Author + information
- Received June 17, 2013
- Revision received August 8, 2013
- Accepted September 9, 2013
- Published online February 18, 2014.
- Yoshihiro Matsumoto, MD∗∗ (, )
- Yasuo Mori, MD†,
- Shinji Kageyama, MD‡,
- Kazuo Arihara, MD§,
- Toshikazu Sugiyama, MD‖,
- Hiromichi Ohmura, MD¶,
- Toru Yakushigawa, MD†,
- Hatsumi Sugiyama, MD‖,
- Yasushi Shimada, MD∗,
- Youichi Nojima, MD∗ and
- Nobuo Shio, MD‡
- ∗Department of Nephrology and Dialysis, Shizuoka City Hospital, Shizuoka, Japan
- †Shibukawa Clinic, Shizuoka, Japan
- ‡Kageyama Urological Clinic, Shizuoka, Japan
- §Ohtemachi Clinic, Shizuoka, Japan
- ‖Sugiyama Clinic, Shizuoka, Japan
- ¶Sugawara Clinic, Shizuoka, Japan
- ↵∗Reprint requests and correspondence:
Dr. Yoshihiro Matsumoto, Department of Nephrology and Dialysis, Shizuoka City Hospital. 10-93 Ohtemachi, Aoi-ku, Shizuoka 420-8630, Japan.
Objectives This study sought to assess whether spironolactone treatment reduces the high incidence of cardiovascular and cerebrovascular (CCV) morbidity and mortality in hemodialysis (HD) patients.
Background Aldosterone receptor blockers reduce cardiac-related events, but the efficacy of the agents in HD patients is unclear.
Methods A 3-year randomized trial involving 5 clinics was performed. Of the 309 oligoanuric HD patients enrolled in the study, 157 patients were randomly assigned to receive 25 mg/day of spironolactone without any restriction on dietary potassium intake (treatment group), and 152 patients were assigned to a control group. The primary outcome was a composite of death from CCV events or hospitalization for CCV events, and the secondary outcome was death from all causes.
Results During the 3-year follow-up, the primary outcome occurred in 5.7% of patients in the treatment group and in 12.5% of patients in the control group. Hazard ratios (HRs) for the primary outcome for treatment were 0.404 (95% confidence interval [CI]: 0.202 to 0.809; p = 0.017) and 0.379 (95% CI: 0.173 to 0.832; p = 0.016) before and after adjustment, respectively. The secondary outcome was significantly reduced in the treatment group compared with the control group (6.4% vs. 19.7%; HRs: 0.355 [95% CI: 0.191 to 0.662; p = 0.002] and 0.335 [95% CI: 0.162 to 0.693; p = 0.003] before and after adjustment, respectively). Gynecomastia or breast pain was reported in 16 patients (10.2%) in the treatment group. Serious hyperkalemia led to treatment discontinuation in 3 patients (1.9%).
Conclusions Aldosterone receptor blockade using spironolactone may substantially reduce the risk of both CCV morbidity and death among HD patients; however, larger-scale studies are recommended to further confirm its efficacy. (Effects of Spironolactone on Cardio- and Cerebrovascular Morbidity and Mortality in Hemodialysis Patients; NCT01687699)
Patients with end-stage renal disease (ESRD) undergoing dialysis are at a particularly high risk of cardiovascular and cerebrovascular (CCV) disease, accounting for 40% to 50% of deaths (1,2). Medical therapy involving the use of agents for the renin-angiotensin-aldosterone system (RAAS) is expected to have a significant impact, because the results of large-scale trials have shown that angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs) are beneficial for CCV events in the general population (3–5). However, data supporting the positive effects of these agents in ESRD patients on dialysis are limited (6). For this, it is necessary to evaluate the RAAS components that need to be blocked, so that favorable outcomes can be obtained in patients with different disease severity levels or complications.
