Relation of Loop Diuretic Dose to Mortality in Advanced Heart Failure

Relation of Loop Diuretic Dose to Mortality in Advanced Heart Failure Shervin Eshaghian, MDa, Tamara B. Horwich, MDb, and Gregg C. Fonarow, MDb,* Although loop diuretics are widely used in heart failure (HF), their effect on outcomes has not been evaluated in large clinical trials. This study sought to determine the dosedependent relation between loop diuretic use and HF prognosis. A cohort of 1,354 patients with advanced systolic HF referred to a single center was studied. Patients were divided into quartiles of equivalent total daily loop diuretic dose: 0 to 40, 41 to 80, 81 to 160, and >160 mg. The cohort was 76% male, with a mean age of 53 ⴞ 13 years and a mean ejection fraction of 24 ⴞ 7%. The mean diuretic dose equivalence was 107 ⴞ 87 mg. The diuretic quartile groups were similar in terms of gender, body mass index, ischemic cause of HF, history of hypertension, and spironolactone use, but the highest quartile was associated with a smaller ejection fraction and lower serum sodium and hemoglobin levels but higher serum blood urea nitrogen and creatinine levels. There was a decrease in survival with increasing diuretic dose (83%, 81%, 68%, and 53% for quartiles 1, 2, 3, and 4, respectively). Even after extensive co-variate adjustment (age, gender, ischemic cause of HF, the ejection fraction, body mass index, pulmonary capillary wedge pressure, peak oxygen consumption, ␤-blocker use, angiotensin-converting enzyme inhibitor or angiotensin receptor blocker use, digoxin use, statin use, serum sodium, blood urea nitrogen, creatinine, hemoglobin, cholesterol, systolic blood pressure, and smoking history), diuretic quartile remained an independent predictor of mortality (quartile 4 vs quartile 1 hazard ratio 4.0, 95% confidence interval 1.9 to 8.4). In conclusion, in this cohort of patients with advanced HF, there was an independent, dose-dependent association between loop diuretic use and impaired survival. Higher loop diuretic dosages identify patients with HF at particularly high risk for mortality. © 2006 Elsevier Inc. All rights reserved. (Am J Cardiol 2006;97:1759 –1764)

There continues to be a “conundrum” in the proper use of diuretics in patients with heart failure (HF).1 Although loop diuretics provide an immediate benefit in congested patients by reducing fluid overload, their long-term effect on the progression of HF is unclear. A recent analysis of patients with chronic HF demonstrated an independent association of large diuretic doses and mortality.2 In addition, in the analyses of the Studies of Left Ventricular Dysfunction, the use of non–potassium-sparing diuretics was associated with increased risk for hospitalization for HF, increased risk for death from HF progression, and increased cardiovascular and all-cause mortality compared with no diuretic therapy or combination therapy with potassium-sparing and non– potassium-sparing diuretics.3 Our study aimed to further investigate the association between loop diuretic dose and HF prognosis in a large cohort of patients with advanced systolic HF of multiple causes.

a

Cedars Sinai Medical Center; and bAhmanson-UCLA Cardiomyopathy Center, Los Angeles, California. Manuscript received October 13, 2005; revised manuscript received and accepted December 21, 2005. This research was supported by the Ahmanson Foundation, Los Angeles, California. Dr. Horwich was supported by Training Grant 401357JI30608 from the National Institutes of Health, Bethesda, Maryland. * Corresponding author: Tel: 310-206-9112; fax: 310-206-9111. E-mail address: [email protected] (G.C. Fonarow). 0002-9149/06/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2005.12.072

Methods Patient population: The study population consisted of 1,354 consecutive patients with advanced systolic HF referred to a single university medical center for HF management and/or transplant evaluation from 1985 to 2004. Patients with left ventricular ejection fractions ⬎40%, those with HF due to valvular disease, and those aged ⬍18 years were excluded from the analysis. All patients were followed in a comprehensive HF management program, as previously described.4 This study was approved by the University of California, Los Angeles, Medical Institutional Review Board. Data collection: Detailed information on patients’ baseline characteristics, including medication and doses, was recorded at their initial visits. The total daily dose of loop diuretics at baseline was assessed for each patient. The formula used to convert other loop diuretics to furosemide equivalents was as follows: furosemide 80 mg ⫽ torsemide 40 mg ⫽ bumetanide 3 mg ⫽ ethacrynic acid 50 mg. Additional as-needed dosing was not included in the daily loop diuretic total. Laboratory testing, echocardiography, and right-sided cardiac catheterization occurred ⬍6 weeks after the initial referral date. The estimated glomerular filtration rate was calculated on the basis of the CockcroftGault formula. Body mass index was calculated using the patients’ “dry” weights after empiric or pulmonary artery catheter-guided HF therapy. Previous left-sided cardiac www.AJConline.org

