Advance Publication by J-STAGE Circulation Journal LATE BREAKING CLINICAL TRIAL (JCS 2012) Official Journal of the Japanese Circulation Society http://www.j-circ.or.jp Superiority of Long-Acting to Short-Acting Loop Diuretics in the Treatment of Congestive Heart Failure – The J-MELODIC Study – Tohru Masuyama, MD; Takeshi Tsujino, MD; Hideki Origasa, PhD; Kazuhiro Yamamoto, MD; Takashi Akasaka, MD; Yutaka Hirano, MD; Nobuyuki Ohte, MD; Takashi Daimon, PhD; Satoshi Nakatani, MD; Hiroshi Ito, MD

Background: Diuretics are the most prescribed drug in heart failure (HF) patients. However, clinical evidence about their long-term effects is lacking. The purpose of this study was to compare the therapeutic effects of furosemide and azosemide, a short- and long-acting loop diuretic, respectively, in patients with chronic heart failure (CHF).

Methods and Results: In this multicenter, prospective, randomized, open, blinded endpoint trial, we compared the effects of azosemide and furosemide in patients with CHF and New York Heart Association class II or III symptoms. 320 patients (160 patients in each group, mean age 71 years) were followed up for a minimum of 2 years. The pri- mary endpoint was a composite of cardiovascular death or unplanned admission to hospital for congestive HF. During a median follow-up of 35.2 months, the primary endpoint occurred in 23 patients in the azosemide group and in 34 patients in the furosemide group (hazard ratio [HR], 0.55, 95% confidence interval [CI] 0.32–0.95: P=0.03). Among the secondary endpoints, unplanned admission to hospital for congestive HF or a need for modification of the treatment for HF were also reduced in the azosemide group compared with the furosemide group (HR, 0.60, 95%CI 0.36–0.99: P=0.048).

Conclusions: Azosemide, compared with furosemide, reduced the risk of cardiovascular death or unplanned admis- sion to hospital for congestive HF.

Key Words: Azosemide; Chronic heart failure; Furosemide; Loop diuretics

iuretics are the most frequently prescribed medicine in tion. However, there has been no randomized clinical trial at patients with chronic heart failure (CHF), not only in all that provides data showing the presence or absence of the D Japan but also in Western countries. Diuretics, par- benefits of long-term administration of loop diuretics. Most ticularly loop diuretics, are even more frequently prescribed recently, subanalyses of megatrials such as the Studies Of Left than renin-angiotensin system (RAS) inhibitors or β-blockers Ventricular Dysfunction and The Digitalis Investigation Group in any cohort study, although any guideline of CHF manage- trial study suggest that prognosis is poorer in CHF patients ment clearly recommends the use of RAS inhibitors and β- given larger doses of loop diuretics.7,8 Moreover, the Japanese blockers prior to diuretics.1–6 This is because any cardiologist Cardiac Registry of Heart Failure in Cardiology (JCARE- knows that persistent congestion is associated with a poor CARD) study revealed that diuretic use at discharge was an prognosis and loop diuretics are effective in lessening conges- independent predictor for all-cause death during long-term

Received January 5, 2012; revised manuscript received February 26, 2012; accepted February 27, 2012; released online March 16, 2012 Time for primary review: 7 days Cardiovascular Division, Department of Internal Medicine, Hyogo College of Medicine, Nishinomiya (T.M.); Department of Pharmacy, School of Pharmacy, Hyogo University of Health Sciences, Kobe (T.T.); Division of Biostatistics and Clinical Epidemiology, Univer- sity of Toyama Graduate School of Medicine, Toyama (H.O.); Division of Molecular Medicine and Therapeutics, Department of Mul- tidiscipinary Internal Medicine, Faculty of Medicine, Tottori University, Yonago (K.Y.); Department of Cardiovascular Medicine, Wakayama Medical University, Wakayama (T.A.); Division of Central Clinical Laboratory, Kinki University School of Medicine, Osakasayama (Y.H.); Department of Cardio-Renal Medicine and Hypertension, Nagoya City University Graduate School of Medical Sciences, Nagoya (N.O.); Division of Biostatistics, Department of Mathematics, Hyogo College of Medicine, Nishinomiya (T.D.); De- partment of Health Sciences, Osaka University Graduate School of Medicine, Suita (S.N.); and Department of Cardiovascular Medicine, Akita University Graduate School of Medicine, Akita (H.I.), Japan This paper was presented at the 76th Annual Scientific Meeting of the Japanese Circulation Society, Late Breaking Clinical Trials 2-4. ClinicalTrials.gov number, NCT00355667; UMIN Clinical trial registry number, UMIN000000528. Mailing address: Tohru Masuyama, MD, PhD, FACC, Cardiovascular Division, Department of Internal Medicine, Hyogo College of Medicine, 1-1 Mukogawa-cho, Nishinomiya 663-8501, Japan. E-mail: [email protected] ISSN-1346-9843 doi: 10.1253/circj.CJ-11-1500 All rights are reserved to the Japanese Circulation Society. For permissions, please e-mail: [email protected] Advance Publication by J-STAGE MASUYAMA T et al.

