%), subpulmonar obstruction (n = 5, 19%) with a peak gradient of 50 ± 7 mm Hg and asymptomatic mild vena cava baffle obstruction (n = 2, 8%) were distributed homogeneously between both groups. Baseline estimates of SRV function are shown in online Supplementary data Table 1. After one year of treatment, there was not any relevant difference within or between groups in terms of SRV function parameters.
As for CPX, baseline parameters are also displayed in online Sup- plementary data Table 1. Patients in the eplerenone group showed significantly higher maximal oxygen uptake and oxygen pulse than the placebo group. After one year of treatment, no significant change from baseline in any of the CPX variables was detected in any of the groups and the difference in change between both groups was neither significant.
4. Discussion
Although this randomized clinical trial has to be considered as non- conclusive regarding its primary endpoint due to insufficient recruit- ment, it is the first to evaluate the effect of an aldosterone antagonist on patients with SRV. In summary, while our cohort of patients with SRV was in a good clinical situation and showed only a modest degree of ventricular hypertrophy when compared to previously published series, they exhibited a pattern of altered collagen turnover and neuro- hormonal activation. In the 13 and 12 patients that completed the one year treatment with eplerenone and placebo respectively, no apparent treatment effect was observed on CMR parameters although a trend to improvement of the CTB profile could be detected.
In the last years, there is growing evidence that myocardial fibrosis may be a pathophysiological substrate behind SRV dysfunction [26]. It has been identified in patients with SRV by CMR and it seems to corre- late with poorer ventricular function and worse prognosis [16,17]. In its turn, aldosterone, one of the most widely studied neurohormones in systemic left ventricular remodeling and failure is known to stimulate the mineralocorticoid receptor in cardiomyocytes, endothelial cells and fibroblasts inducing myocardial and vascular fibrosis. Experimental studies have demonstrated that, regardless the dose, treatment with aldosterone antagonists can reduce myocardial fibrosis [19–23]. The cur- rent study is the first clinical trial to evaluate, based on the antifibrotic properties of these drugs, the effect of an aldosterone antagonist on patients with SRV.
According to our findings, even being a cohort with much lower ventricular mass than previously published series [16,17,25] our patients with SRV exhibit a pattern of altered collagen turnover and neurohor- monal activation. Both CICP (a marker of type I collagen synthesis) an ICTP (a marker of type I collagen degradation) were significantly elevated in our patient population when compared to healthy controls without a significant difference in the ratio of CICP/ICTP between both groups. This suggests a situation of increased collagen turnover without (in a small cohort of clinically stable and mostly asymptomatic patients like ours) a detectable net collagen accumulation. Nonetheless, it has been reported that hypertensive patients show significant elevations of CICP, particularly those that exhibit severe fibrosis on endomyocardial biopsy [27]. Furthermore, high CICP concentrations have been proposed as predictor of heart failure with preserved ejection fraction in hyper- tensive patients [28]. On the other hand, a sub-study of the EPHESUS trial showed that a combination of elevated brain natriuretic peptide (also documented in our population) and ICTP at baseline was associat- ed with increased all-cause mortality and the composite endpoint of cardiovascular death and hospitalization during follow-up [29]. These results were confirmed in a subsequent study on heart failure patients of mixed etiologies [30]. In a recent study performed in elderly popula- tion ICTP has been reported as a predictor of heart failure not only in cardiac patients but also in control subjects [31]. Conversely, SRV patients showed significantly lower levels of NT-proANP than healthy volunteers. The explanation for this finding might be sought on the scar- city of remaining healthy atrial tissue after an extensive surgery during their repair in childhood. Perhaps a similar explanation pertaining the idiosyncratic nature of the atrial switch might underlie the lower con- centrations of galectin 3 and TIMP1 found in the patient population, but a definite interpretation remains elusive for the authors.
It is also noteworthy that although NT-proBNP was significantly increased in our sample, plasma renin activity (excluding patients on ACEi) and aldosterone levels were comparable to those of healthy age- and sex-matched controls. Elevated NT-proBNP levels in asymp- tomatic patients with SRV have been previously reported [32] but plasma concentrations of aldosterone and renin activity correlate with the severity of symptoms both in acquired heart disease [33] and in congenital heart disease [34]. A recent clinical trial with valsartan in patients with SRV showed normal aldosterone levels in asymptomatic patients while the subgroup in NYHA functional class ≥ 2, which actually benefited from the study treatment, had significantly higher aldosterone concentrations [15]. A previous trial with losartan in SRV [10] with 93% of patients in NYHA functional class I showed a neurohormon- al profile similar to ours (88.5% of patients in NYHA functional class I). After one year of treatment, a trend toward reduction of CICP, NT- proMMP1, TIMP1 and galectin 3 levels and a lower increase in ICTP in patients under eplerenone when compared to placebo was observed. The sub-analysis excluding patients on ACEi strengthened this ten- dency with NT-proMMP1 reaching statistical significance. A reduction of markers of collagen synthesis in patients under treatment with eplerenone has been previously reported in patients with systemic left ventricle. In the previously mentioned sub-study of the EPHESUS trial [29] (in which the addition of eplerenone to optimal medical therapy reduced morbidity and mortality among patients with acute myocardial infarction complicated by left ventricular dysfunction and heart failure), levels of aminoterminal propeptide of type I and type III procollagen were found to be significantly lower in the eplerenone group beginning at 6 months after the index hospitalization. In a recent trial on patients with heart failure with preserved ejection fraction, Deswal et al. [35] showed a reduction of procollagen type I aminoterminal peptide and ICTP along with an improvement in echocardiographic measures of diastolic function in the eplerenone group.
