Author + information
- Received June 2, 2000
- Revision received January 31, 2001
- Accepted February 13, 2001
- Published online June 1, 2001.
- ↵*Reprint requests and correspondence: Dr. K. M. Yee, Cardiac Unit, Papworth Hospital, Papworth Everard, Cambridge, United Kingdom
The study was designed to comprehensively evaluate the circadian effects of aldosterone blockade on autonomic tone and QT dispersion in chronic heart failure (CHF).
Spironolactone therapy given in addition to angiotensin-converting enzyme inhibitors improved survival in CHF, but the mechanism of its benefit is uncertain. Experimental evidence suggests that aldosterone may have detrimental effects on the autonomic nervous system, especially during the morning hours.
Twenty-eight patients with New York Heart Association class II to IV CHF received spironolactone 50 mg daily and placebo for four weeks each in a double-blind crossover fashion. After each treatment phase, a full circadian assessment was undertaken of spironolactone’s autonomic effects. The assessment included monitoring heart rate, QT dispersion, continuous Holter recordings, heart rate variability (HRV) and norepinephrine kinetics.
Spironolactone significantly reduced all indices of QT dispersion. The reductions in QTcmax, QTd and QTcd were greatest at 6 am. In addition, spironolactone had favorable autonomic effects, which were limited to the morning (6–10 am), including heart rate reduction and an improvement in HRV.
Spironolactone reduced heart rate and improved HRV and QT dispersion in CHF. Its effects were particularly prominent during the morning hours.
Despite the introduction of angiotensin-converting enzyme (ACE) inhibitors, mortality rates remain high in chronic heart failure (CHF) (1–3). This may be partly attributed to the fact that neurohormonal suppression of the renin-angiotensin-aldosterone system by long-term ACE inhibition is inadequate. The suppressive effect of long-term ACE inhibitors on aldosterone is weak, variable and unsustained (4,5). The clinical significance of these residual levels of aldosterone was recently highlighted in the multicenter Randomized Aldactone Evaluation Study (RALES), where a 30% reduction in mortality was demonstrated in those patients with CHF who received the aldosterone antagonist spironolactone, compared with those receiving placebo (6).
Elevated aldosterone levels in CHF may be harmful for a number of reasons, including the recently recognized phenomenon that aldosterone may have a detrimental effect on the autonomic nervous system. First, Wang et al. found that aldosterone infusion directly reduced baroreceptor discharge from the carotid sinus in dogs (7). Second, we have recently shown that aldosterone blunts the human baroreflex response (8). Third, aldosterone may potentiate the effects of catecholamines (9). Aldosterone has been shown to block myocardial uptake of norepinephrine (NA) in vivo in an animal model (10). Furthermore, in patients with CHF, the addition of spironolactone was found to increase myocardial NA uptake by metaiodobenzylguanidine (MIBG) scanning (10). Finally, we recently observed that spironolactone also improved the time-domain measures of heart rate variability (HRV) in CHF (11).
Intriguingly, aldosterone secretion (12,13)and the autonomic nervous system (14,15)exhibit a circadian variation, with both increasing in the early morning. This early-morning peak in aldosterone and sympathetic activity occurs at the same time as cardiac events peak (16–18). Hence, the hypothesis arises that the dawn surge in aldosterone may unfavorably modulate the sympathovagal balance, which then contributes to the high incidence of cardiac arrhythmias (16), ischemia (17)and sudden deaths (18)seen particularly during these early hours. In this study, we carried out a full circadian assessment of the influence of aldosterone blockade on sympathovagal balance. In addition, we investigated the effects on cardiac arrhythmias and the circadian variation of QT dispersion indices. The effect of aldosterone on the QT interval dispersion (QTd), which is the difference between the longest and shortest QT interval measured on the 12-lead electrocardiogram (ECG), is of particular interest because it has recently been identified as a sensitive predictor of sudden cardiac death (SCD) (19–21).
Subjects and protocol
Twenty-eight patients with CHF (New York Heart Association [NYHA] class II to IV) and left ventricular ejection fraction ≤40% were recruited from the CHF clinic of our institute. All patients gave informed written consent, and the protocol was approved by the local Tayside Ethics Committee on Medical Research. The baseline demographic characteristics are listed in Table 1. All patients were on optimal therapy and had been treated with a diuretic and ACE inhibitor for at least six weeks. Left ventricular ejection fraction was measured by either echocardiography or radionuclide ventriculography. We excluded subjects whose predominant cardiac rhythm was not sinus or who had a bundle branch block or paced complexes on the ECG. All cardioactive drugs were unchanged during the study.
