Author + information
- Received October 16, 2002
- Revision received March 19, 2003
- Accepted March 27, 2003
- Published online August 20, 2003.
- Michael Domanski, MD*,* (, )
- James Norman, PhD*,
- Bertram Pitt, MD†,
- Mark Haigney, MD‡,
- Stephen Hanlon, MD‡ and
- Eliot Peyster*
- ↵*Reprint requests and correspondence:
Dr. Michael Domanski, National Heart, Lung, and Blood Institute, 6701 Rocklege Avenue, Bethesda, Maryland 20892-7936, USA.
Objectives We sought to determine whether non–potassium-sparing diuretics (PSDs) in the absence of a PSD may result in progressive heart failure (HF).
Background Angiotensin-converting enzyme (ACE) inhibitors incompletely suppress ACE activity in HF patients. Furthermore, non-PSDs are activators of aldosterone secretion. We reasoned that non-PSDs, in the absence of a PSD, might result in progressive HF.
Methods In the 6,797 patients in the Studies Of Left Ventricular Dysfunction (SOLVD), we compared the risk of hospitalization for, or death from, HF between those taking a PSD and those who were not, adjusting for known covariates.
Results The risk of hospitalization from worsening HF in those taking a PSD relative to those taking only a non-PSD was 0.74 (95% confidence interval [CI] 0.55 to 0.99; p = 0.047). The relative risk for cardiovascular death was 0.74 (95% CI 0.59 to 0.93; p = 0.011), for death from all causes 0.73 (95% CI 0.59 to 0.90; p = 0.004), and for hospitalization for, or death from, HF 0.75 (95% CI 0.58 to 0.97; p = 0.030). Compared with patients not taking any diuretic, the risk of hospitalization or death due to worsening HF in patients taking non-PSDs alone was significantly increased (risk ratio [RR] = 1.31, 95% CI 1.09 to 1.57; p = 0.0004); this was not observed in patients taking PSDs with or without a non-PSD (RR = 0.99, 95% CI 0.76 to 1.30; p = 0.95).
Conclusions The use of PSDs in HF patients is associated with a reduced risk of death from, or hospitalization for, progressive HF or all-cause or cardiovascular death, compared with patients taking only a non-PSD.
Pharmacologic antagonists of the neurohormonal response to heart failure (HF) reduce death due to progressive left ventricular dysfunction (1–7). The Randomized ALdactone Evaluation Study (RALES) demonstrated reduced HF mortality with complete suppression of aldosterone secretion effected by adding spironolactone to an angiotensin-converting enzyme (ACE) inhibitor (3).
Sudden cardiac death is increased when non-PSDs are used without PSDs in HF (8), even when no clinically significant difference in serum potassium is present in patients taking PSDs, compared with those only taking non-PSDs. Because aldosterone secretion causes pathologic remodeling, treatment with non-PSD alone likely increased the underlying arrhythmogenic substrate. We reasoned that non-PSDs, in the absence of a PSD, might result in progressive HF. This study compares the rates of death or HF hospitalization in the Studies Of Left Ventricular Dysfunction (SOLVD) between patients taking a PSD at randomization and those who were not.
The SOLVD study has been described previously (9,10). Briefly, 6,797 symptomatic and asymptomatic patients with a left ventricular ejection fraction <0.36 were randomly assigned to double-blinded treatment with enalapril or placebo. Only drug class was ascertained; specific medications were not recorded.
Incidence rates were compared using the result that if X and Y have Poisson distributions, the conditional distribution of X, given X + Y = n is binomial with parameters n and p = PYR(X)/(PYR[X] + PYR[Y]) for equal incidence rates. The Cox proportional hazards regression model was applied to estimate hazard ratios for hospitalization and/or death from worsening heart failure and death from cardiovascular disease and any cause. The models included those co-variates recorded at baseline that were judged as likely predictors of these events (Table 1). The rates per 100 person-years of follow-up for each of the previously defined events were calculated for each of the four patterns of diuretic use: 1) no diuretic use (n = 3,915) (omitted in later analyses); 2) PSD only (n = 68); and 3) non-PSD only (n = 2469); and 4) both types of diuretic (n = 338). Group 2 was too small to permit a meaningful test of an interactive effect from taking both types of diuretic. The definitive comparison for evaluating the effect of using a PSD was made by estimating the risks of those patients in groups 2 and 4 relative to those in group 3, using the aforementioned Cox model. Analyses were performed using version 6.12 of PC SAS (SAS Institute, Cary, North Carolina).
At baseline (Table 1), compared with patients not taking any diuretic, patients taking diuretics were older and more often female, had a lower left ventricular ejection fraction, higher New York Heart Association functional class, and higher creatinine levels, and were more frequently taking digitalis and anti-arrhythmic drugs.
Distribution of events by diuretic use
Rates of hospitalization for HF, death from cardiovascular disease, death from all causes, and either hospitalization or death due to worsening HF in all baseline diuretic use pattern groups appear in Table 2. All are significantly higher for patients taking a diuretic than for those who were not (p < 0.01 for all events). The event rates for those patients taking a PSD are significantly lower than for those taking only a non-PSD (p < 10−5to p = 0.0008).
Table 3shows the risks for the same set of events as in Table 2, after adjusting for known covariates. Compared with patients not taking any diuretic, the risk of hospitalization for, or death from, worsening HF in patients taking non-PSDs alone was significantly increased (risk ratio [RR] = 1.31, 95% confidence interval [CI] 1.09 to 1.57; p = 0.004). In patients taking PSDs with or without a non-PSD, this relative risk was not significantly elevated (RR = 0.99, 95% CI 0.76 to 1.30; p = 0.71). Table 4shows the risks (adjusted for the covariates listed earlier) for the four outcomes in Table 3in patients taking a PSD relative to those taking only a non-PSD. They show significant reductions in the risks of hospitalization, death, or both, due to worsening HF, as well as death from cardiovascular disease and all causes, when the patient is taking a PSD.
