Author + information
- Received October 1, 2013
- Accepted November 18, 2013
- Published online June 24, 2014.
- Kristen K. Patton, MD∗,
- Anne S. Hellkamp, MS†,
- Kerry L. Lee, PhD†,
- Daniel B. Mark, MD, MPH†,
- George W. Johnson, BSEE‡,
- Jill Anderson, RN‡,
- Gust H. Bardy, MD‡,
- Jeanne E. Poole, MD∗∗ (, )
- SCD-HeFT Investigators
- ∗Department of Medicine, University of Washington, Seattle, Washington
- †Duke Clinical Research Institute, Durham, North Carolina
- ‡Seattle Institute for Cardiac Research, Seattle, Washington
- ↵∗Reprint requests and correspondence:
Dr. Jeanne E. Poole, University of Washington, Box 356422, 1959 NE Pacific Street, Seattle, Washington 98115.
Objectives This study sought to determine whether circadian patterns in ventricular arrhythmias (VAs) occur in a current primary prevention defibrillator (implantable cardioverter-defibrillator [ICD]) population.
Background Cardiovascular events, including VAs, demonstrate biorhythmic periodicity.
Methods We tested for deviation from the previously described occurrences of a morning peak, early morning nadir, and peak on Mondays in ICD therapies using generalized estimating equations and Student t tests. All hypothesis tests were performed in the entire cohort of patients with VAs as well as pre-specified subgroups.
Results Of 811 patients with an ICD, 186 subjects experienced 714 ICD therapy episodes for life-threatening VA. There was no morning (6 am to 12 pm) peak in therapies for the entire cohort or any subgroups. The overall cohort and several subgroups had a typical early morning (12 am to 6 am) nadir in therapies, with significantly less than 25% of therapies occurring during this 6-h block (all p < 0.05). A significant peak in therapies on Mondays occurred only in patients not on beta-blocker therapy (22% of events for the week, p = 0.029).
Conclusions In the SCD-HeFT (Sudden Cardiac Death in Heart Failure Trial) population, the distribution of life-threatening VA failed to show a typical early morning peak or increased VA events on Mondays. A typical early morning nadir was seen in the entire cohort. An increased rate of events on Mondays was found in the subgroup of subjects not on beta-blocker therapy. These findings may indicate suppression of the neurohormonal triggers for VA by current heart failure therapy, particularly the use of beta-blockers in heart failure.
The timing of acute myocardial infarction, sudden cardiac death, and ventricular arrhythmia (VA) shows circadian and septadian periodicity, with a prominent peak in the morning, a sleep time nadir, and increased event rates on Mondays (1–8). Sudden death and implantable cardioverter-defibrillator (ICD) shocks have been reported to occur with higher frequency in the early morning hours in both ischemic and nonischemic cardiomyopathy (9,10). These findings suggest that some triggering mechanisms for arrhythmic events are influenced by normal alterations in biorhythms. Biorhythmic fluctuations in platelet aggregation, blood pressure, heart rate, cortisol levels, catecholamine levels, vascular tone, and ischemia are believed to underlie the circadian patterning of myocardial infarction, ICD shocks, and sudden cardiac death (11–13).
Clinical management of heart failure, indications for ICD placement, and our ability to interrogate ICD electrograms have evolved significantly since earlier studies on timing and ICD shocks were performed. Recent analyses of temporal patterns of VA in subjects with hypertrophic cardiomyopathy, Brugada syndrome, and early repolarization syndrome showed substantial differences in timing of arrhythmia occurrence that may indicate differences in arrhythmia triggering mechanisms in varying pathologies (14,15). This may have therapeutic implications, because a vagally mediated arrhythmogenic process would be targeted differently than a sympathetically driven process. Although current conceptual models of chronobiologic rhythms and their impact on ventricular arrhythmogenesis have highlighted rhythmic variability in sympathetic nervous system input as a major triggering mechanism, the data supporting this are conflicting (16). Notably, adrenergic blockade has been shown in small studies to abolish the typical early morning VA peak (17) and in other studies to leave typical circadian patterns unaffected (18). In an analogous area of cardiovascular disease, evidence has suggested that hypertension chronotherapy—reestablishing normal circadian patterns in blood pressure by systematic scheduling of medication dosing—is associated with a significant reduction in cardiovascular events (19). Using the SCD-HeFT (Sudden Cardiac Death in Heart Failure Trial) primary prevention ICD population, we sought to determine whether the timing of ICD therapies for life-threatening VA manifests circadian and septadian variation and whether that pattern is importantly modified by specific patient characteristics or heart failure therapies.