1760

The American Journal of Cardiology (www.AJConline.org)

Table 1 Baseline characteristics of the study cohort divided by diuretic quartile Variable

Age (yrs) Men NYHA class III/IV Ejection fraction Left ventricular end-diastolic dimension (mm) Severe mitral regurgitation Ischemic cause of HF Smoking history Hypertension Body mass index (kg/m2) Peak oxygen consumption (ml/kg/min) Systolic blood pressure (mm Hg) Pulmonary capillary wedge pressure (mm Hg) Cardiac index (L/min ⫻ m2) Serum sodium (mmol/L) Blood urea nitrogen (mg/dl) Creatinine (mg/dl) Estimated glomerular filtration rate Albumin (g/dl) Hemoglobin (g/dl) Total cholesterol (mg/dl) ACE-l or ARB ␤ blocker Statin Spironolactone Digoxin

Quartile

p Value

Total Cohort (n ⫽ 1,354)

1 (0–40 mg) (n ⫽ 465)

2 (41–80 mg) (n ⫽ 365)

3 (81–160 mg) (n ⫽ 320)

4 (⬎160 mg) (n ⫽ 204)

53 ⫾ 13 1,029 (76%) 1,164 (86%) 24 ⫾ 7 70 ⫾ 11 339 (25%) 636 (47%) 826 (61%) 542 (40%) 27 ⫾ 5 14 ⫾ 5 104 ⫾ 18 15 ⫾ 6 2.6 ⫾ 0.9 136 ⫾ 4 28 ⫾ 18 1.4 ⫾ 0.8 80 ⫾ 35 3.9 ⫾ 0.6 13 ⫾ 2 174 ⫾ 55 1,219 (90%) 569 (42%) 501 (37%) 352 (26%) 948 (70%)

54 ⫾ 12 344 (74%) 367 (79%) 24 ⫾ 7 67 ⫾ 11 88 (19%) 214 (46%) 274 (59%) 177 (38%) 26 ⫾ 5 14 ⫾ 5 107 ⫾ 18 14 ⫾ 5 2.7 ⫾ 0.7 137 ⫾ 4 23 ⫾ 15 1.3 ⫾ 0.9 80 ⫾ 32 3.9 ⫾ 0.6 13 ⫾ 2 180 ⫾ 54 423 (91%) 251 (54%) 191 (41%) 130 (28%) 293 (63%)

53 ⫾ 13 270 (74%) 318 (87%) 24 ⫾ 7 70 ⫾ 10 91 (25%) 175 (48%) 208 (57%) 157 (43%) 27 ⫾ 5 14 ⫾ 5 104 ⫾ 18 15 ⫾ 6 2.7 ⫾ 1.3 137 ⫾ 4 28 ⫾ 17 1.4 ⫾ 0.6 78 ⫾ 34 3.9 ⫾ 0.6 14 ⫾ 2 180 ⫾ 54 332 (91%) 150 (41%) 135 (37%) 91 (25%) 270 (74%)

53 ⫾ 13 250 (78%) 282 (88%) 23 ⫾ 6 71 ⫾ 10 86 (27%) 144 (45%) 202 (63%) 125 (39%) 27 ⫾ 6 13 ⫾ 4 104 ⫾ 18 16 ⫾ 6 2.6 ⫾ 0.7 136 ⫾ 5 31 ⫾ 20 1.4 ⫾ 0.9 80 ⫾ 38 3.9 ⫾ 0.6 13 ⫾ 2 167 ⫾ 57 288 (90%) 115 (36%) 112 (35%) 83 (26%) 237 (74%)