Figure 1. Trial profile.

follow-up among elderly patients hospitalized with HF.9 Some uncontrolled hypertension; current symptomatic hypotension; experimental studies have demonstrated the adverse effects of serum creatinine >2.5 mg/dl; acute coronary syndrome; im- furosemide in HF models.10,11 This may be explained by the plantable cardiac defibrillator; hemodynamically significant fact that loop diuretics activate the RAS and sympathetic left ventricular outflow tract obstruction; acute myocardial in- nerve system, resulting in the poor prognosis in patients given farction within the past 3 months; percutaneous coronary inter- loop diuretics.10–12 Diuretics are certainly essential in most vention or open heart surgery within the past 3 months; any patients with CHF to minimize symptoms related to conges- change in cardiovascular drug therapy within the month prior tion and therefore should not be withdrawn. However, the use to randomization; requirement for intravenous inotropes; and of a long-acting rather than a short-acting loop diuretic may any serious non-cardiovascular disease including malignancy. well provide a better prognosis in patients with CHF. Yoshida et al reported that the administration of azosemide, a long-act- Study Procedure ing loop diuretic, improved the mortality rate in a hypertensive The study used a multicenter, prospective, randomized, open, HF model.13 However, there has been no randomized clinical blinded endpoint (PROBE) design. After screening for eligibil- trial of loop diuretics in patients with CHF. Thus, we designed ity and obtaining written informed consent, patients were ran- a multicenter prospective study, J-MELODIC (Japanese Mul- domized to either azosemide or furosemide treatment in a 1:1 ticenter Evaluation of LOng- versus short-acting Diuretics In ratio. In any arm, patients were treated with standard therapy Congestive heart failure) to obtain clinical evidence of the including digitalis, mineralocorticoid receptor blockers, angio- effects of diuretics in HF patients.14 This is the first prospec- tensin-converting enzyme inhibitors, angiotensin II type 1 re- tive randomized study of the effect of long-term administra- ceptor blockers, β-blockers, and calcium-channel blockers. tion of a long-acting vs. a short-acting diuretics on the prog- Patients discontinued taking any previous loop diuretic(s) nosis in patients with CHF. and were directly rolled over to one of the 2 arms with either azosemide 30–60 mg/day or furosemide 20–40 mg/day, with- out a placebo run-in period. We considered 40 mg furosemide Methods to be equivalent to approximately 60 mg azosemide, based on Patients the result of a previous clinical pharmacology study.17 The The design of the J-MELODIC study has been published in dose of each diuretic was appropriately adjusted within 8 weeks detail.14 The trial was approved by each center’s ethics com- according to the symptoms of each patient, and patients were mittee and all patients provided written informed consent. maintained for the rest of the study. When it was difficult Eligibility criteria were as follows: 20 years or older; clinical to stabilize symptoms of HF, azosemide was increased up to diagnosis of HF based on a slight modification of the Framing- 120 mg/day, and furosemide was increased up to 80 mg/day. ham criteria15 as previously described16 within 6 months before Diuretics other than furosemide, azosemide, thiazides and min- study entry; New York Heart Association (NYHA) functional eralocorticoid inhibitors were not allowed to be taken. class II or III symptoms; loop diuretic(s) user; and no change in Thereafter, patients were reviewed every 2–8 weeks. The baseline therapy or symptoms of HF within 1 month. planned minimum follow-up period for each patient was 2 Key exclusion criteria were: uncontrolled diabetes mellitus; years, and electrocardiography, echocardiography, chest X-ray Advance Publication by J-STAGE Long-Acting Loop Diuretic in Congestive HF