Recent studies suggest that the protection mechanism of these blockers could be attributed to a change in either the qualitative or quantitative pathological functions of aldosterone rather than that of angiotensin II (7). In fact, a series beginning with RALES (Randomized Aldactone Evaluation Study) has shown a strong association of aldosterone with a risk of cardiovascular events, including sudden cardiac death (SCD) (8–10). The interactions between cardiovascular disease and cerebrovascular disease have been widely developed over the past 2 decades, and aldosterone blockade has shown protective effects on ischemic cerebral infarct size and cerebrovascular remodeling in stroke-prone rats (11,12). Therefore, clinicians have attempted to improve the outcomes of dialysis patients by managing CCV disease using aldosterone blockade, including spironolactone. However, lack of detailed information on the adverse-effects profile of spironolactone for dialysis patients limits its use.The major concern with administrating spironolactone to patients with renal failure is life-threatening hyperkalemia. Nevertheless, oligoanuric patients requiring dialysis may not be at a risk of spironolactone-induced hyperkalemia when extrarenal potassium disposable (e.g., colonic transport) is marginal. Our pilot study (13) and several previous small studies (14,15) have suggested that spironolactone can be safely administered to patients under hemodialysis (HD) without the development of serious hyperkalemia.
On the basis of these data, we designed DOHAS (Dialysis Outcomes Heart Failure Aldactone Study) to test the hypothesis that daily treatment with low-dose spironolactone would significantly reduce the risk of death from all causes and CCV morbidity in patients with ESRD undergoing HD.
Eligibility criteria were determined by reviewing the medical records and laboratory data of patients. HD patients who were at least 30 years of age, had undergone 4-h-long HD 3 times a week for at least 2 years, and had an average serum potassium level (immediately before dialysis on the first day of the week) of <6.5 mEq/l over the previous 2 months and a 24-h urinary output of <500 ml were included in the study. The exclusion criteria included a history of noncompliance (including noncompliance with HD), unstable vascular access, hypotension, hepatic failure, active malignancy, or any life-threatening disease other than ESRD. Patients were recruited from 5 Japanese outpatient HD clinics affiliated with a teaching hospital. Recruitment began in April 2008 and was completed in December 2008. Finally, 309 patients were registered in the study. Patients were able to receive standard medications such as vitamin D, phosphate binders, erythropoietin, and antihypertensive agents (including ACEIs, ARBs, and beta-blockers).
Study design and procedure
DOHAS was a prospective, multicenter, randomized, controlled, open-label trial. After registration, patients were randomly assigned to either the spironolactone treatment group or the control group. The treatment group was administered an oral spironolactone dose of 25 mg/day as a 3-year plan. Study medication was withheld in the event of life-threatening hyperkalemia (serum potassium level >6.8 mEq/l), gynecomastia, breast pain, or any condition wherein discontinuation of spironolactone treatment was deemed medically necessary by the physician responsible for the patient. Patients who moved to other dialysis facilities outside of the 5 clinics included in the study or who were found to have a malignancy were discontinued from the study. Pre-enrollment medication that may have affected serum potassium levels (e.g., ACEIs, ARBs, loop diuretics, and ion exchange resins) were not changed during the study. HD prescriptions, including dialyzer type, dialysate composition, blood flow rate, and dialysate flow rate, were not changed. Dialysate potassium concentration was 2.0 mEq/l for all patients. No additional dietary potassium restrictions beyond the usual recommendations were imposed. Routine laboratory assessments, including serum potassium levels, were repeated twice a month for all patients. Adverse events other than hyperkalemia were also assessed during the treatment period.
The trial complied with the principles of the Declaration of Helsinki and was approved by the ethics committee at each study clinic. All patients provided written informed consent. An independent data and safety monitoring board periodically reviewed the results in a blinded fashion. Event committees whose members were unaware of treatment assignments assessed the causes of death and reasons for hospitalization. There was no commercial support or involvement in this trial.
