Increased Monday Incidence of Life-Threatening Ventricular Arrhythmias
Experience With a Third-Generation Implantable Defibrillator
Background Examination of the timing of cardiac events provides important pathophysiological information. Previous studies have shown that the onset of acute myocardial infarction occurs most frequently on Monday. The septadian (day of the week) pattern of occurrence of sudden cardiac death and lethal ventricular arrhythmias has not been examined previously.
Methods and Results We examined the septadian distribution of life-threatening (cycle length <280 ms) ventricular arrhythmias in 683 consecutive patients receiving a Ventak PRx, a third-generation implantable defibrillator with an event recorder. There was a prominent Monday peak, with a midweek decline and a secondary peak later in the week. A marked trough is apparent on both weekend days. The observed pattern was independent of age, sex, ejection fraction, NYHA functional class, type of heart disease, and the use of antiarrhythmic drugs but was not observed in patients receiving β-blockers.
Conclusions Potentially lethal arrhythmias are not random events but occur in a daily pattern suggesting a relationship to the beginning and end of the work week. The absence of a Monday peak in patients receiving β-blockers suggests that the pattern may be influenced by β-blockers. This information may be useful in devising strategies to prevent sudden cardiac death.
Analysis of the chronology of acute cardiovascular events provides important pathophysiological information and may ultimately lead to an improved preventive strategy. Circadian periodicity has been demonstrated in the occurrence of potentially lethal ventricular arrhythmias and sudden cardiac death.1 2 3 However, one of the major limitations of these studies has been the inability to delineate event chronology accurately. This problem has been especially apparent in the setting of unwitnessed sudden cardiac death. Some of the third-generation implantable cardioverter/defibrillators (ICDs) obviate this difficulty by providing the date and time of device activation as well as some confirmatory information about the nature of the arrhythmia.
Several recent studies have suggested that acute myocardial infarction has its onset most frequently on Mondays.4 5 In addition, an excess in cardiac mortality has been reported to occur on Mondays, although the mechanism of this finding is unclear.6 In the present study, we used the clinical database from a third-generation ICD (Ventak PRx) to assess the septadian (day of the week) frequency of rapid life-threatening ventricular arrhythmias.
The subjects for this study were obtained from the database of CPI Guidant and consist of 683 consecutive patients receiving a Ventak PRx (model 1700 or 1705) between December 4, 1990, and February 8, 1994, in whom a spontaneous tachyarrhythmia at a cycle length <280 ms was recorded. The Ventak PRx is a multiprogrammable ICD that is capable of high- and low-energy shocks and antitachycardia and antibradycardia pacing. At a cycle length of <273 ms, however, only high-energy shocks can be delivered. Incorporated into the device is an event recorder that stores the date, time, and cycle length of each tachycardic event up to a total of 128 therapy attempts, after which more recent attempts replace the older ones.
After device implantation, patients were followed at intervals of every 2 to 3 months. At each follow-up visit, the internal clock time was validated and corrected if necessary, and the event history (tachycardia detections) was transferred to a floppy disk, which was forwarded to the sponsor for inclusion in the data bank. Specifically excluded for ICD implantation were patients with a life expectancy of less than 6 months because of other medical conditions, patients who were thought to be unable to receive regular follow-up at approved investigational sites, patients who required a permanent pacemaker in addition to the ICD, and patients whose arrhythmias were thought to be due to reversible causes (eg, digitalis toxicity).
The Ventak PRx detection algorithm is “activated” when the heart rate exceeds a preset (but programmable) rate but “reconfirms” this rate before delivering therapy. Enhanced detection features are available to help prevent inappropriate shocks for sinus tachycardia and atrial fibrillation with a rapid ventricular response. These features include “onset” criteria to detect a sudden decrease in cycle length rather than a more gradual one that might be expected in sinus tachycardia or atrial fibrillation and “stability enhancement criteria” to exclude rhythms with marked cycle length variability, such as atrial fibrillation or multifocal atrial tachycardia. The number of beats required for device activation is programmable and ranges from 8 to 250 cycles (the nominal setting is 10 cycles). By design, only tachycardias with cycle lengths <280 ms (a rate of >214 bpm) that were “reconfirmed” were included in the present analysis. This cycle length was chosen prospectively on the basis of previous studies of inducible arrhythmias in cardiac arrest survivors7 8 9 10 and was aimed at including patients with potentially life-threatening events. Patients were considered to be taking antiarrhythmic drugs if they were receiving either a class I, class II, or class III drug. For purposes of analysis, we considered sotalol to be a class II drug because of its β-blocking properties. Patients receiving a drug with class II properties were grouped separately, even if they were also receiving a class I or class III drug.