Table 1. Baseline Characteristics of the Patients According to Study Group Furosemide group Azosemide group P value (n=160) (n=160) Age (years) 71±11 71±11 0.45 Female 65 (41) 59 (37) 0.49 Smoking 25 (16) 22 (14) 0.75 NYHA class III 19 (12) 16 (10) 0.76 SAS questionnaire score (Mets) 4.7±1.5 4.6±1.7 0.30 Weight (kg) 59.0±12.8 59.8±13.3 0.56 CTR (%) 55.7±7.3 55.3±7.4 0.61 BMI (kg/m2) 23.2±4.2 23.6±4.5 0.40 SBP (mmHg) 125±17 126±18 0.58 DBP (mmHg) 71±11 71±11 0.74 Heart rate (beats/min) 71±14 72±13 0.54 LVEF (%) 52±15 51±16 0.50 LVDd (mm) 53±9 54±9 0.95 BUN (mg/dl) 22.2±9.1 21.6±9.3 0.55 eGFR (ml · min–1 · 1.73 m–2) 55±19 53±17 0.18 Hb (g/dl) 12.8±2.0 12.7±1.9 0.72 Potassium (mEq/L) 4.3±0.5 4.3±0.4 0.11 BNP (pg/ml) 115 (52, 285) 121 (44, 260) 0.92 NE (pg/ml) 694 (470, 1,000) 631 (440, 950) 0.28 AF/AFL 70 (44) 59 (37) 0.25 Etiology of HF 0.2 Ischemic 47 (30) 60 (37) DCM 34 (21) 36 (23) Other 79 (49) 64 (40) History Hypertension 102 (64) 103 (64) 0.91 Diabetes mellitus 56 (35) 45 (28) 0.23 Dyslipidemia 81 (51) 81 (51) 1.00 Stroke 23 (14) 21 (13) 0.87 Medication ACEI or ARB 108 (68) 118 (74) 0.27 β-blocker 78 (49) 88 (55) 0.31 Aldosterone-antagonist 65 (41) 63 (39) 0.90 Digitalis 39 (24) 30 (19) 0.28 Statin 56 (35) 65 (41) 0.35 Warfarin 71 (44) 63 (39) 0.42 Data are mean±standard deviation, median (25 percentile, 75 percentile) or patients number (%). NYHA, New York Heart Association; CTR, cardiothoracic ratio; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; LVEF, left ventricular ejection fraction; LVDd, left ventricular end-diastolic dimension; BUN, blood urea nitrogen; eGFR, estimated glomerular filtration rate; Hb, hemoglobin; BNP, brain natriuretic peptide; NE, norepinephrine; AF, atrial fibrillation; AFL, atrial flutter; HF, heart failure; DCM, dilated cardiomyopathy; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin type 1 receptor blocker. and blood sampling were conducted at study entry and every with BNP ≥200 pg/ml at randomization; unplanned admission 12 months after the randomization. to hospital for congestive HF, or a need for modification of the treatment for HF (changes in oral medicine for at least 1 month Study Endpoints or the addition of intravenous drug(s) for at least 4 h). Hospi- The primary endpoint was a composite of cardiovascular death talization was defined as care at an acute hospital lasting for at and unplanned admission to hospital for congestive HF. Car- least 24 h. Hospitalization for HF required documentation that diovascular death was defined as death from worsening of worsening HF was the principal reason for admission. Patients congestive HF, coronary artery disease, cardiac arrest, cardiac who discontinued treatment continued to be followed for arrhythmia, myocardial infarction, stroke, or sudden death. The events. Endpoints were assessed by the independent endpoint secondary endpoints were: all-cause mortality; worsening of committee for whom the allocated group was blinded to all the the symptoms (defined as either a decrease by ≥1 Mets in the committee members. Specific Activity Scale questionnaire score or an increase by ≥I class in the NYHA functional class for at least 3 months as Statistical Analysis compared with baseline); an increase in brain natriuretic pep- Sample Size Determination The short-acting loop diuret- tide (BNP) by ≥30% of the value at randomization in patients ics, furosemide, has produced data of a 2-year incidence of the Advance Publication by J-STAGE MASUYAMA T et al.

Table 2. Clinical and Laboratory Parameters 2 Years After Randomization Furosemide group Azosemide group P value (n=133) (n=131) Weight (kg) 60.1 ±12.5 59.7±11.0 0.75 CTR (%) 55.5±7.0 54.7±6.5 0.34 SBP (mmHg) 123±17 124±19 0.67 DBP (mmHg) 70±11 73±12 0.05 Heart rate (beats/min) 71±13 71±13 0.89 LVEF (%) 53±14 53±14 0.89 LVDd (mm) 53±8 54±9 0.60 BUN (mg/dl) 21.9±8.6 20.9±9.65 0.38 eGFR (ml · min–1 · 1.73 m–2) 60.8±22.6 58.2±22.5 0.36 Hemoglobin (g/dl) 13.0±2.1 12.9±2.0 0.55 Potassium (mEq/L) 4.3±0.4 4.3±0.5 0.24 BNP (pg/ml) 101 (44, 262) 83 (38, 168) 0.12 NE (pg/ml) 668 (465, 870) 680 (474, 945) 0.65 Data are mean±standard deviation or median (25 percentile, 75 percentile). Abbreviations see in Table 1. primary endpoint of 35%, whereas the 2-year incidence with was 71±11 years (±SD); 39% were female and 11% were in the use of the long-acting loop diuretics, torasemide, was NYHA class III. Median left ventricular ejection fraction (EF) demonstrated to be 15%.18 It suggests that the hazard ratio was 51%. The etiology of HF was ischemic heart disease (HR) of azosemide might be 0.38 against furosemide. How- (33%), dilated cardiomyopathy (22%), and others (45%). The ever, the 2-year incidence of death and hospitalization for HF 2 treatment groups had similar characteristics at study entry in Japanese patients who were treated with standard therapy (Table 1). was reported to be 15.3% for 52 weeks.19 Therefore, it was assumed to be 25% in furosemide and the HR was conserva- Study Drug Administration and Follow-up tively set at 0.5. Assuming a 2-year recruitment and a maxi- 3 patients in the furosemide group and 2 patients in the mum 4-year follow-up, a total of 132 per group was required azosemide group did not start the study medication. At the to detect a statistical significance between groups with the use study’s cutoff date, the study drug had been discontinued in 20 of a log-rank test, with a power of 80% under the method by patients (12.5%) receiving furosemide and 25 patients (15.6%) Lakatos and Lan.20 Considering a small portion of attritions, receiving azosemide. At the cutoff date, 5 patients (3.1%) in we ended up with a total of 160 per group. the furosemide group and 3 patients (1.9%) in the azosemide Analyses Values are expressed as mean ± standard devia- group were lost to follow-up (Figure 1). Clinical and labora- tion (SD). Differences between groups were assessed with the tory parameters 2 years after randomization were similar in the χ2 test for categorical data. Differences in continuous variables 2 groups (Table 2). When the data were censored in the anal- between groups were assessed with unpaired Student’s t-tests ysis at the time of the last contact (Analysis A), the median when normally distributed, and with the Wilcoxon rank-sum duration of follow-up among all patients was 35.2 months, test when not normally distributed. Cumulative event-free with 913 patient-years of follow-up. When the data were cen- curves were constructed by means of Kaplan-Meier methods, sored for patients were lost to follow-up or discontinuing the and differences between treatment groups were evaluated with intervention (Analysis B), the median duration of follow-up the use of log-rank tests. The Cox proportional hazards regres- was 35.1 months, with 883 patient-years of follow-up. sion model was used to examine the effect of treatment in the presence of prespecified covariates. All analyses were con- Study Endpoints ducted in accordance with the intention-to-treat principle. If When we censored the data for patients lost to follow-up only, contact could not be made at the trial cutoff date, the data were the composite of cardiovascular death and unplanned admis- censored in the analysis at the time of the last contact (Analy- sion to hospital for congestive HF (primary endpoint) occurred sis A). We also performed another analysis by censoring the in 23 patients in the azosemide group and 34 patients in the data when patients were lost to follow-up or discontinued the furosemide group (Figure 2A, Table 3). The HR for the pri- intervention (Analysis B). Proportions such as worsening BNP mary endpoint in the azosemide group, as compared to the and symptoms were compared with either the chi-square test furosemide group, was 0.55 (95% confidence interval [CI] or Fisher’s exact test. Statistical significance was defined by 0.32–0.95: P=0.03). It was consistent across all subgroup anal- 2-sided 5% and the analyses were performed by SAS statisti- ysis (Figure 3). Among the secondary endpoints, unplanned cal software version 9.1. admission to hospital for congestive HF or the need for modi- fication of the treatment for HF occurred in 26 patients in the azosemide group and 36 patients in the furosemide group (HR Results 0.60, 95%CI 0.36–0.99: P=0.048) (Figure 2B, Table 3). All- Patients cause mortality was recorded for 17 patients in the azosemide From June 2, 2006 to August 31, 2008, a total of 320 patients group and for 17 patients in the furosemide group (no differ- underwent randomization at 15 sites in Japan. Of these pa- ence between the 2 groups) (Figure 2C, Table 3). tients, 160 were assigned to receive furosemide and 160 to When we censored the data for patients lost to follow-up or receive azosemide (Figure 1). The mean age of all patients discontinuing the intervention, the value for the primary out- Advance Publication by J-STAGE Long-Acting Loop Diuretic in Congestive HF