The primary outcome of the study was a composite of death from CCV events or hospitalization for CCV events. CCV events included new occurrence or exacerbation of heart failure that was not improved by water removal through dialysis (clinical symptoms together with left ventricular dysfunction by echocardiography according to the American Heart Association [AHA]/American College of Cardiology [ACC] guidelines), malignant ventricular arrhythmias (ventricular fibrillation or sustained ventricular tachycardia), new or recurrent acute myocardial infarction (changes on electrocardiography [ECG] and biomarkers for myocardial infarction), new occurrence or exacerbation of angina pectoris (ECG change corresponding to chest symptoms and coronary angiography showing >75% stenosis according to AHA/ACC guidelines), dissecting aneurysm of the aorta (diagnosed by imaging techniques), stroke (diagnosed by computed tomography [CT] and/or magnetic resonance imaging [MRI]), new or recurrent transient ischemic attack (TIA) (diagnosed by CT and/or MRI, and sudden onset of neurological deficit persisting for <24 h), and SCD. According to the Hemodialysis Study, SCD was defined as an unexpected death, with a preceding symptom duration of <24 h for witnessed deaths and less than the interval since the last dialysis session for unwitnessed deaths (16). The secondary outcome was death from any cause. The causes of death were classified into cardiovascular events (CCV events other than stroke and TIA), cerebrovascular events (stroke and TIA), and others.
Few epidemiological data about CCV risk in Japan were available; hence, the sample size was determined such that the study will be adequately powered for the analysis of the secondary endpoint, time from randomization to death. On the basis of the results of a previous report (17), mortality in the control group was anticipated to be 10% per year. Our original calculation suggested that a sample size of 600 patients was appropriate to detect a hazard-ratio reduction of 40% in the mortality of the treatment group compared with that of the control group at a 2-sided alpha level of 0.05, with 80% power. The study commenced in April 2008; however, because of slow recruitment owing to participation refusal, recruitment closed in December 2008.
Baseline characteristics were compared between the 2 groups. A Fisher’s exact test or chi-square test was used for categorical data, and a Student t test was used for quantitative data. A p value of <0.05 was considered statistically significant.
The analyses of primary and secondary outcomes were conducted by an independent statistician on the basis of the intention-to-treat principle. For multiple CCV events observed in a single patient, only the first CCV event was analyzed as a primary outcome. First, univariate analyses were constructed to determine the difference between the treatment and control group in the primary outcome and its components (cardiovascular events, cerebrovascular events, and sudden death) and in the secondary outcome and its components (death from CCV events, cardiovascular death, and cerebrovascular death). Kaplan-Meier curves were drawn to visualize the primary and secondary outcome distribution over time for the treatment and control groups, and a comparison between the 2 groups was made on the basis of a Cox-Mantel test. Second, a multivariate analysis using a Cox proportional hazards regression model was constructed to determine the difference between the 2 groups in the primary and secondary outcomes with adjustment for sex, duration of dialysis, and cardiothoracic ratio (CTR) to explore the effects of baseline variables. Hazard ratios (HRs), 95% confidence intervals (CIs), and p values were calculated. Lastly, the consistency of the treatment effect on the 2 main outcomes was assessed among 8 pre-defined subgroups, namely, age, sex, duration of dialysis, systolic and diastolic blood pressure, the use of ACEIs or ARBs, CTR, and serum albumin concentration. The effect in each subgroup was analyzed with the use of an unadjusted Cox model, and p values for the interaction were calculated using a Cox proportional hazards regression model with adjustment for the subgroup item. All analyses of outcomes were conducted using DANS version 7.1 (Sugimoto Data Analysis Service, Nagoya, Japan).
Some significant differences in baseline characteristics between the treatment group (n = 157) and the control group (n = 152) were observed (Table 1), with average dialysis duration 27 months longer, serum albumin level 0.08 g/dl lower, and CTR 1.5% greater in the control group than in the treatment group.
There were 45 patients in the treatment group and 24 in the control group who were discontinued from the study. There were 28 patients (12 in the treatment group and 16 in the control group) who discontinued because of transfer to other facilities for reasons that included decreased activities of daily life, progressed dementia, or for nonmedical reasons. There were 13 patients (6 in the treatment group and 7 in the control group) who discontinued because of a new malignancy and 1 patient who discontinued because of renal transplantation. An additional 27 patients discontinued spironolactone therapy because of adverse events (n = 23) or because of withdrawn consent for nonmedical reasons (n = 4). Figure 1 shows the random assignment and the follow-up of the patients.