As for myocardial MMP1 expression (NT-proMMP1 or interstitial collagenase is an enzyme with high specificity for degradation of fibril- lar collagens [36]) it has been reported to be altered in patients with cardiac diseases associated with either fibrosis such as aortic stenosis or loss of collagen scaffold like idiopathic dilated cardiomyopathy or ischemic heart disease [37]. Serum MMP1 has been found to correlate directly with left ventricular dimensions and inversely with left ventric- ular ejection fraction in a population of hypertensive patients with sys- tolic heart failure [38]. Although this is a matter of extensive research, it seems that in early stages of pressure overload there is an inhibition of metalloproteinases favoring matrix accumulation and fibrosis. How- ever, in advanced stages there is an increase of metalloproteinases leading to matrix degradation (particularly mysial collagen matrix which provides support, geometric alignment and coordination of adja- cent cardiomyocyte fascicle contraction) with the subsequent ventricu- lar dilatation and systolic dysfunction [36,38]. In a recent study, high MMP1 and ICTP concentrations were associated with increased risk of appropriate implantable cardioverter defibrillator therapy in a popula- tion of patients with dilated cardiomyopathy [39]. Noteworthy, it has been reported that metalloproteinase inhibition attenuates ventricular enlargement and dysfunction in an animal model of spontaneously hy- pertensive rats with heart failure [40].
Whether this reduction of CTB may be considered as a guide for medical treatment or even a therapeutic target is still a matter of debate but collagen turnover biomarkers is a subject of intense study with a prolific production in the literature over the past years [36]. Moreover, all the above referenced literature is based on patients with systemic left ventricle. It remains to be proven whether all this evidence would be applicable to patients with SRV since, to the best of our knowledge, this is the first study to address the subject of CTB in patients with SRV or even in patients with congenital heart disease in general.
Regarding structural RV parameters, after one year of treatment, CMR studies showed a non-significant trend toward reduction of myocardial mass and SRV function improvement in the eplerenone group. This find- ing would be in keeping with the results of a previous study in which hypertensive patients treated with eplerenone 200 mg during 9 months significantly reduced the systemic left ventricular mass as measured by CMR [24]. However, the magnitude of the effect of eplerenone on ven- tricular mass and ejection fraction observed in our study could underes- timate the potential actual effect of eplerenone in SRV population, since our patients were in much better clinical condition than the use. In fact, comparing other series our patients had better NYHA functional class [41,42], maximal VO2 [16,43], ejection fraction [16,41] and less ventricu- lar hypertrophy [16,17,25]. In addition, the measurement of ventricular mass might not be a parameter precise enough to detect the changes attributable to the treatment in a small cohort with only a modest degree of ventricular hypertrophy. Undoubtedly, it would have been more accurate to quantify a fibrosis index as proposed by Broberg et al.
[44] but, unfortunately, this technique was not available by the time our trial was designed and could not be applied to our recorded images.
5. Limitations and strengths of the study
The main limitation of our study is the low number of patients en- rolled because of the refusal of the patients to participate. Their very good clinical situation, the fear of side-effects and a self-negation of their chronic condition likely influenced the low rate of participation despite our efforts. It has to be pointed out, however, that baseline char- acteristics of consenting patients and those who refused to participate compared well in terms of age (26.4 ± 5.4 vs 26.6 ± 4.2, p = 0.9), sex (female: 38.5% vs 35.1%, p = 0.81) and complex anatomy (11.5% vs 10.5%, p = 1).
A second limitation is the good clinical situation of our sample, which could not entirely represent the actual status of SRV population, thus limiting the external validity and probably penalizing the magni- tude of the results. The explanation to this good baseline status of our cohort when compared to others should be sought on the exclusion criteria of our study which could have led to an excessively selected patient population. It has been proven that pacemaker stimulation (one of our exclusion criteria) has a deleterious effect on functional class, exercise performance and ventricular function in patients with systemic right ventricle [45]. On the other hand, despite the lack of evidence, most of the patients with symptoms or poor ventricular func- tion were already on aldosterone antagonists by the time the study was undertaken.
Finally, the presence of myocardial fibrosis in our sample was based on previously published MRI studies [16,17,44] in patients with AtS for TGA. However, although late gadolinium enhancement has been docu- mented in our patients, no histological evidence of myocardial fibrosis has been provided.
Although this randomized clinical trial has to be considered as non- conclusive regarding its primary endpoint due to insufficient recruit- ment (the statistical power of the study to detect differences on the primary outcome between both groups was as low as 33%), it is the first to evaluate the effect of an aldosterone antagonist on patients with SRV. It could be, therefore, considered as a pilot study since several results might be useful to orient future research.
The study has also additional strengths. All patients underwent an exhaustive RV and biochemical study according to a systematic protocol. In addition, we included a cohort of age–sex matched controls for the neurohormone and collagen turnover study, which helps to put in per- spective the findings of these parameters in SRV patients.
6. Conclusion
Despite an overall baseline good situation, our cohort of patients with SRV exhibits an altered baseline CTB profile and neurohormonal activation. Although the small sample size prevents from any conclusive statement, our results suggest that patients with SRV could hypotheti- cally benefit from the treatment with eplerenone in terms of SRV hyper- trophy reduction, and SRV function and CTB profile improvement. Moreover, the reduction of myocardial fibrosis in these patients might be a therapeutic target in future research. However, further studies including larger numbers and more deteriorated patients are Olitigaltin required to prove this hypothesis.