Figure 2. Cumulative Kaplan-Meier estimates of the rates of the primary endpoint and other endpoints, according to study group. The hazard ratios for azosemide vs. furosemide are shown for cardiovascular death or un- planned hospitalization for congestive heart failure (the primary endpoint) (A), unplanned hospitalization or modification of treatment for congestive heart failure (B), and death from any cause (C). Advance Publication by J-STAGE MASUYAMA T et al.

Table 3. Primary Outcome, Secondary Outcomes, and Component Events (Analysis A) Azosemide Furosemide Adjusted HR Adjusted Outcome group group (95%CI) P value Primary outcome Cardiovascular death and unplanned admission to 23 34 0.55 (0.32–0.95) 0.03 hospital for congestive HF Secondary outcomes Unplanned admission to hospital for congestive HF 26 36 0.60 (0.36–0.997) 0.048 or a need for modification of the treatment for HF All-cause mortality 17 17 0.93 (0.47–1.85) 0.99 Component events Cardiovascular death 7 11 0.64 (0.24–1.67) 0.36 Unplanned admission to hospital for congestive HF 19 29 0.53 (0.30–0.96) 0.04 Need for modification of the treatment for HF 7 10 0.68 (0.26–1.80) 0.44 without admission HR, hazard ratio; CI, confidence interval; HF, heart failure.

Table 4. Primary Outcome, Secondary Outcomes, and Component Events (Analysis B) Azosemide Furosemide Adjusted HR Adjusted Outcome group group (95%CI) P value Primary outcome Cardiovascular death and unplanned admission to 18 32 0.47 (0.26–0.83) 0.01 hospital for congestive HF Secondary outcomes Unplanned admission to hospital for congestive HF 22 35 0.52 (0.30–0.90) 0.02 or a need for modification of the treatment for HF All-cause mortality 11 15 0.67 (0.30–1.47) 0.31 Component events Cardiovascular death 5 10 0.49 (0.17–1.45) 0.20 Unplanned admission to hospital for congestive HF 15 28 0.45 (0.24–0.84) 0.01 Need for modification of the treatment for HF without 7 10 0.68 (0.26–1.79) 0.43 admission Abbreviations see in Table 4.