53 ⫾ 13 161 (79%) 196 (96%) 23 ⫾ 6 70 ⫾ 10 63 (31%) 108 (53%) 137 (67%) 80 (39%) 27 ⫾ 5 13 ⫾ 5 100 ⫾ 17 17 ⫾ 6 2.5 ⫾ 0.7 135 ⫾ 5 32 ⫾ 20 1.5 ⫾ 1.0 79 ⫾ 37 3.8 ⫾ 0.6 13 ⫾ 2 163 ⫾ 54 169 (83%) 169 (23%) 57 (28%) 43 (21%) 139 (68%)

0.91 0.39 0.0001 0.03 0.0001 0.01 0.41 0.06 0.43 0.25 0.02 0.0001 0.0001 0.15 0.0001 0.0001 0.04 0.28 0.15 0.03 0.0001 0.01 0.0001 0.01 0.14 0.002

Data are presented as mean ⫾ SD or percentage of patients. ACE-I ⫽ angiotensin-converting enzyme inhibitor; ARB ⫽ angiotensin receptor blocker, NYHA ⫽ New York Heart Association.

catheterization reports and angiographic films were reviewed, or, if not done previously, left-sided cardiac catheterization was performed. Significant coronary artery disease was defined as any single stenosis ⬎70% of the cross-sectional luminal diameter of the involved artery on angiography. All-cause mortality was the primary end point of the study. Heart transplantation (statuses IA, IB, and II) was coded as a nonfatal end of follow-up at the time of transplantation, as previously described.4 The composite end point of death or urgent transplant (status IA) was analyzed as a secondary end point.5 Death was considered sudden if it was unexpected on the basis of the patient’s clinical status and if it occurred out of the hospital ⬍15 minutes after the onset of unexpected symptoms or during sleep. Death during hospitalization for worsening congestive symptoms was considered death due to HF. Statistical analysis: Patients were grouped on the basis of loop diuretic dose equivalence: quartile 1, 0 to 40 mg (n ⫽ 465); quartile 2, 41 to 80 mg (n ⫽ 365); quartile 3, 81 to 160 mg (n ⫽ 320); and quartile 4, ⬎160 mg (n ⫽ 204). Because there were a large number patients receiving diuretic doses at each of the quartile dividing points of 40, 80, and 160 mg, the number of patients in each group significantly differed. Results are presented as mean ⫾ SD for

continuous variables and as percentages of the total for categorical variables. The independent-samples Student’s t test, analysis of variance, and the chi-square test were used for the comparison of variables when appropriate. One- and 2-year product-moment survival estimates were calculated using the Kaplan-Meier method, and differences between the curves were evaluated with the log-rank statistic. Cox proportional-hazards modeling was performed to analyze the risk associated with diuretic dose, in addition to a variety of other prospectively recorded parameters, to survival or urgent transplant-free survival. A forward method was used for the multivariate model. SPSS for Windows version 14.0 (SPSS, Inc., Chicago, Illinois) was used for all analyses. Results Baseline characteristics of the cohort: The cohort was 76% male and ranged in age from 18 to 84 years (mean 53 ⫾ 13). New York Heart Association class III and IV HF constituted 39% and 47% of the population, respectively. The mean left ventricular ejection fraction was 24 ⫾ 7% (Table 1). Causes of HF were ischemic (47%) and idiopathic (33%); the remaining causes included alcohol induced, hypertrophic, and postpartum cardiomyopathy.

Heart Failure/Diuretics and Mortality in Heart Failure Table 2 Characteristics of survivors and nonsurvivors at 2 years

Age (yrs) Men NYHA class III or IV Ejection fraction Left ventricular end diastolic dimension (mm) Severe mitral regurgitation Ischemic cause of HF Smoking history Hypertension Body mass index (kg/m2) Peak oxygen consumption (ml/kg/min) Systolic blood pressure (mm Hg) Pulmonary capillary wedge pressure (mm Hg) Cardiac index (L/min ⫻ m2) Serum sodium (mmol/L) Blood urea nitrogen (mg/dl) Creatinine (mg/dl) Estimated glomerular filtration rate Albumin (g/dl) Hemoglobin (g/dl) Total cholesterol (mg/dl) ACE-l or ARB ␤ blocker Statin Spironolactone Digoxin Loop diuretic mean dose (mg)