Primary outcome: death from CCV causes or hospitalization for CCV causes
During the study, a total of 9 patients (5.7%) in the treatment group and 19 patients (12.5%) in the control group reached the primary outcome. Kaplan-Meier curves (Fig. 2A) showed that the patients in the treatment group had fewer CCV events compared with the control group during follow-up. The unadjusted HR for spironolactone treatment was 0.404 (95% CI: 0.202 to 0.809; p = 0.017) and the HR adjusted by sex, duration of dialysis, and CTR was 0.379 (95% CI: 0.173 to 0.832; p = 0.016) (Fig. 3). Unadjusted analyses separating cardiovascular and cerebrovascular causes showed similar reductions in the risk of each outcome among patients in the treatment group compared with those in the control group (HRs of 0.428 and 0.379, respectively), although the findings were not significant (Fig. 3A).
The HRs for 8 pre-defined subgroups (age, sex, duration of dialysis, systolic and diastolic blood pressure, the use of ACEIs or ARBs, CTR, and serum albumin concentration) were calculated. The reduction in the risk of the primary outcome in the treatment group was consistent among subgroups, and none were significant in the interaction tests.
Secondary outcome: death from all causes
There were 10 deaths (6.4%) in the treatment group and 30 (19.7%) in the control group. Kaplan-Meier curves (Fig. 2B) showed that the patients in the treatment group had lower all-cause mortality compared with the control group during follow-up. The unadjusted HR for spironolactone treatment was 0.355 (95% CI: 0.191 to 0.662; p = 0.002) and the HR adjusted by sex, duration of dialysis, and CTR was 0.335 (95% CI: 0.162 to 0.693; p = 0.003) (Fig. 3). Unadjusted analyses for death from CCV events, cardiovascular death, and cerebrovascular death showed similar reductions in the risk of each outcome among patients in the treatment group compared with those in the control group (HRs: 0.430, 0.572, and 0.256, respectively), although the findings were not significant (Fig. 3A).
The HRs for the 8 pre-defined subgroups were calculated in the same manner as that for the primary outcome. The reduction in the risk of the secondary outcome in the treatment group was consistent among subgroups, and none were significant in the interaction tests.
Blood pressure immediately before dialysis was measured at each dialysis session. Spironolactone treatment did not significantly affect blood pressure. The blood pressure values in the 98 patients who survived without CCV events under spironolactone treatment were 152.8 ± 22.7/77.8 ± 14.5 mm Hg at baseline and 152.7 ± 22.0/77.9 ± 10.9 mm Hg at 3 years. In those patients, the average potassium concentration did not increase 3 years after administration of 25 mg/day spironolactone (5.16 mEq/l at baseline vs. 5.14 mEq/l after 3 years). The adverse reactions in the treatment group are shown in Figure 1. During the study, only 3 patients discontinued spironolactone treatment because of hyperkalemia (potassium concentration: 7.3, 6.7, and 6.6 mEq/l; period from treatment initiation: 35, 0.5, and 1 month, respectively). Gynecomastia or breast pain was reported by 10% of patients in the treatment group.
DOHAS is a randomized, prospective clinical trial comparing the efficacy of the addition of spironolactone to the recommended treatment in patients with ESRD on HD. The primary outcome, which was a composite of death or hospitalization because of CCV events, was 5.7% in the treatment group and 12.5% in the control group. There was also a significant >60% reduction in all-cause mortality (secondary outcome) in the treatment group. These reductions, however, cannot be attributed to only the use of spironolactone because there were some differences (including duration of dialysis) in baseline characteristics between the 2 groups. However, a Cox model analysis after adjustment for 3 variables (sex, duration of dialysis, and cardiothoracic ratio) demonstrated significant reductions in primary and secondary outcomes, with corresponding HRs being lower than unadjusted HRs (Fig. 3). Furthermore, the impact of dialysis duration on each outcome appeared to be low, since the HRs for the primary and secondary outcomes were 0.998 and 0.998, respectively (Fig. 3B). Hence, these results suggest that spironolactone can reduce the number of CCV events and deaths in HD patients.