Details of SCD-HeFT have been previously reported (20). In brief, from September 16, 1997, to July 18, 2001, 2,521 subjects were randomly assigned to receive placebo (847 subjects), amiodarone (845 subjects), or a single-chamber ICD programmed to shock-only mode (Medtronic 7223; 829 subjects, of which 811 received an ICD). Patients enrolled were at least 18 years of age and had New York Heart Association (NYHA) functional class II or III, chronic stable heart failure due to ischemic or nonischemic causes, and a left ventricular ejection fraction (EF) of no more than 35%. All patients signed an informed consent form, and the study was approved by the human subjects committee of each participating institution.
The ICD was programmed to have a single zone of therapy, and an episode of tachycardia was defined as at least 18 of 24 sequential beats at a detection rate of 187.5 beats/min or more. Antitachycardia pacing was not specified in the protocol but was programmed in a small number of subjects. In contrast to prior reports (21), this analysis included subjects who received either shocks or antitachycardia pacing therapy for events classified by the electrogram committee as an appropriate therapy for ventricular tachycardia or ventricular fibrillation. Subjects were followed up every 3 months with alternating clinic visits and telephone calls. Data from the ICD memory log were regularly downloaded at these visits. The time and date of each episode was collected. For subjects with arrhythmic storm, events separated by <1 min were counted as the same event.
Baseline characteristics are summarized as median (25th, 75th percentile) for continuous variables and percent (number) for categorical variables. The timing of episode occurrence was examined in 2 ways: 1) on a per-episode basis, considering each episode as a separate observation and giving each episode equal weight in the analysis; and 2) on a per-patient basis, considering each patient as a separate observation and giving each patient equal weight in the analysis. The latter approach circumvents the issue of the statistical results being dominated by a few patients with a large number of events. On the basis of time of occurrence, each ICD therapy was assigned to one of 4 time intervals: 12 am to 6 am, 6 am to 12 pm, 12 pm to 6 pm, and 6 pm to 12 am. For each 6-h interval, the proportion of all episodes occurring during that interval was calculated for the per-episode approach. For the per-patient approach, we calculated a weighted frequency for each patient equal to the proportion of episodes weighted by the inverse of the total episodes (for example, a patient with 1 episode in a given interval out of 4 total episodes would have a weighted frequency of 0.25 for that interval).
We tested 3 hypotheses: 1) morning peak: that the morning interval of 6 am to 12 pm, representing 0.25 of the 24-h period, would have more than 0.25 of the episodes; 2) early morning nadir: that the early morning interval of 12 am to 6 am, again representing 0.25 of the 24-hour period, would have less than 0.25 of the episodes; and 3) Monday peak: that Monday, representing 0.143 (one-seventh) of the week, would have more than 0.143 of the episodes. Each of these hypotheses was tested for both the per-episode approach, using generalized estimating equations to account for intrapatient correlation between episodes, and the per-patient approach, using a Student t test to compare the sample mean weighted frequency with the stated value. However, because the results of the 2 sets of analyses were quite similar, only the results of the per-patient analysis are shown. All hypothesis tests were performed in the entire cohort of patients with VA as well as subgroups defined by heart failure etiology, NYHA functional class, age (≤50 years and >50 years), sex, EF (≤25% and >25%), and baseline beta-blocker use. All hypothesis tests were one sided, and p < 0.05 was considered significant. All analyses were conducted with SAS version 9.2 (SAS Institute, Cary, North Carolina).
A total of 714 ICD therapies for life-threatening VA occurred in 186 subjects. Baseline characteristics are presented in Table 1. Among the patients experiencing appropriate ICD therapy for VA, the median age was 63 years and the majority were male (79%). At enrollment, 56% of the subjects were treated with a beta-blocker and 82% with an angiotensin-converting enzyme inhibitor. Ischemic cardiomyopathy was present in 51%, and 64% had NYHA functional class II chronic heart failure.
The distribution of ICD therapies over the 24-h period, in 3-h increments, is shown in Figure 1. No clear pattern is evident. The proportion of ICD therapies occurring in the typical 6 am to 12 pm interval was not significantly >0.25, either overall or for any subgroup (all p > 0.2) (Table 2). In fact, for almost all subgroups, both the proportion of episodes and the mean weighted frequency during these morning hours were <0.25.
The proportion of ICD therapies occurring in the typical nadir of the 12 am to 6 am interval is shown in Table 3 and Figure 2. In the overall cohort and among patients who were ischemic, NYHA functional class II, older, or not on a beta-blocker, the proportion of events occurring during this interval was significantly less than 0.25 (all p < 0.05), with a weighted frequency of 0.21 for all patients and a range of 0.17 to 0.19 for the 4 subgroups. The proportions are closer to 0.25 than the weighted frequencies in most subgroups, indicating that the proportions are influenced by a subset of patients with a larger number of episodes during the interval.