All patients were recruited into a double-blind, randomized crossover study. Baseline hemodynamic, biochemical and ECG measurements were obtained before each patient was randomized to receive spironolactone 50 mg daily or placebo for a one-month period for each treatment. Between each treatment, there was a washout period of 30 days. Patients were reviewed weekly for clinical evaluation of tolerability and biochemical measurements.
At the end of each treatment phase, the patients were admitted to our research unit and studied over 24 h. All patients had two intravenous 18G indwelling cannulae inserted (one in each forearm) and a 24-hr continuous Holter monitor attached to their chest. After 30 min of bed rest, baseline values of blood pressure and heart rate were determined in triplicate using a semi-automatic sphygmomanometer (Dinamap, Critikon, Tampa, Florida).
Repeated ECGs and venous blood samples (10 ml) for plasma NA were obtained at the following time points: 1500 h, 1700 h, 2300 h, 0300 h, 0600 h, 0700 h, 0800 h and 1100 h. The patients were asked to lie supine for 30 min before each time point. In addition, at 1600 h and 0700 h the patients were administered an intravenous infusion of 1-(2,5,6-H)-NA for assessment of NA kinetics.
Between the study time points, the patients were allowed to sit or move about the bedside. Excessive movements were kept to a minimum, and patients were instructed to record all their movements. Bedtime and mealtimes were standardized as follows: breakfast 0930 h, lunch 1300 h, dinner 1830 h and bedtime 2300 to 0700 h.
Continuous ECG monitoring and heart rate variability analysis
Twenty-four hour continuous ECG recordings were obtained for analysis in all subjects using a standard two-channel (four leads) Tracker 2 analog tape recorder (Reynolds, Hertford, United Kingdom) recording standard leads CM1 and CM5.
Semiautomatic analysis of arrhythmias was performed using the Pathfinder 500 Series analyzer system (Reynolds, Hertford, United Kingdom). In addition, the recordings were also visually checked, and RR intervals and QRS configuration manually edited, to ensure correct arrhythmia recognition and classification. Heart rate variability was assessed according to standardized guidelines (22). The RR interval was automatically analyzed statistically as a function of time (time domain analysis) by the Pathfinder using a commercially available and validated software program (Reynolds Medical Software version 4.63). The following time-domain indices were evaluated: standard deviation of all RR intervals, 24 h triangle index, standard deviation of 5-min mean RR intervals and the root mean square of differences of successive RR intervals.
Frequency domain (power spectral) analysis was undertaken by fast Fourier transformation. This process operated on data of 5 min segments of each hour over the 24-h recording while the subjects were at rest. Time periods where there were excessive movements were excluded. Spectral plots were used to identify the low-frequency (LF) component (0.03 to 0.14 Hz) and the high-frequency (HF) component (0.18 to 0.40 Hz). Indices are expressed in normalized units (n.u.), or the relative percentage compared to the total oscillatory power.
Recordings shorter than 18 h or with <40% of the tape suitable for analysis were discarded. Separate analyses were undertaken for the dawn period (0600 h to 1000 h) and the rest of the day (1000 h to 0600 h). These times were specified before the study began, as our hypothesis was that the dawn period might be different from the rest. All analyses were carried out by a single observer (KMY) blind to conditions. The intra-observer variability for HRV analysis was 6% to 7% for power spectral analyses and 5% to 6% for time-domain indices.
QT interval and dispersion analysis
QT interval analysis was performed on ECGs (with simultaneous 12-lead acquisition) recorded with a Hewlett-Packard 4700A electrocardiogram machine (Palo Alto, California). All QT intervals and dispersion (QTd) were analyzed blindly on a Calcomp digitizing tablet (Twyford, United Kingdom) using customized software (Medical Computing Unit, Ninewells Hospital). QT interval was taken from the onset of the QRS to the end of the T wave (i.e., return to the T/P baseline). If U waves were present, the QT interval was measured to the nadir of the curve between the T and U waves. Three consecutive cycles were usually measured for each lead. QT intervals were corrected with Bazett’s formula (QTc= QT/RR1/2).