100

Survivors (n ⫽ 1,085)

Nonsurvivors (n ⫽ 269)

p Value

53 ⫾ 12 814 (75%) 911 (84%) 24 ⫾ 7 70 ⫾ 10

56 ⫾ 14 207 (77%) 253 (94%) 23 ⫾ 7 70 ⫾ 11

0.0001 0.25 0.0001 0.94 0.75

80

250 (23%) 477 (44%) 651 (60%) 423 (39%) 27 ⫾ 5 14 ⫾ 5

83 (31%) 164 (61%) 175 (65%) 116 (43%) 26 ⫾ 6 12 ⫾ 4

0.01 0.0001 0.08 0.09 0.47 0.0001

105 ⫾ 18

101 ⫾ 16

0.001

15 ⫾ 5

17 ⫾ 6

0.0001

2.6 ⫾ 1.0 137 ⫾ 4 26 ⫾ 16 1.4 ⫾ 0.9 82 ⫾ 35

2.5 ⫾ 0.7 135 ⫾ 5 35 ⫾ 22 1.5 ⫾ 0.6 69 ⫾ 34

0.13 0.0001 0.0001 0.002 0.0001

3.9 ⫾ 0.6 14 ⫾ 2 177 ⫾ 54 987 (91%) 488 (45%) 412 (38%) 282 (26%) 770 (71%) 98

3.7 ⫾ 0.6 13 ⫾ 2 163 ⫾ 57 221 (82%) 81 (30%) 81 (30%) 59 (22%) 178 (66%) 140

0.0001 0.01 0.001 0.0001 0.0001 0.01 0.10 0.08 0.0001

Data are presented as mean ⫾ SD or percentage of patients. Abbreviations as in Table 1.

Diuretic dose and patient characteristics: There were 465 patients (34%) in the first quartile, with dose equivalence of 0 to 40 mg; 365 patients (27%) in the second quartile, with dose equivalence of 40 to 80 mg; 320 patients (24%) in the third quartile, with dose equivalence of 80 to 160 mg; and 204 patients (15%) in the fourth quartile, with dose equivalence of ⬎160 mg. In this cohort, 1,257 patients (93%) were treated with furosemide, 55 (4%) with torsemide, and 42 (3%) with bumetanide. Differences in patient characteristics among the loop diuretic quartiles were analyzed (Table 1). There were no statistically significant differences among the diuretic dose quartiles in terms of age, gender, body mass index, ischemic cause of HF, history of hypertension, cardiac index, serum albumin, estimated glomerular filtration rate, and spironolactone use. However, the highest diuretic dose quartile was associated with a smaller ejection fraction; lower blood pressure, serum sodium, hemoglobin, and total cholesterol; and less angiotensin-converting enzyme inhibitor or angiotensin receptor blocker, ␤ blocker, digoxin, and statin use.

Survival (%)

Variable

1761

60 40

0-40 (n=465)

P < 0.0001

40-80 (n=365) 80-160 (n=320)

20

>160 (n=204)

0 0

2

4

6

8

10 12 14 16 18 20 22 24

Time (months) Figure 1. Kaplan-Meier survival estimates for loop diuretic dose quartiles in patients with advanced systolic HF over 2-year follow-up.

Hazard Ratios and 95% Confidence Interval for Endpoints

Death From Any Cause HR 3.4 (95% CI 2.4-4.7) Death or Urgent Transplant HR 2.7 (95% CI 2.0-3.5) Progressive Heart Failure Death HR 3.8 (95% CI 2.1-6.8) Sudden Death HR 3.6 (95% CI 1.9-6.8)

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

Highest Diuretic Quartile Worse

Figure 2. Two-year HRs and 95% CIs for death from any cause, death or urgent transplantation, progressive HF death, and sudden death for the highest quartile of loop diuretic versus the lowest quartile.