Data were grouped according to the day of the week on which therapy was delivered, with each day starting at midnight and extending for a 24-hour period. Because the marked variability between patients in the number of episodes might bias the relationship between arrhythmic episodes and days of the week (eg, one patient might have 200 episodes in a single 24-hour period), statistical analysis was also performed with adjustment for the number of episodes by use of patient-episodes per day (the number of patients having a device activation in a given day). With the response variable of patient-episodes per day of the week, χ2 tests were performed to determine whether the distribution was uniform across the 7 days of the week. Next, a χ2 test for heterogeneity was performed to test whether the observed pattern was affected by eight possible effect modifiers: use of class I or class III antiarrhythmic drugs; use of class II antiarrhythmic drugs (β-blockers); sex; age group (<50, 50 to 65, >65 years old); NYHA functional classification; ejection fraction (<20%, 20% to 29%, 30% to 39%, ≥40%); primary cardiac disease process (arteriosclerotic, idiopathic dilated cardiomyopathy, other); and the timing of arrhythmia detections (the daily pattern in the first 30 days after implantation was compared with the remainder of the follow-up period to exclude the possibility that the perioperative period, including the predischarge testing, was responsible for the Monday peak). Because some patients changed antiarrhythmic drug status between follow-up periods, use of antiarrhythmic drugs was made a time-dependent variable that changed according to whether or not a patient was receiving the drug at the time of an arrhythmia detection. Thus, arrhythmia detections that occurred during a time period when a patient was receiving an antiarrhythmic drug were considered to have occurred on that drug, whereas those that occurred when a patient was not receiving a drug were counted as “off drug.” Last, a “fixed-effects” model ANOVA was performed to look for statistically significant differences between days by each level of stratified variable. A value of P<.05 was considered statistically significant.
The demographics of the study population were recorded at the time of implant and are shown in the Table⇓. Not surprisingly, the group is predominantly male and in late middle age, with a high incidence of organic heart disease and diminished left ventricular function. Approximately one third of patients (n=241) were receiving antiarrhythmic drugs; 57 were receiving class I drugs, 123 were receiving class II drugs, 95 were receiving class III drugs, and 62 were receiving a combination of classes I and III. A total of 5270 rapid (cycle length <280 ms) arrhythmic episodes were detected among the 683 patients.
Figs 1⇓ and 2⇓ show the daily distribution of total device therapy deliveries and therapy deliveries arranged according to patient-episode-days (the number of patients experiencing a device activation on a given day), respectively. The distribution is significantly nonuniform (P<.001) by χ2 analysis. Specifically, there is a prominent Monday peak (21% of episodes), with a midweek decline and a secondary peak later in the week. A marked trough is apparent on both days of the weekend; there were almost twice as many events on Monday as there were on either Saturday or Sunday.
The χ2 test for independence for the response variable (episode on a given day of the week) by the eight possible effect modifiers revealed that there was no demonstrated association between the daily distribution of events and any of the tested variables (P>.05) except for the use of class II drugs (β-blockers). All of the subgroups displayed a prominent event peak on Monday except for the patients receiving class II drugs. In this relatively small subgroup (n=123), only 9.4% of events occurred on Monday compared with 21.2% in the remainder of the population (P<.001). The distribution of therapy deliveries in patients receiving class II drugs arranged according to patient-episode-days is shown in Fig 3⇓. In contrast, in the patients receiving class I or III antiarrhythmic drugs, Monday accounted for 24.7% of total events.
In the present study, analysis of the frequency distribution of arrhythmia detections in a large cohort of patients with life-threatening ventricular arrhythmias by means of the Ventak PRx, a third-generation ICD, revealed a prominent Monday peak, a relative midweek trough with a secondary peak later in the week, and a marked weekend nadir. The observed pattern has proved to be independent of a variety of possible effect modifiers in our analysis, including age (P=.81), sex (P=.77), cause of cardiac disease (P=.95), ejection fraction (P=.54), congestive heart failure (P=.54), and the use and type of antiarrhythmic drugs (P=.34). This pattern implies a relationship to the work week and is consistent with previous studies suggesting that stress may play an important role in the genesis of acute cardiovascular events.11 12 It is also interesting to note that the observed Monday peak was not present in patients receiving β-blockers, since this class of drugs has been demonstrated to affect circadian patterns of acute cardiac events.3 This finding implies that the trigger factors of septadian rhythms may be similar to those of circadian rhythms and that the Monday peak in ventricular arrhythmias may be related to surges in catecholamine levels. We urge caution in drawing definitive conclusions about the relationship between β-blockers and septadian rhythms, however, because the number of patients receiving these drugs in our study was relatively small.