QTd was calculated in ECGs in which ≥9 leads were measurable. QTd is defined as the difference between maximum and minimum QT intervals (QTd = QTmax− QTmin). Corrected QT dispersion (QTcd) is defined as the difference between maximum and minimum QTc (QTcd = QTcmax− QTcmin). The intra-observer variability was 3% and 8% to 9% for measurements of QT intervals and QT dispersion, respectively.
Studies were performed according to the well-established technique of Esler et al. (23). An initial loading dose of tritiated NA infusion (NEN-DuPont, Herts, United Kingdom) was given intravenously as a bolus (12.5 μCi over 2 min) followed by a constant infusion (0.7 μCi/min/m2) for 60 min. This infusion protocol has been shown to achieve constant plasma concentration of tritiated NA within 30 min (24). Venous blood samples (20 ml) were taken for determination of resting plasma NA and tritiated NA levels after 40 and 50 min infusion (two samples were taken to ensure that steady state had been achieved). Blood samples were taken in chilled lithium-heparin tubes centrifuged at 3,000 rpm at 4°C, and the plasma was separated and stored at −70°C until assayed as a batch. Two 5 ml samples of the infusate were also collected (one before and the other after completion of infusion), stored and assayed as described for the blood samples, to allow determination of the actual rate of tritiated NA infusion.
After the addition of dihydroxybenzylamine used as an internal standard, catechols were extracted through adsorption on alumina. After elution from the alumina through the addition of 0.2 mol/L HClO4, the catechols were then separated and measured by high-performance liquid chromatography (HPLC). The HPLC effluent coinciding with the NA peak was collected, and (3H)NA radioactivity levels were determined by liquid scintillation spectroscopy. The intra-assay coefficient of variability was 6.1%. Norepinephrine kinetics are calculated as follows:
All the data were analyzed using Statgraphics software package (STSC Software Publishing Group, Maryland). Repeated measures analysis of variance was carried out on the HRV and QT outcome measures. The within-subject factors were treatment and time; the between-subject factor was order of treatment. A full factorial model was fitted to test the significance of these factors. The effects of treatment on the other measures were analyzed using the paired Student ttest. A p value < 0.05 was regarded as significant.
Both spironolactone and placebo given in addition to ACE inhibitors were well tolerated. There were no significant changes in baseline hemodynamic parameters. After four weeks of treatment, spironolactone caused a small but significant rise in plasma potassium and magnesium (Table 2). Three of the patients had significant hyperkalemia (potassium >5.5 mmol/l) and elevation of plasma creatinine (>300 μmol/l), which necessitated the dose of spironolactone to be halved (25 mg). Plasma aldosterone levels measured between 6 amand 11 amrose from 90 ± 39 pg/ml to 137 ± 60 pg/ml and from 220 ± 80 pg/ml to 316 ± 108 pg/ml on the placebo and spironolactone study days, respectively. The effects of spironolactone on the various autonomic indices are displayed in Tables 2 and 3. ⇓Arrhythmic activity was unaffected.
Circadian pattern of heart rate variability
Spironolactone therapy caused a significant overall increase in the mean HF component and decrease in the LF/HF ratio throughout the 24-h period compared to placebo (Table 2). There was also a highly significant time and treatment interaction for these measures. The effects on the HRV indices were greatest from 0600 to 1000 (Table 3, Figs. 1 and 2). ⇓⇓During these hours, standard deviation of the RR intervals and the HF component were significantly increased, whereas both heart rate and the LF/HF ratio were reduced. The LF component was not significantly affected.
Circadian pattern of QT dispersion
QT interval dispersion indices were significantly higher during the day compared with the sleeping hours (Fig. 3). All measured QT indices including QTcmaxand QTd were significantly reduced by spironolactone (Table 2, Fig. 4). In addition, there were also significant time and treatment interactions for these outcome measures, with the treatment effects of spironolactone greatest during the morning hours.
There were no significant differences in the diurnal profile of plasma NA levels between the two treatments. Whole-body NA clearance and spillover did not appear to be significantly affected by spironolactone therapy compared with placebo (Table 3).
There were no significant differences in treatment effects between NYHA classes or between the NIDDM and non-NIDDM subgroups. It must however be pointed out that the study was notdesigned (and therefore was not statistically “powered”) for subgroup analyses.