The distribution of ICD therapies over the 7-day week for all patients is shown in Figure 3. In general, the lowest proportion occurred on Sundays and Thursdays and highest proportion on Tuesdays or Wednesdays. The proportion of therapies on Mondays was not significantly greater than one-seventh for the entire cohort or any subgroup except those patients not on beta-blocker therapy (Table 4). Among those patients, the mean weighted frequency was 0.22, indicating that nearly one-fourth of their episodes occurred on Mondays.
Analysis of the timing of ICD therapy for life-threatening VA in the SCD-HeFT cohort revealed a typical early morning nadir of VA but no early morning peak or Monday predominance. Additionally, subgroup analysis showed the early morning nadir in patients who were ischemic, NYHA functional class II, younger than 50 years of age, and not on beta-blocker therapy. Conversely, we did not see a typical morning peak of arrhythmia in any subgroup stratified by sex, heart failure etiology, heart failure class, or beta-blocker use. Septadian (weekly) variation was seen only in the subgroup not on beta-blocker therapy.
Multiple prior studies have shown a circadian pattern of ICD firing, with a prominent morning peak, often a smaller secondary late afternoon peak, and an early morning nadir (4,5,18,22,23). Tofler et al. (8), in the first large study of the timing of ICD shocks, analyzed data from 483 patients with a Ventak PRx ICD (Guidant, St. Paul, Minnesota) implanted between 1990 and 1993. A circadian pattern with a peak between 9 am and 12 pm and a nadir between 3 am and 6 am was seen in both groups (p < 0.001). The results did not differ within subgroups of age, sex, EF, NYHA functional class, or heart failure etiology. The group with EF <20% or NYHA functional class III chronic heart failure was notable for an attenuated nighttime decrease in arrhythmia frequency and a blunted morning peak.
Mallavarapu et al. (24) analyzed the timing of VA episodes in 390 patients with healed myocardial infarction and assessed whether the circadian pattern was influenced by sex, age, left ventricular function, or ventricular tachycardia cycle length <300 ms or >300 ms. A total of 15,731 device detections occurred, of which 2,692 electrograms were available and consistent with VA (2,488 episodes treated, 208 episodes with therapy aborted). Circadian variation in the occurrence of VA was evident, with the highest rate between 10 am and 11 am and the lowest rate between 2 am and 3 am. This pattern persisted regardless of subgroup.
In 2 analyses, Peters et al. (7,25) investigated the septadian distribution of ICD shocks and the interaction of septadian and circadian rhythms in 683 patients with a Ventak PRx ICD. The circadian pattern of ICD shock for episodes with a cycle length <280 ms showed a broad peak between 9 am and 6 pm and a long nadir between 9 pm and 6 am. The daily event timing distribution was strikingly similar for each day of the week, despite twice as many discharges on Mondays as compared with Saturdays. The circadian variation was independent of age, sex, EF, NYHA functional class, etiology of heart failure, and use of antiarrhythmic drugs but was not seen in patients on beta-blocker therapy.
Prior large and small studies predominantly included subjects with coronary artery disease. Englund et al. (26) reported an analysis of a more mixed population of ischemic and nonischemic chronic heart failure. Between 1993 and 1994, 820 patients received an ICD, and 310 had an episode of VA. Both groups had similar circadian patterns of events, with a pronounced morning and a less pronounced afternoon peak. Analogous data showing similar circadian distributions of ventricular tachycardia episodes in ischemic and nonischemic cardiomyopathy were also reported by Grimm et al. (10).
Several prior studies have investigated the relationship between NYHA functional class or EF and circadian pattern of VA (4,5,8,22,24). Three studies did not show a difference in the timing of VA in subgroup analysis that included EF or NYHA functional class (5,22,24), whereas Behrens et al. (4) found no typical circadian pattern in class III subjects and Tofler et al. (8) described a blunted morning peak and attenuation of the nighttime nadir in class IV subjects and those with EF <20%. In our study, an early nadir was found in the overall cohort but on subgroup analysis only in subjects with class II heart failure, ischemic cardiomyopathy, age younger than 50 years, and beta-blocker therapy. It is curious that both the older and sicker patients had an attenuated circadian pattern, as well as the nonischemic patients and those treated with adrenergic blockade. A critical difference in our population compared with prior studies is the mean EF of 20%; the mean EF in all prior studies but one was >30%. This difference may play a fundamental role in the dissimilarity of specific triggers for episodes of VA or may reflect other pathways of autonomic dysregulation, such as severe heart failure.