Relation between diuretic dose and mortality: There were 269 deaths during the 2-year follow-up, with 182 deaths by year 1 and 87 deaths during year 2. Of the 269 deaths, progressive HF death accounted for 91 of the deaths (34%), whereas 72 deaths (27%) were sudden, 8 (3%) occurred secondary to myocardial infarction, and 101 (34.9%) occurred from unclassifiable, unknown, or other causes. Of the 1,354 patients followed, 431 received heart transplants by the end of the 2-year follow-up: 223 urgent (status IA) and 208 elective (status IB or II). No patients were lost to follow-up. Table 2 lists the characteristics for 2-year survivors compared with nonsurvivors. The mean diuretic dose was significantly higher in nonsurvivors than survivors. There was a stepwise linear decrease in survival with increasing diuretic dose. Patients with HF in the highest diuretic dose quartile were found to have significantly impaired survival compared with patients in the lowest quartile. Survival estimates at 1 year were 91%, 88%, 80%, and 69% for

1762

The American Journal of Cardiology (www.AJConline.org)

Table 3 Hazard ratio of death for increasing diuretic dose at one or two years: univariate and multivariate analysis Variable

1-yr 1-yr 1-yr 1-yr 2-yr 2-yr 2-yr 2-yr

Quartile 1 (0–40 mg) (n ⫽ 465)

2 (41–80 mg) (n ⫽ 365)

3 (81–160 mg) (n ⫽ 320)

4 (⬎160 mg) (n ⫽ 204)

37 (8%) 9% 1.0 1.0 60 (13%) 17 % 1.0 1.0

40 (11%) 12% 1.4 (0.9–2.2) 1.6 (0.6–4.4) 58 (16%) 19% 1.2 (0.8–1.7) 1.2 (0.6–2.5)

55 (17%) 20% 2.4 (1.6–3.6) 2.2 (0.8–6.0) 80 (25%) 32% 2.1 (1.5–2.9) 2.6 (1.3–5.2)

53 (26%) 31% 3.8 (2.5–5.8) 4.2 (1.5–11.3) 71 (35%) 47% 3.4 (2.4–4.7) 4.0 (1.9–8.4)

Death, n mortality unadjusted HR (95% CI) multivariate HR* (95% CI) Death, n mortality rate unadjusted HR (95% CI) multivariate HR* (95% CI)

* Multivariate analysis included age, gender, ischemic cause of HF ejection fraction, body mass index, ␤ blocker use, angiotensin-converting enzyme inhibitor or angiotensin receptor blocker use, digoxin use, statin use, serum sodium, blood urea nitrogen, creatinine, hemoglobin, total cholesterol, blood pressure, smoking history, and peak oxygen consumption.

Quartile 1

Two Year Mortality, %

70

P<0.0001 P<0.0001 P<0.0001

Quartile 3

Quartile 2

P<0.0001 P<0.0001 P<0.0001

Quartile 4

P=0.01

P=0.001 P=0.006

60 50 40 30 20 10

5)

6)

40 (n =

(n = PC W

P>

15

15 P<

e> in in at

Cr e

PC W

1.

5

5 1. e< in in at

Cr e

52

29

(n =

(n =

98

=7 1 (n AD

C No

9)

3)

3)

1)

) =3

31

CA D

(n

(n =6 4

23 W

om en

(n =1 0 en

M

To

ta

l( n=

13

54

)

)

0

Figure 3. Two-year mortality rates in the different loop diuretic dose quartiles in the cohort of patients with advanced systolic HF. The increase in mortality was compared with subgroups of men and women, patients with coronary artery disease (CAD) and those without CAD, creatinine ⬍1.5 or ⬎1.5 mg/dl, and pulmonary capillary wedge pressure (PCWP) ⬍15 or ⬎15 mm Hg.

quartiles 1, 2, 3, and 4, respectively (p ⬍0.0001). Survival estimates at 2 years were 83%, 81%, 68%, and 53%, respectively (p ⬍0.0001) (Figure 1). The hazard ratios (HRs) for death from any cause, death and urgent transplantation, death from progressive HF, and sudden death for quartile 4 compared with quartile 1 were similar (Figure 2). On univariate analysis, compared with the lowest quartile, increasing loop diuretic dose quartiles were associated with a progressive increase in mortality (second quartile, HR 1.2, 95% confidence interval [CI] 0.8 to 1.7; third quartile, HR 2.1, 95% CI 1.5 to 2.9; and fourth quartile, HR 3.4, 95% CI 2.4 to 4.7). Diuretic dose quartiles were associated with increased mortality independent of other covariates. After adjustment for age, gender, ischemic cause of HF, the ejection fraction, pulmonary capillary wedge pressure, peak oxygen consumption, blood pressure, smoking