Consistent with previous findings (6,10,24,25), the addition of a small dose of spironolactone (50 mg/daily) to maintenance treatment with a loop diuretic and ACE inhibitor in CHF appears to be well tolerated. Spironolactone improved not only heart rate itself but also HRV and QT dispersion. Perhaps the most intriguing aspect of our findings is that these effects appear to be maximal between 6 amand 10 amwhen aldosterone secretion is also maximal (12,13). It is unlikely that the observed heightened early morning effect of spironolactone can be attributed to the drug pharmacokinetics, as spironolactone has a very gradual and prolonged action (26). Furthermore the observed beneficial effects were observed from 6 amonwards, well beforethe patients took the study medications (8–9 am).
Aldosterone and circadian sympathovagal modulation
To date, there have been very limited data regarding the diurnal variation of the autonomic tone in CHF. Spectral analysis of 24-h HRV recordings has shown that in normal subjects, LF and HF indices exhibit a circadian pattern and reciprocal fluctuations with higher values of LF in the daytime and of HF at night, corresponding with a sympathetic predominance during the day and vagal predominance during the night (15).
In CHF, this normal circadian variation of the autonomic tone is disrupted. Both time-domain and frequency-domain parameters of HRV are reduced (14,27–29). Reduced HRV has recently been shown to be a good predictor of SCD in CHF (30). In our study, we observed a reduced but still preserved circadian pattern in the spectral components in CHF, a finding also reported by some (14)but not others (28,29).
We found that the circadian variation of the autonomic tone is further improved or preserved following spironolactone therapy. In particular, the HF spectral component was significantly increased whereas the LF/HF ratio was significantly reduced during the morning hours (6–10 am). Interestingly, heart rate was also significantly reduced during this crucial time period, a finding that we had observed in another study (11).
Vagal activity is thought to be the major contributor to the HF component (22). Whereas some studies suggest that LF, when expressed as n.u., is a quantitative marker for sympathetic modulations, others view LF as reflecting both sympathetic and vagal activity (22). Consequently, the LF/HF ratio is considered by some to mirror sympathovagal balance (22). Thus the favorable impact of spironolactone on both HF and LF/HF (but not LF) suggests that it works predominantly on the vagal limb of the autonomic tone. There is certainly a growing body of evidence that aldosterone has detrimental effects on vagal activity (7,8).
The role of aldosterone in sympathetic modulation is controversial. Although there is some evidence to suggest that aldosterone may potentiate sympathetic activity (10,11), we were unable to confirm this. The diurnal plasma NA profiles were similar with both treatments, and whole-body NA kinetics was unaffected. These findings are not necessarily a contradiction to previous observations where cardiac NA uptake as assessed by MIBG scanning was increased by spironolactone (10). The sympathetic nervous system is highly differentiated, and the effects observed in the body as a whole may not be representative of adrenergic drive specifically to the heart (31). Newton et al. (32)made similar observations with digoxin therapy in CHF when they found that digoxin reduced cardiac NA spillover but did not affect total-body NA kinetics.
Aldosterone and QT dispersion
The circadian pattern of the QTc interval and dispersion in normal humans is well established (33,34), but we believe that ours is the first study to show that QT dispersion occurs in a circadian fashion in patients with CHF. Following spironolactone treatment, both QTcmaxand QT dispersion were significantly reduced at nearly all time points (Table 3, Fig. 4). This effect was greatest in the morning hours where there was a 19% reduction in QTd compared to placebo. Even more important is that of the six patients who had a QTd > 80 ms on placebo, none of them had a QTd > 80 ms on spironolactone. QTc interval prolongation has repeatedly been shown to be an independent risk factor for SCD (35–37). Similarly, QTd > 80–85 ms has been identified as a crucial threshold in predicting SCD (37). QTd is thought to be a measure of electrical inhomogeneity, which decreases an individual threshold for ventricular arrhythmias (38). These data would suggest that spironolactone may have properties that reduce malignant arrhythmias. Although arrhythmic events were not significantly influenced in this study, the sample size and the number of potentially lethal arrhythmic events recorded were small.