The reason for the high risk of CCV disease in HD patients is not entirely understood (18). The risk factor profile in these HD patients is markedly different from that of the general population; this is because these patients have ill-defined pathophysiologic processes, such as persistent inflammation, endothelial dysfunction, oxidative stress, autonomic dysfunction, and vascular calcification, that are associated with the development of uremic cardiomyopathy or uremic vascular disease (18,19). Multiple therapeutic strategies, such as ACEI treatment (6), homocysteine-lowering therapy (20), intensified nutrition (21), and increased dialysis dose (22), have been thought to reduce this risk, but no survival benefit has been observed with these strategies. Since the seminal data from RALES, the role of aldosterone blockade in the reduction of cardiovascular event risk associated with ESRD has become a topic of interest. Although the concept of DOHAS is challenging because spironolactone is currently contraindicated in dialysis patients, this study is unique because it is the “HD patient version” of RALES and provides evidence for improved outcomes in HD patients. A novel perspective on aldosterone identifies its pathophysiological role, including the stimulation of cytokine production and inflammatory cell recruitment (23), augmentation in the production of transforming growth factor and plasminogen activator inhibitor-1 (24), and key regulation of body ion homeostasis (25). Mineralocorticoid receptors (MRs) exist in not only the epithelium but also in several nonepithelial sites (including the brain, heart, and vasculature), and cardiovascular tissues can also synthesize aldosterone (26). Hence, aldosterone exerts both paracrine and autocrine effects, and consequently induces endothelial dysfunction (27), reduces vascular compliance (28), and causes myocardial and vascular fibrosis (29). This wide range of aldosterone–MR interactions may be implicated in the pathophysiology of CCV disease in HD patients.
An exciting report using spironolactone in HD patients has recently been published. Administration of 50 mg spironolactone 3 times per week for 2 years resulted in a significant decrease or even reversal of carotid intima-medial thickness (30). Carotid intima-medial thickness is a well-documented marker for CCV risk (31,32) and correlates with the presence of myocardial infarction and stroke by cross-sectional analysis (33). Interestingly, the protective effect of spironolactone on the progression of carotid intima-medial thickness was observed in the absence of ACEIs or ARBs (30). In the current study, the use of ACEIs or ARBs did not affect the superiority of spironolactone in the occurrence of CCV events. These data strongly suggest that low-dose spironolactone, even when used alone, reduces the progression of atherosclerosis and/or medial enlargement, possibly resulting in the prevention of CCV events in HD patients.
Aldosterone blockade may play a prominent role in the prevention of cardiac arrest or arrhythmia in HD patients. According to the United States Renal Data System database, 62% of all cardiac deaths among HD patients are attributable to cardiac arrest/cause unknown or arrhythmia (34). One of the mechanisms for this is activation of MRs, which may directly affect the electrical properties of the ventricle. Expression of MRs in the myocardium is enhanced in congestive heart failure (35) and myocardial infarction (36). Furthermore, under conditions of increased reactive oxygen species, such as during heart failure and uremia, the conversion of cortisol to cortisone is inhibited and aldosterone can activate MRs together with cortisol in these tissues (37). At the cellular level, aldosterone has been shown to alter the expression of several different ion channels. These changes may be important contributors to arrhythmogenesis by the modification of cardiac action potential (38). Activation of MRs also affects ionic current levels, such as calcium and potassium (25). Taken together, aldosterone blockade may inhibit ion channel remodeling associated with overstimulation of MRs in cardiomyocytes, resulting in reduced malignant ventricular arrhythmias in HD patients, who easily react electrophysiologically to HD procedures (39).