history, serum sodium, blood urea nitrogen, creatinine, hemoglobin, cholesterol level, ␤-blocker use, angiotensinconverting enzyme inhibitor use, digoxin use, and statin use, the highest diuretic quartile remained a significant predictor of increased mortality at 1 year (HR 4.2, 95% CI 1.5 to 11.3) and at 2 years (HR 4.0, 95% CI 1.9 to 8.4) (Table 3). Furthermore, Kaplan-Meier survival estimate curves were assessed in clinically relevant subgroups to determine whether the increased loop diuretic dose quartile maintained its prognostic value. The association between increased loop diuretic dose and mortality persisted in men and women, patients with and without coronary artery disease, those with high or low pulmonary capillary wedge pressure, and those with creatinine ⬍1.5 or ⬎1.5 mg/dl (Figure 3). Discussion This study suggests that in patients with advanced systolic HF, the use of higher doses of loop diuretics is associated with significantly increased all-cause mortality. We observed that the highest diuretic doses (⬎160 mg) were associated with a significant increase in 1- and 2-year allcause mortality compared with the lowest loop diuretic doses (0 to 40 mg). The association was not limited to the mode of death; death from any cause, death and urgent transplantation, progressive HF death, and sudden death were all significantly increased in patients taking the highest doses of loop diuretics. Although it may appear obvious that patients with HF requiring higher loop diuretic doses to prevent fluid retention and control symptoms might be sicker than patients receiving lower doses, the powerful and independent association with mortality warrants further consideration. Over the past few decades, clinical trials have altered the management of systolic HF with the introduction of medical therapy shown to reduce morbidity and mortality in HF, including angiotensin-converting enzyme inhibitors, ␤ blockers, and aldosterone.6 – 8 Of all the medications used in

Heart Failure/Diuretics and Mortality in Heart Failure

HF, loop diuretics are the most efficacious in relieving clinical symptoms of shortness of breath and signs of peripheral edema.9 –11 Diuretics also improve cardiac performance by reducing left ventricular preload. The diureticinduced reductions in ventricular filling pressures and ventricular cavity size, decreases in mitral regurgitation, and resulting increases in forward cardiac output may all be favorable in HF.12 When administered alone or in combination with other short-term anti-HF drugs, diuretics improve patients’ physical activity levels and overall assessments of the quality of life.12–14 However, the impact of loop diuretics on the prognosis of patients with HF has not been assessed in a large-scale, randomized clinical trial. Loop diuretics’ activation of the renin-angiotensin-aldosterone system and sympathetic nervous system has been shown to play an important role in the pathophysiology of HF and may be associated with the progression of HF.15–18 When diuretics are prescribed to patients who do not have significant fluid retention, intravascular volume contraction and decreased left ventricular filling pressure may worsen cardiac performance. Loop diuretics, especially furosemide, activate the renin-angiotensin-aldosterone system in patients with HF.11,19 Ikram et al20 demonstrated significant neurohormonal activation by demonstrating a significant increase in plasma renin, angiotensin II, and aldosterone concentration as a result of the administration of loop diuretics. In addition, Bayliss et al11 reported that neither plasma renin nor aldosterone was increased in 12 subjects with untreated symptomatic HF, whereas there was substantial neurohumoral activation after 1 month of treatment with furosemide. Prolonged aldosterone exposure has deleterious effects on left ventricular function, causing reactive myocardial fibrosis and a variety of other potentially adverse effects.21,22 Furthermore, loop diuretics are associated with significant decreases in the glomerular filtration rate and may contribute to cardiorenal syndrome.23 It is now recognized that the kidneys play a vital role in the prognosis of patients with HF, and increases in creatinine or blood urea nitrogen are known predictors of mortality. Thus, it could also be hypothesized that larger doses of loop diuretics result in unfavorable outcomes in patients with HF because of adverse effects on renal function. Although no large, randomized clinical trials have examined the role of loop diuretics on HF prognosis, retrospective analyses of the Studies of Left Ventricular Dysfunction data have examined the association of diuretics and mortality in patients with mild to moderate systolic HF. Domanski et al3 demonstrated that compared with patients not taking any diuretics, the risk for hospitalization or death due to worsening HF was significantly increased in patients taking non–potassium-sparing diuretics. Moreover, Cooper et al24 demonstrated that the use of non–potassium-sparing diuretics was associated with an increased risk for cardiovascular death, arrhythmic death, and all-cause mortality.24 The association of larger doses of loop diuretics with poor