Table 5. Adverse Events Furosemide group Azosemide group New onset of hypokalemia 1 year after randomization 2 6 Hypotension 0 1 Digitalis intoxication 1 0 Eruption 1 0 Arthritis 1 0

come was lower in the azosemide group than in the frosemide group and 15% in the furosemide group at 2 years after ran- group (18 vs. 32, P=0.01) (Table 4). Among the secondary domization (P=0.75). endpoints, unplanned admission to hospital for congestive HF or the need for modification of the treatment for HF occurred Safety less frequently in the azosemide group than in the furosemide Relevant adverse events are summarized in Table 5. The in- group (22 vs. 35, P=0.02) (Table 4). cidence of hypotension, hypokalemia, digitalis intoxication or Worsening of symptoms occurred in 29 patients in the eruption did not differ between the 2 groups. azosemide group and in 34 patients in the furosemide group at 1 year after randomization (P=0.67) and in 34 patients in the azosemide group and 33 patients in the furosemide group at 2 Discussion years after randomization (P=0.89). Plasma BNP concentra- We compared the effect of long-term administration of tion was ≥200 pg/ml in 49 patients in the azosemide group and azosemide, a long-acting loop diuretic, to furosemide, a short- in 51 patients in the furosemide group. An increase in BNP by acting diuretic, on the prognosis in patients with CHF. The >30% of the value at randomization occurred in 11% in the rate of the primary endpoint, a composite of cardiovascular azosemide group and 15% in the furosemide group at 1 year death and unplanned admission to hospital for congestive HF, after randomization (P=0.75) and in 11% in the azosemide was significantly smaller in the azosemide group than in the Advance Publication by J-STAGE Long-Acting Loop Diuretic in Congestive HF