This study shows that, in patients with moderate or severe left ventricular dysfunction, the use of a PSD is associated with a reduced risk of death or hospitalization due to progressive HF, relative to patients taking only a non-PSD.
In HF, initially compensatory molecules are released, eventually contributing to progressive left ventricular dysfunction (1,2,4–6). An important component of this neurohormonal response is activation of the renin-angiotensin-aldosterone system. In HF patients, there is as much as a 20-fold increase in the aldosterone level (11,12). Aldosterone causes reactive perivascular and interstitial myocardial fibrosis (13)that can impair both systolic and diastolic function (12,14)and serve as a substrate for ventricular arrhythmias (5,14,15). These effects can be prevented by spironolactone (1,14). In the RALES substudy, the greatest benefit from spironolactone occurred in the patients with the highest baseline measures of collagen synthesis and cardiac fibrosis (16). Other PSDs that are not direct antagonists of aldosterone also appear to have beneficial myocardial effects. Amiloride, for instance, prevents aldosterone-mediated myocyte necrosis and replacement fibrosis (17). The exact etiology of these positive effects remains unclear; it may be due to a direct effect on the Na+/H+exchanger, which amiloride antagonizes, versus prevention of intracellular potassium depletion. If the maintenance of potassium homeostasis is the critical mechanism behind these findings, then potassium depletion secondary to non-PSDs, per se, represents a potential effector of myocardial necrosis and fibrosis, which PSDs could prevent, whether they directly antagonize aldosterone or not.
Elevated aldosterone levels occur in HF patients, despite treatment with ACE inhibitors (2,18–21). Aldosterone elimination is reduced in HF due to reduced hepatic clearance (22,23). Also, tolerable ACE inhibitor doses may not fully suppress aldosterone secretion (2,18–21). Further, peripheral mechanisms that are not suppressed by ACE inhibitors contribute to angiotensin II production (24,25). Importantly, activation of angiotensin-independent aldosterone secretion is stimulated by sodium losses caused by loop diuretics (25,26). This mechanism could explain the role of non-PSDs in causing progressive HF and, therefore, the therapeutic benefit of aldosterone antagonists and other PSDs.
This is the first report demonstrating increased mortality from progressive HF in patients treated with only a non-PSD, compared with those treated with a PSD. The RALES study demonstrated that all-cause mortality and progressive HF mortality are reduced by treatment with the aldosterone antagonist spironolactone (3). This beneficial effect was additive to ACE inhibitor treatment, consistent with the failure of ACE inhibitors to fully suppress aldosterone production.
The American College of Cardiology/American Heart Association guidelines for HF treatment recommend diuretics for treating fluid overload (27). Our data suggest that HF treatment with non-PSDs may cause more rapid HF progression, and that this can be aborted by co-treatment with a PSD. Although many patients should be taking spironolactone because of the RALES results, a PSD should be added to the regimen of other patients being treated with a non-PSD.
This study is retrospective and, therefore, not definitive proof that non-PSDs cause progressive HF. However, the SOLVD database is extensive, and all available co-variates and confounders were considered in the multivariate analysis. The number of patients taking only a PSD was too small to provide a powerful test of the interaction effect of taking both diuretic types on the risk of the end points considered. Nevertheless, we believe that the results of our comparisons of patients taking both types of diuretics with those taking only non-PSDs are robust.
Because the diuretic dosage was not available, we cannot draw conclusions about a dose-response relationship. Also, baseline data were used, and diuretic treatment status may have changed over time. This effect, however, would tend to make the groups look more, rather than less, alike.
The use of PSDs in HF patients is associated with a reduced risk of death from, or hospitalization for, progressive HF, all-cause death, or cardiovascular death, compared with patients taking only a non-PSD.
- angiotensin-converting enzyme
- confidence interval
- heart failure
- potassium-sparing diuretic
- Randomized ALdactone Evaluation Study
- risk ratio
- Studies Of Left Ventricular Dysfunction
- Received October 16, 2002.
- Revision received March 19, 2003.
- Accepted March 27, 2003.
- American College of Cardiology Foundation
- Weber K.T.,
- Brilla C.G.
- Packer M.
- SAVE Investigators,
- Pfeffer M.A.,
- Braunwald E.,
- Moye L.A.,
- et al.
- Cooper H.,
- Dries D.,
- Davis C.,
- Li Shen Y.,
- Domanski M.
- Davis J.O.
- Laragh J.H.
- ↵Brilla CG, Weber KT. Prevention of myocardial fibrosis in renovascular hypertension and hyperaldosteronism: anti-fibrotic effects of spironolactone. In: 17th International Aldosterone Conference, Evansville, IN. Meetings and Events Communications, 1991:56–7
- Cleland J.G.F.,
- Dargie H.J.,
- Hodsman G.P.,
- et al.
- Duprez D.A.,
- De Buyzere M.L.,
- Rietzschel E.R.,
- et al.
- Urata H.,
- Healy B.H.,
- Stewart R.,
- Bumpus F.M.,
- Husain A.
- Hunt S.A.,
- Baker D.W.,
- Chin M.H.,
- et al.
- RALES Investigators,
- Zannad F.,
- Alla F.,
- Dousset B.,
- Perez A.,
- Pitt B.