1763

outcomes was demonstrated in an analysis of patients enrolled in the Prospective Randomized Amlodipine Survival Evaluation trial, in which significant associations between larger-than-median doses (⬎80 mg) of loop diuretics and all-cause mortality and sudden death but not HF death were observed.2 Recently, McCurley et al25 demonstrated in an experimental model of HF that animals treated with furosemide developed left ventricular dysfunction and increases in serum aldosterone levels significantly faster than animals receiving placebo. There are thus a variety of pathophysiologic mechanisms by which loop diuretics may accelerate the progression of HF and contribute to an increased mortality risk. Observational data and experimental models of HF support a deleterious effect of loop diuretics. Although the previous studies examined and delineated an association between loop diuretic use and HF prognosis, to our knowledge, this is the first study to report a stepwise, dose-dependent effect of loop diuretics on mortality. We acknowledge a number of potential limitations of our study. The cohort was a select population of patients with advanced systolic HF referred to a single university center. These patients were significantly younger and differed in other characteristics from patients with HF in community cohort studies and hospital registries. Diuretic dose was examined at only a single point in time, and clearly, many patients undergo significant up- or down-titration of diuretic doses over time. However, this would have served only to diminish the chance of observing differences in outcomes by loop diuretic dose at the time of referral. There were significant differences in baseline characteristics and other HF treatments among the diuretic dose quartiles. Even after adjustment for multiple covariates, larger loop diuretic doses could still be a surrogate for other measured and unmeasured variables that reflect more severe HF. There may be interactions between different variables and clinical outcomes that multivariate methods cannot adjust for. In addition, information regarding serum potassium and magnesium levels was not available. Propensity matching was not performed. Therefore, whether the relation between loop diuretic dose and increased mortality is causative is unknown. The present findings warrant further investigations into the relation between diuretic dose and prognosis in patients with HF. 1. Gupta S, Neyses L. Diuretic usage in heart failure: a continuing conundrum in 2005. Eur Heart J 2005;7:644 – 649. 2. Neuberg GW, Miller AB, O’Connor CM, Belkin RN, Carson PE, Cropp AB, Frid DJ, Nye RG, Pressler ML, Wertheimer JH, Packer M. Diuretic resistance predicts mortality in patients with advanced heart failure. Am Heart J 2002;144:31–38. 3. Domanski M, Norman J, Pitt B, Haigney M, Hanlon S, Peyster E. Diuretic use, progressive heart failure, and death in patients in the Studies of Left Ventricular Dysfunction (SOLVD). J Am Coll Cardiol 2003;42:705–708. 4. Fonarow GC, Stevenson LW, Walden JA, Livingston NA, Steimle AE, Hamilton MA, Moriguchi J, Tillisch JH, Woo MA. Impact of a comprehensive heart failure management program on hospital read-

1764

5.

6.

7.

8.

9.

10.

11.

12.

13. 14. 15.

The American Journal of Cardiology (www.AJConline.org)