Figure 3. Subgroup analysis. Hazard ratios for cardiovascular death or unplanned admission to hospital for congestive heart failure (the primary endpoint) with furosemide vs. azosemide, according to subgroups. The subgroups are based on baseline demographic and clinical characteristics. The size of each square corresponds to the number of patients with an event. Data are missing for some patients in some subgroups. eGFR, estimated glomerular filtration rate; LVEF, left ventricular ejection fraction; LVDd, left ventricular end-diastolic dimension; AF, atrial fibrillation; AFL, atrial flutter; ACEI, angiotensin-converting enzyme in- hibitor; ARB, angiotensin type 1 receptor blocker. Advance Publication by J-STAGE MASUYAMA T et al. furosemide group. The superiority of azosemide to furosemide hibitors and β-blockers if tolerated” in the guideline from the was consistent across all subgroups. With azosemide, there European Society of Cardiology.4 One might think combining was also a reduction in unplanned admissions to hospital for the administration of ACE inhibitors/angiotensin-receptor congestive HF or the need for modification of the treatment blockers (ARBs) and/or β-blockers might prevent loop di- for HF. Adverse effects occurred equally in both groups. uretic-induced neurohumoral activation. Long-acting loop di- Azosemide and furosemide were equally safe in the treatment uretics were superior to short-acting loop diuretics even when of CHF. ACE inhibitors or ARBs and β-blockers were administered in The mechanisms by which azosemide provided a better 226 (71%) and 166 (52%) of patients, respectively. Currently, prognosis than furosemide are not fully understood. Loop di- we do not have data about the doses of the ACE inhibitors/ uretics may be divided into short- and long-acting, and furo- ARBs and β-blockers, and even if administered to most of the semide, a short-acting drug, is the most widely used. McCur- patients, the doses might have been too small to prevent loop ley et al demonstrated that the acceleration of left ventricular diuretic-induced neurohumoral activation. Alternatively, the dysfunction by furosemide was associated with an increase in difference may not be explained solely by loop diuretic-in- plasma aldosterone level in a tachycardia-induced HF model.11 duced neurohumoral activation and another mechanism should Yoshida et al have shown that sympathetic activation might be considered. explain the lack of improvement in mortality for HF treatment We included in the study not only patients with HF and re- with furosemide, despite the reduction in preload and after- duced EF but also those with HF with preserved EF (HFPEF). load.13 These experimental studies suggested that neurohu- Regarding the etiologies of HF, “others” was the most frequent moral activation by short-acting loop diuretics partly explains category. Enrollment of HFPEF patients might cause the lower the lack of improvement in mortality for patients treated with prescription rates of ACE inhibitors/ARBs and β-blockers in loop diuretics. Azosemide, a long-acting loop diuretic, in- J-MELODIC than in contemporary trials of CHF because these duced diuresis without activating sympathetic nerves in a rat medicines have not been proven to be effective in HFPEF. model of HF.13 However, plasma norepinephrine levels 2 When we compared the superiority of long-acting diuretics to years after randomization were not different between the 2 short-acting diuretics between those with and without a re- groups. We did not collect blood samples after 30 min of bed duced EF, there was no difference between the subsets. rest nor did we request the patients to take diuretics in the morning of blood sampling. These sampling conditions might Study Limitations obscure differences between the 2 groups. Electrolyte or met- 3 limitations of the study are noted. First, unplanned admission abolic disturbances induced by loop diuretics could have del- to hospital for congestive HF, part of the primary endpoint, eterious effects on the prognosis on CHF. The incidence of depended on the subjective judgment of the admitting doctor. hypokalemia was similar between the 2 groups. However, we One might doubt that the superiority of azosemide to furose- did not evaluate serum levels of sodium, chloride, calcium, mide was caused by a bias of researchers toward azosemide magnesium, glucose or lipids in this study. Azosemide might because the current study used a PROBE design. However, have different effects on these factors from furosemide. Fur- unplanned admission to hospital for congestive HF or the need ther studies are needed. for modification of the treatment for HF, one of the secondary Western readers might think the doses of loop diuretics endpoints, was also reduced in the azosemide group, compared were too low in this study. Japanese doctors usually hesitate to the furosemide group. This finding suggests that intentional to prescribe higher doses of loop diuretics because they fre- avoidance of admission was not the cause of the reduced num- quently induce worsening of renal function and hyponatremia. ber of primary endpoints in the azosemide group. Second, the In a recent post-hoc analysis of the Beta-Blocker Evaluation number of patients was small and the incidence rate of the of Survival Trial study, blood urea nitrogen was less than primary endpoint was low at 6.2%/year, which was much less 21 mg/dl in approximately one-third of the CHF patients who than expected. The reasons for the low incidence rate of the received high-dose loop diuretics (furosemide ≥160 mg/day).21 primary endpoint were that patients with NYHA III symp- If we prescribe such a high dose of loop diuretics in Japan, toms at randomization comprised only 11% in this study and almost all the patients would show azotemia. the rates of patients who received ACE inhibitors, ARB, and The incidence of worsening of symptoms and of an increase β-blockers were higher than in previous Japanese trials of in plasma BNP by ≥30% of the value at randomization in pa- CHF.19,28,29 The low incidence rate of the primary endpoint tients with BNP ≥200 pg/ml at randomization did not differ weakened the statistical power of this study, together with the between the azosemide and furosemide groups. This might be small number of patients. All subset analyses were negative; related to the reinforcement of the therapy in the subjects however, we still need further studies to conclude that whose symptoms worsened. azosemide is better than furosemide for all types of CHF. Previous randomized or non-randomized observational Third, most patients were assumed to be taking furosemide studies have suggested the superiority of torasemide, a long- before randomization because it is by far the most prescribed acting loop diuretic, to furosemide.17,22,23 Torasemide inhibits loop diuretic in Japan. Thus, most patients in the azosemide aldosterone binding to its receptor24,25 and several clinical stud- group needed dose adjustment of the newly prescribed medi- ies have demonstrated the usefulness of aldosterone blockers cine after randomization, whereas patients in the furosemide in patients with HF.26,27 Thus, it is unclear whether the benefits group did not. The difference in the ratio of patients who of torasemide are provided through its longer half-life or changed medicine at randomization could have affected the through an associated aldosterone blockade action. Our results incidence of the endpoints. Fortunately, this might not be the indicate that the benefits of torasemide are provided at least case in this study because the numbers of endpoints were partially through its longer half-life. Azosemide, which has a similar in both groups within 8 weeks after randomization. longer half-life and longer duration of action than torasemide, may be the first choice of loop diuretic in CHF treatment. It is stated that “diuretics should always be administered in Conclusion combination with angiotensin-converting enzyme (ACE)-in- Azosemide, compared to furosemide, reduced the risk of car- Advance Publication by J-STAGE Long-Acting Loop Diuretic in Congestive HF diovascular death and unplanned admission to hospital for con- 13. Yoshida J, Yamamoto K, Mano T, Sakata Y, Nishio M, Ohtani T, et gestive HF. In spite of recent progress in the treatment of HF al. Different effects of long- and short-acting loop diuretics on sur- following a number of prospective clinical trials during the past vival rate in Dahl high-salt heart failure model rats. Cardiovasc Res 30 2005; 68: 118 – 127. decades, the prognosis of HF patients is still poor. Switching 14. J-MELODIC Program Committee. Rationale and design of a random- short-acting diuretics to long-acting ones may improve the ized trial to assess the effects of diuretics in heart failure: Japanese prognosis of patients with CHF. Multicenter Evaluation of Long- vs Short-Acting Diuretics in Con- gestive Heart Failure (J-MELODIC). Circ J 2007; 71: 1137 – 1140. 15. McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural Acknowledgments history of congestive heart failure: The Framingham study. N Engl J This work was supported by a grant from the Ministry of Health, Labor Med 1971; 285: 1441 – 1446. and Welfare, Japan [H18-Junkanki-tou (Seishuu)-Ippan-046)], and a grant 16. Senni M, Tribouilloy CM, Rodeheffer RJ, Jacobsen SJ, Evans JM, from the Japan Heart Foundation. Bailey KR, et al. Congestive heart failure in the community: A study The authors gratefully acknowledge the secretarial assistance of Ms of all incident cases in Olmsted County, Minnesota, in 1991. Circu- Satomi Fujita and Hiroko Nakashima (Hyogo College of Medicine, Nishi- lation 1998; 98: 2282 – 2289. nomiya, Japan) and the clinical assistance of clinical research coordinators 17. Krück F, Bablok W, Besenfelder E, Betzien G, Kaufmann B. Clinical Ms Hiroko Nakamura and Ms Yumiko Hashimoto. and pharmacological investigations of the new saluretic azosemid. Eur J Clin Pharmacol 1978; 14: 153 – 161. 18. Cosin J, Diez J. Torasemide in chronic heart failure: Results of the Disclosures TORIC study. Eur J Heart Fail 2002; 4: 507 – 513. Conflict of Interest: None declared. 19. Effects of Pimobendan on Chronic Heart Failure Study (EPOCH Study). Effects of pimobendan on adverse cardiac events and physi- cal activities in patients with mild to moderate chronic heart failure: References The effects of pimobendan on chronic heart failure study (EPOCH 1. Lenzen MJ, Scholte op Reimer WJ, Boersma E, Vantrimpont PJ, study). Circ J 2002; 66: 149 – 157. Follath F, Swedberg K, et al. Differences between patients with a 20. Lakatos E, Lan KKG. A comparison of sample size methods for the preserved and a depressed left ventricular function: A report from the logrank statistic. Stat Med 1992; 11: 179 – 191. EuroHeart Failure Survey. Eur Heart J 2004; 25: 1214 – 1220. 21. Testani JM, Cappola TP, Brensinger CM, Shannon RP, Kimmel SE. 2. Tsutsui H, Tsuchihashi-Makaya M, Kinugawa S, Goto D, Takeshita Interaction between loop diuretic-associated mortality and blood A; JCARE-GENERAL Investigators. Characteristics and outcomes urea nitrogen concentration in chronic heart failure. J Am Coll Car- of patients with heart failure in general practices and hospitals. Circ J diol 2011; 58: 375 – 382. 2007; 71: 449 – 454. 22. Stroupe KT, Forthofer MM, Brater DC, Murray MD. Healthcare 3. Hunt SA, Abraham WT, Chin MH, Feldman AM, Francis GS, costs of patients with heart failure treated with torasemide or furose- Ganiats TG, et al; American College of Cardiology; American Heart mide. Pharmacoeconomics 2000; 17: 429 – 440. Association Task Force on Practice Guidelines; American College of 23. Murray MD, Deer MM, Ferguson JA, Dexter PR, Bennett SJ, Perkins Chest Physicians; International Society for Heart and Lung Trans- SM, et al. Open-label randomized trial of torasemide compared with plantation; Heart Rhythm Society. ACC/AHA 2005 Guideline Up- furosemide therapy for patients with heart failure. Am J Med 2001; date for the Diagnosis and Management of Chronic Heart Failure in 111: 513 – 520. the Adult: A report of the American College of Cardiology/American 24. Uchida T, Yamanaga K, Nishikawa M, Ohtaki Y, Kido H, Watanabe Heart Association Task Force on Practice Guidelines (Writing Com- M. Anti-aldosteronergic effect of torasemide. Eur J Pharmacol 1991; mittee to Update the 2001 Guidelines for the Evaluation and Manage- 205: 145 – 150. ment of Heart Failure): Developed in collaboration with the Ameri- 25. Uchida T, Yamanaga K, Kido H, Ohtaki Y, Watanabe M. Diuretic and can College of Chest Physicians and the International Society for vasodilating actions of torasemide. Cardiology 1994; 84: 14 – 17. Heart and Lung Transplantation: Endorsed by the Heart Rhythm 26. Pitt B, Zannad F, Remme WJ, Cody R, Castaigne A, Perez A, et al. Society. Circulation 2005; 112: e154 – e235. The effect of spironolactone on mortality and morbidity in patients 4. European Society of Cardiology. Task force for the Diagnosis and with severe heart failure. N Engl J Med 1999; 341: 709 – 717. Treatment of CHF: Guidelines for the diagnosis and treatment of 27. Zannad F, McMurray JJ, Krum H, van Veldhuisen DJ, Swedberg K, chronic heart failure. 2005. http://www.escardio.org/knowledge/ Shi H, et al; EMPHASIS-HF Study Group. Eplerenone in patients guidelines/Chronic_Heart_Failure.htm (accessed February 22, 2012). with systolic heart failure and mild symptoms. N Engl J Med 2011; 5. Matsuzaki M, Ishii M, Izumi T, Itoh H, Itoh H, Ueda Y, et al. Guide- 364: 11 – 21. lines for treatment of chronic heart failure (JCS 2010). http://www. 28. Hori M, Sasayama S, Kitabatake A, Toyo-oka T, Handa S, Yokoyama j-circ.or.jp/guideline/pdf/JCS2010_matsuzaki_h.pdf (in Japanese) M, et al; MUCHA Investigators. Low-dose carvedilol improves left (accessed February 22, 2012). ventricular function and reduces cardiovascular hospitalization in 6. Nair AP, Timoh T, Fuster V. Contemporary medical management of Japanese patients with chronic heart failure: The Multicenter Carve- systolic heart failure. Circ J 2012; 76: 268 – 277. dilol Heart Failure Dose Assessment (MUCHA) trial. Am Heart J 7. Domanski M, Norman J, Pitt B, Haigney M, Hanlon S, Peyster E. 2004; 147: 324 – 330. Diuretic use, progressive heart failure, and death in patients in the 29. Matsumori A, on behalf of the assessment of response to candesartan studies of left ventricular dysfunction (SOLVD). J Am Coll Cardiol in heart failure in Japan (ARCH-J) study investigators. Efficacy and 2003; 42: 705 – 708. safety of oral candesartan cilexetil in patients with congestive heart 8. Domanski M, Tian X, Haigney M, Pitt B. Diuretic use, progressive failure. Eur J Heart Fail 2003; 5: 669 – 677 (erratum in: Eur J Heart heart failure, and death in patients in the DIG study. J Card Fail Fail 2004; 6: 523). 2006; 12: 327 – 332. 30. Levy D, Kenchaiah S, Larson MG, Benjamin EJ, Kupka MJ, Ho KK, 9. Hamaguchi S, Kinugawa S, Goto D, Tsuchihashi-Makaya M, Yokota et al. Long-term trends in the incidence of and survival with heart T, Yamada S, et al; JCARE-CARD Investigators. Predictors of long- failure. N Engl J Med 2002; 347: 1397 – 1402. term adverse outcomes in elderly patients over 80 years hospitalized with heart failure: A report from the Japanese Cardiac Registry of Heart Failure in Cardiology (JCARE-CARD). Circ J 2011; 75: 2403 – Appendix 2410. The J-MELODIC Program Committee 10. Seeland U, Kouchi I, Zolk O, Itter G, Linz W, Böhm M. Effect of Executive Committee Members: Tohru Masuyama (Chair), Hyogo Col- ramipril and furosemide treatment on interstitial remodeling in post- lege of Medicine: Hiroshi Ito, Akita University; Nobuyuki Ohte, Nagoya infarction heart failure rat hearts. J Mol Cell Cardiol 2002; 34: 151 – City University; Takashi Akasaka, Wakayama Medical University; 163. Satoshi Nakatani, Osaka University Graduate School of Medicine; Yutaka 11. McCurley JM, Hanlon SU, Wei SK, Wedam EF, Michalski M, Hirano, Kinki University; Kazuhiro Yamamoto, Tottori University. Haigney MC. Furosemide and the progression of left ventricular Endpoint Subcommittee Members: Mitsumasa Ohyanagi (Chair), dysfunction in experimental heart failure. J Am Coll Cardiol 2004; 44: Hyogo College of Medicine; Keiko Takahashi, Hyogo College of Medi- 1301 – 1307. cine; Yutaka Takeda, Nagoya City University; Shin-ichi Fujimoto, Nara 12. Francis GS, Siegel RM, Goldsmith SR, Olivari MT, Levine TB, Cohn Medical University; Toshiaki Mano, Osaka University Graduate School JN. Acute vasoconstrictor response to intravenous furosemide in pa- of Medicine; Yoshio Yasumura, National Hospital Organization Osaka tients with chronic congestive heart failure: Activation of the neuro- National Hospital. humoral axis. Ann Intern Med 1985; 103: 1 – 6. Statistics and Data Management: Hideki Origasa, Motonori Hatta, Advance Publication by J-STAGE MASUYAMA T et al.