A previous paper also noted the safety of spironolactone in HD patients, with spironolactone treatment (50 mg, 3 times per week) showing a tendency to increase serum potassium concentration by only 0.012 mEq/l per month during treatment in patients whose serum potassium levels were <6.0 mEq/l before the administration of spironolactone, and none of the 33 patients treated with spironolactone experienced hyperkalemic events during the 24-month follow-up (30). On the other hand, in our pilot study, the median potassium concentration increased by 0.22 mEq/l in 2 weeks and did not show a subsequent concentration increase in 50 HD patients who were administered a spironolactone dose of 25 mg/day (13). In the current study, 98 patients who survived without CCV events for 3 years under spironolactone treatment did not show an increase in average potassium levels, but 3 of 157 patients in the treatment group discontinued spironolactone treatment because of hyperkalemia. It is an important consideration that our inclusion criteria for potassium concentrations may not have been strict enough. Our study excluded patients with an average serum potassium level of ≥6.5 mEq/l in the 2 months before treatment, while other studies excluded those with a serum potassium level of >6.0 mEq/l (14,30). It should be emphasized that there was a difference in the time of blood sampling between these studies. Samples in our study were collected on the first day of the week and those in the other studies were collected on the second day of the week (14,30). Because the average increase in the serum potassium levels immediately before dialysis from the first day to the second day of the week was 0.4 mEq/l in our hospital (Y. Matsumoto, unpublished data, November 2008), our criterion of <6.5 mEq/l may be comparable to the <6.0 mEq/l level used in other studies. Regardless of the consideration of appropriate inclusion criteria, the use of spironolactone in HD patients appear to be safer than in patients with chronic kidney disease not undergoing dialysis.
First, this study was not blinded. Further, a placebo was not administered to the control group. Additionally, the interaction tests were underpowered, and the number of endpoints was small. Therefore, further large-scale prospective, randomized, blinded trials with adequate power are required to confirm these results.
This study showed that low-dose spironolactone can be safely administered for a long duration to HD patients, and it may reduce CCV events and all-cause mortality. We believe that spironolactone may be a useful therapy for improving the currently poor prognosis of HD patients; however, larger-scale studies are necessary to confirm our results.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- angiotensin-converting enzyme inhibitor
- angiotensin receptor blocker
- cardiovascular and cerebrovascular
- confidence interval
- computed tomography
- cardiothoracic ratio
- end-stage renal disease
- hazard ratio
- magnetic resonance imaging
- renin-angiotensin-aldosterone system
- sudden cardiac death
- transient ischemic attack
- Received June 17, 2013.
- Revision received August 8, 2013.
- Accepted September 9, 2013.
- American College of Cardiology Foundation
- Schrader J.,
- Luders S.,
- Kulschewski A.,
- et al.
- Struthers A.D.,
- MacDonald T.M.
- Dorrance A.M.,
- Osborn H.L.,
- Grekin R.,
- Webb R.C.
- Saudan P.,
- Mach F.,
- Perneger T.,
- et al.
- Stenvinkel P.,
- Carrero J.J.,
- Axelsson J.,
- Lindholm B.,
- Heimburger O.,
- Massy Z.
- Cano N.J.,
- Fouque D.,
- Roth H.,
- et al.
- Perrier E.,
- Kerfant B.G.,
- Lalevee N.,
- et al.
- Schafer A.,
- Fraccarollo D.,
- Hildemann S.K.,
- Tas P.,
- Ertl G.,
- Bauersachs J.
- Sun Y.,
- Ramires F.J.,
- Weber K.T.
- Vukusich A.,
- Kunstmann S.,
- Varela C.,
- et al.
- Tsivgoulis G.,
- Vemmos K.,
- Papamichael C.,
- et al.
- O'Leary D.H.,
- Polak J.F.,
- Kronmal R.A.,
- Manolio T.A.,
- Burke G.L.,
- Wolfson S.K. Jr..,
- for the Cardiovascular Health Study Collaborative Research Group
- Lorenz M.W.,
- Markus H.S.,
- Bots M.L.,
- Rosvall M.,
- Sitzer M.