mission and functional status of patients with advanced heart failure. J Am Coll Cardiol 1997;30:725–731. Aaronson KD, Mancini DM. Mortality remains high for outpatient transplant candidates with prolonged (⬎6 months) waiting list time. J Am Coll Cardiol 1999;33:1189 –1195. The SOLVD Investigators. Effect of enalapril on survival in patients with reduced left ventricular ejection fractions and congestive heart failure. N Engl J Med 1991;325:293–302. Packer M, Bristow MR, Cohn JN, Colucci WS, Fowler MB, Gilbert EM, Shusterman NH, U.S. Carvedilol Heart Failure Study Group. The effect of carvedilol on morbidity and mortality in patients with chronic heart failure. N Engl J Med 1996;334:1349 –1355. Pitt B, Zannad F, Remme WJ, Cody R, Castaigne A, Perez A, Palensky J, Wittes J. The effect of spironolactone on morbidity and mortality in patients with severe heart failure. N Engl J Med 1999;341:709 –717. Stampfer M, Epstein SE, Beiser GD, Braunwald E. Hemodynamic effects of diuretics at rest and during intense upright exercise in patients with impaired cardiac function. Circulation 1968;37:900 – 911. Anand IS, Kalra GS, Harris P, Poole-Wilson PA, Panzali A, De Giuli F, Ferrari R. Diuretics as essential and sole treatment in chronic heart failure. Cardioscience 1997;18:852– 857. Bayliss J, Norell M, Canepa-Anson R, Sutton G, Poole-Wilson P. Untreated heart failure: clinical and endocrine effects of introducing diuretics. Br Heart J 1987;57:17–22. Stevenson LW, Brunken RC, Belil D, Grover-McKay M, Schwaiger M, Schelbert HR, Tillisch JH. Afterload reduction with vasodilators and diuretics decreases mitral regurgitation during upright exercise in advanced heart failure. J Am Coll Cardiol 1990;15:174 –180. Silke B. Central hemodynamic effects of diuretic therapy in chronic heart failure. Cardiovasc Drug Ther 1993;7(suppl):45–53. Anand IS, Florea VG. Diuretics in chronic heart failure-benefits and hazards. Eur Heart J 2001;3(suppl):1–18. Benedict CR, Weiner DH, Johnstone DE, Bourassa MG, Ghali JK, Nicklas J, Kirlin P, Greenberg B, Quinones MA, Yusuf S. Comparative neurohormonal responses in patients with preserved and impaired left ventricular ejection fraction: results of the Studies of Left Ven-

16.

17.

18.

19.

20.

21.

22. 23.

24.

25.

tricular Dysfunction (SOLVD) registry. J Am Coll Cardiol 1993; 22(suppl):146 –153. Dzau VJ, Colucci WS, Hollenberg NK, Williams GH. Relations of the rennin-angiotensin-aldosterone system to clinical state in congestive heart failure. Circulation 1981;63:645– 651. Packer M. The neurohormonal hypothesis: a theory to explain the mechanism of disease progression in heart failure. J Am Coll Cardiol 1992;20:248 –254. Swedberg K, Eneroth P, Kjekshus J, Wilhelmsen L. Hormones regulating cardiovascular function in patients with sever congestive heart failure and their relation to mortality. CONSENSUS Trial Study Group. Circulation 1990;82:1730 –1736. Francis GS, Benedict C, Johnstone DE, Kirlin PC, Nicklas J, Liang CS, Kubo SH, Rudin-Toretsky E, Yusuf S. Comparison of neuroendocrine activation in patients with left ventricular dysfunction with and without congestive heart failure. Circulation 1990;82:1724 –1729. Ikram H, Chan W, Espiner EA, Nicholls MG. Hemodynamic and hormone responses to acute and chronic frusemide therapy in congestive heart failure. Clin Sci 1980;59:443– 449. Weber KT, Brilla CG. PA. Pathologic hypertrophy and cardiac interstitium; fibrosis and rennin-angiotensin-aldosterone system. Circulation 1991;83:1849 –1865. Lijnen P, Petrov. Induction of cardiac fibrosis by aldosterone. J Mol Cell Cardiol 2000;32:865– 879. Gottlieb SS, Brater DC, Thomas I, Havranek E, Bourge R, Goldman S, Dyer F, Gomez M, Bennett D, Ticho B, et al. BG9719 (CVT-124), an A1 adenosine receptor antagonist, protects against the decline in renal function observed with diuretic therapy. Circulation 2002;105:1348 – 1353. Cooper HA, Dries DL, Davis CE, Shen YL, Domanski MJ. Diuretics and risk of arrhythmic death inpatients with left ventricular dysfunction. Circulation 1999;100:1311–1315. McCurley JM, Hanlon SU, Wei SK, Wedam EF, Michalski M, Haigney MC. Furosemide and the progression of left ventricular dysfunction in experimental heart failure. J Am Coll Cardiol 2004;44: 1301–1307.