University of Toyama School of Medicine; Tatsuyuki Kakuma, Satoshi Shinji Morimoto, Yoshiro Naito, Toshio Naka, Ayumi Nakaboh, Shinji Hattori, Kurume University Biostatistics Center; Takeshi Tsujino, Hyogo Nakao, Mitsumasa Ohyanagi, Katsumi Oka, Takahiro Okumura, Tsuyoshi University of Health Sciences; Masaaki Lee-Kawabata, Hyogo College Sakoda, Hiroki Shimizu, Keiko Takahashi, Takeshi Tsujino, Chikako of Medicine Sasayama Hospital. Yoshida, Masao Yuba; Ibaraki Iseikai Hospital; Fumiaki Nakamura; Data Safety and Monitoring Committee: Akira Kitabatake, Japanese Iseikai Hospital; Hirotaka Murata; Kawasaki Hospital; Toshiaki Hasuike; Circulation Society; Hirohide Matsuo, Shikoku Electric Power Co Inc; Kinki University School of Medicine; Yutaka Hirano, Mitsugu Taniguchi; Toshihiko Morikawa, Kurume University; Takashi Daimon, Hyogo Col- Nagoya City University Graduate School of Medical Sciences; Kaoru lege of Medicine. Asada, Hidekatsu Fukuta, Tatsuya Ito, Seiji Mukai, Nobuyuki Ohte, Investigators of J-MELODIC: Akita University Graduate School of Kazuaki Wakami; Nara Medical University; Shin-ichi Fujimoto; Osaka Medicine; Hiroshi Ito, Masaru Ishida; Higashi Takarazuka Satoh Hospital; University Graduate School of Medicine; Toshiaki Mano, Yasushi Sakata, Shin Takiuchi; Honjyo Daiichi Hospital; Yasushi Suzuki; Hyogo College Kazuhiro Yamamoto; Takarazuka Municipal Hospital; Kenji Hiromoto, of Medicine; Ken-ichi Fujii, Yoshio Furukawa, Akiko Goda, Tetsuzou Keiji Tanabe, Atsunori Ueda; Tsunan Metropolitan Hospital; Shinichiro Kanemori, Daizo Kawasaki, Masaaki Lee-Kawabata, Miho Masai, Moto- Ishikawa; Wakayama Medical University; Takashi Akasaka, Takashi maru Masutani, Mika Matsumoto, Tohru Masuyama, Yoshitomo Mori, Kubo, Takashi Tanimoto.