Our study population differs from prior reports in ways important to the underlying pathophysiological processes involved in arrhythmia generation. SCD-HeFT was a prospective randomized clinical trial designed to evaluate ICD therapy in the primary prevention of sudden cardiac death. Prior studies of biorhythmic effects on VA occurrence have examined predominantly secondary prevention patients, with mostly ischemic heart disease and higher EFs. In our study, stored electrograms for each episode were available and adjudicated by the SCD-HeFT Electrogram Committee, data that were similarly available in only one prior study. Because the majority of the foundational research in the area of circadian variation in ICD shocks was performed more than a decade ago, major changes in heart failure therapy have occurred. More than 90% of the SCD-HeFT ICD subjects were treated with angiotensin-converting enzyme inhibitors or angiotensin receptor blockade, the majority was treated with a beta-blocker, and 38% were treated with a statin. Given the myriad effects of these agents, which include anti-ischemic, anti-inflammatory, and vasodilatory properties, any of these agents, or a synergistic combination, may account for the attenuation in usual circadian patterns seen in this study.
The influence of circadian rhythms on the occurrence of cardiovascular events is often explained by a morning surge of sympathetic nervous system activity resulting in physiological changes that produce myocardial ischemia (11,13). This has also led to the hypothesis that a subset of episodes of VAs and sudden cardiac death corresponding to these circadian surges may be aborted by treating patients with beta-blockers. Notably, the significant result in the full cohort is driven by the pattern in the patients with ischemic cardiomyopathy. Our findings support the ischemia hypothesis in that we found an attenuation of circadian and septadian rhythms in subjects on beta-blocker therapy; in the SCD-HeFT ICD arm, 57% of the subjects who received ICD shocks for VAs and 70% of all patients who received an ICD were treated with beta-blockade. It is tempting to speculate that the increased rate of neurohormonal suppression of adrenergic triggers of VAs modulated the previously observed circadian and septadian patterns. Additional evidence for this is that the SCD-HeFT subgroup of subjects not on beta-blocker therapy showed more traditional variation in timing of arrhythmia, with the early morning nadir, and increased prevalence on Mondays.
Limitations of this study include the post-hoc nature of the analysis. Additionally important to note is the lack of data regarding work schedules or sleep-wake schedules, timing of medication ingestion, or consumption of alcohol, caffeine, and other sympathomimetic substances.
In the SCD-HeFT primary prevention population, the time distribution of ICD therapies for life-threatening VA shows an early morning nadir in the occurrence of VA but no overall typical morning peak or increase in events on Mondays. Subgroup analysis showed a more prominent typical early morning nadir in ischemic patients, those with NYHA functional class II heart failure, and younger subjects. Subjects not on beta-blocker therapy also had an increased event rate on Mondays. These findings may indicate suppression of neurohormonal triggers with current heart failure therapy.
Dr. Patton has served as a site investigator in a Cameron Health clinical trial. Dr. Lee has received research grants and consulting fees from Medtronic. Dr. Mark has received consulting fees and research grants from Eli Lilly and Company, Gilead Sciences, Bristol-Myers Squibb, AGA Medical Corporation, AstraZeneca, and Medtronic. Ms. Anderson has received consulting fees from Boston Scientific. Dr. Bardy has received research grants from the National Heart, Lung, and Blood Institute and St. Jude Medical; and holds equity and intellectual property rights with Cameron Health. Dr. Poole has received lecture fees from Medtronic, Boston Scientific/Guidant, Biotronik, and St. Jude Medical; has received consulting fees from Physio Control; holds equity in Cameron Health; and serves on the advisory board for Boston Scientific. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- ejection fraction
- implantable cardioverter-defibrillator
- New York Heart Association
- ventricular arrhythmia
- Received October 1, 2013.
- Accepted November 18, 2013.
- American College of Cardiology Foundation
- Muller J.,
- Ludmer P.,
- Willich S.,
- et al.
- Lampert R.,
- Rosenfeld L.,
- Batsford W.,
- Lee F.,
- McPherson C.
- Arntz H.-R.,
- Willich S.N.,
- Schreiber C.,
- Bruggemann T.,
- Stern R.,
- Schultheiss H.-P.
- Tofler G.H.,
- Gebara O.C.E.,
- Mittleman M.A.,
- et al.
- Muller J.E.,
- Ludmer P.L.,
- Willich S.N.,
- et al.
- Muller J.,
- Tofler G.,
- Stone P.
- Willich S.,
- Maclure M.,
- Mittleman M.,
- Arntz H.,
- Muller J.
- Bardy G.,
- Lee K.,
- Mark D.,
- et al.
- D'Avila A.,
- Wellens F.,
- Andries E.,
- Brugada P.
- Wood M.A.,
- Simpson P.M.,
- London W.B.,
- et al.
- Peters R.W.,
- McQuillan S.,
- Resnick S.K.,
- Gold M.R.
- Englund A.,
- Behrens S.,
- Wegscheider K.,
- Rowland E.