There are a number of possible mechanisms in which aldosterone may contribute to prolongation and dispersion of the QT interval. As our HRV data suggest, aldosterone may have detrimental autonomic modulating properties, and the only drugs to have reduced QTd are those that modulate the autonomic tone, such as beta-blockers (39,40)and ACE inhibitors (41–43). Aldosterone-induced potassium and magnesium depletion may also contribute to dispersion of the QT intervals. Both hypokalemic and hypomagnesemic states are arrhythmogenic, associated with prolongation of the QT intervals and increased likelihood of developing ventricular arrhythmias (44–46). Conversely, magnesium replacement shortens the QT interval and has been shown to be effective in terminating torsade de pointes (47). The effects of magnesium on QT dispersion are unknown. Potassium is one of the main determinants of the QT interval, as it is responsible for the outward repolarization currents. Hypokalemia results in slower repolarization and prolonged QT intervals (48). Recently, Choy et al. (49)observed that intravenous potassium infusion not only normalized QT prolongation but also reduces QT dispersion in CHF.
First, the technical difficulty of obtaining a precise and reproducible determination of the QT interval is well recognized. Care has been taken to use consistent criteria to define the end of the T wave and to exclude U waves. In any case, this was not a major issue for this study as we were looking at treatment-induced intraindividualchanges in QTd. Second, only a minority of subjects in this trial were on beta-blockers. This reflects current practice in the United Kingdom. Although beta-blockers could alter the autonomic tone, many patients with CHF cannot tolerate them. In any case, the main effect of beta-blocker therapy is on the sympathetic system, whereas the main effect of spironolactone appears to be on the parasympathetic tone. Hence, the use of bothtreatments may actually be synergisticin reducing mortality, as suggested by the RALES subgroup analysis (6).
Aldosterone blockade has a number of beneficial effects in CHF. It reduces heart rate and improves the parasympathetic component of HRV and QT dispersion. Its effects appear to be greatest during the morning hours, coinciding with the circadian pattern of cardiac events.
☆ Dr. K.M. Yee is supported by a grant from the Scottish Office and Home Health Department, United Kingdom.
- angiotensin-converting enzyme
- chronic heart failure
- corrected QT dispersion
- high frequency
- high-performance liquid chromatography
- heart rate variability
- low frequency
- New York Heart Association
- QT interval dispersion
- Randomized Aldactone Evaluation Study
- sudden cardiac death
- Received June 2, 2000.
- Revision received January 31, 2001.
- Accepted February 13, 2001.
- American College of Cardiology
- Staesson J.,
- Lingen P.,
- Fagard R.,
- Verschueren L.J.,
- Amery A.
- MacFadyen R.J.,
- Barr C.S.,
- Struthers A.D.
- Armbruster H.,
- Vetter W.,
- Beckerhoff R.,
- Nussberger J.,
- Vetter H.,
- Seigenthaler W.
- Davidson N.C.,
- Coutie W.J.,
- Webb D.J.,
- Struthers A.D.
- Adamopoulos S.,
- Ponikowsky P.,
- Cerquetani E.,
- et al.
- Furlan R.,
- Guzzetti S.,
- Crivellaro W.,
- et al.
- Moser D.K.,
- Stevenson W.G.,
- Woo M.A.,
- Stevenson L.W.
- Darbar D.,
- Luck J.,
- Davidson N.,
- et al.
- Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology
- Dahlstrom U.,
- Karlsson E.
- Nolan J.,
- Batin P.D.,
- Andrews R.,
- et al.
- Haskings G.J.,
- Esler M.D.,
- Jennings G.L.,
- et al.
- Newton G.E.,
- Tong J.H.,
- Schofield A.M.,
- Baines A.D.,
- Floras J.S.,
- Parker J.D.
- Schwartz P.J.,
- Wolf S.
- Algra A.,
- Tijsren J.G.,
- Roelandt J.R.,
- Paul J.,
- Lubsen J.
- Naas A.,
- Davidson N.C.,
- Thompson C.,
- et al.
- Pye M.P.,
- Cobbe S.M.
- Priori S.G.,
- Napolitano C.,
- Diehl L.,
- Schwartz P.J.
- Day C.P.,
- McComb J.M.,
- Matthews J.,
- Campbell R.W.F.
- Fisch C.
- Nabauer M.,
- Beuckelmann D.J.,
- Erdmann E.
- Choy A.M.,
- Lang C.C.,
- Chomsky D.M.,
- Rayos G.H.,
- Wilson J.R.,
- Roden D.M.