Septadian patterns of the onset of acute myocardial infarction have been described previously. Peters and associates13 found a Monday peak in the onset of acute myocardial infarction in the very large database of the Cardiac Arrhythmia Suppression Trial (CAST). In a group of patients with acute myocardial infarction in Berlin, Germany, Willich and coworkers4 found that a Monday peak occurred only in the subgroup who were currently employed. It is tempting to speculate that an increase in catecholamine levels triggered by the stress of returning to work and perhaps by anticipating the end of the work week may affect the stability of an atherosclerotic plaque, helping to trigger acute myocardial infarction. The pathogenesis of life-threatening ventricular arrhythmias is less clear. We are not aware of previous reports of septadian periodicity of ventricular arrhythmias.
Although information regarding employment status was not available for our population, the advanced age as well as the presence of severe symptomatic cardiac disease in most of the patients and the morbidity and activity restriction often associated with ICD implantation suggest that the majority were probably unemployed. Thus, it is difficult to explain the apparent relationship between the work week and arrhythmias.
It is also interesting to speculate about our findings regarding ejection fraction and congestive heart failure. A circadian distribution of ventricular ectopy has been described,14 but this pattern is apparently absent in patients with low ejection fractions,15 presumably because of the persistently elevated sympathetic tone associated with heart failure. In contrast, we found that ejection fraction or the presence of clinical congestive heart failure did not significantly affect the septadian pattern of device activations.
Limitations of the Study
Our data must be interpreted in light of certain methodological limitations. We cannot be certain that all of the arrhythmias analyzed in this study were ventricular in origin. It has been reported that up to 20% of device activations in third-generation ICDs are “spurious,” ie, due to factors other than ventricular tachycardia or fibrillation, most commonly atrial fibrillation.16 17 We attempted to deal with this problem by excluding activations due to arrhythmias with cycle lengths >280 ms (214 bpm). We think it likely, in light of the nature of our study population (sustained ventricular arrhythmias and a high incidence of severe left ventricular dysfunction), that the vast majority of detections at this rapid rate were due to ventricular arrhythmias. Our rate cutoff was selected on the basis of previous studies of patients with sustained ventricular tachycardia and/or aborted sudden cardiac death7 8 9 10 and was designed to exclude patients with arrhythmias that were hemodynamically well tolerated. In so doing, we included a total of 37 episodes (0.7% of the total number of episodes analyzed) in which antitachycardia pacing was initially delivered rather than a high-energy shock. We think it unlikely that these episodes affected our results. Another limitation is that not all of the arrhythmias detected would necessarily have been sustained. Thus, it is probable that some of the episodes would have self-terminated in the absence of an ICD shock. However, since we consider all ventricular arrhythmias at this rapid rate to be potentially life-threatening, we chose to include all activations that fulfilled our study criteria. Third, we have limited information about the patients in our study and the situations associated with arrhythmia detections. In particular, we do not have data regarding employment. Additional information would be invaluable in helping to delineate the pathophysiology of the factors that trigger ventricular arrhythmias. This information, we hope, will be forthcoming in future studies, probably involving a smaller number of subjects but designed to collect data in considerably greater detail. Fourth, the number of events occurring in patients receiving β-blockers is relatively small, so that our findings in this area will require validation by larger, prospectively designed trials. Finally, our data may not be applicable to all patients with ventricular arrhythmias but only to a selected subset who were considered appropriate to receive the Ventak PRx ICD.
Using the large database from the Ventak PRx, a third-generation ICD with memory capabilities, we found a prominent Monday peak in rapid (cycle length <280 ms) ventricular arrhythmias, with a secondary peak later in the week. In contrast, the number of arrhythmias detected on weekend days was less than half of that seen on Mondays. This pattern was observed in all patient subgroups except patients receiving β-blockers. This information has important pathophysiological and therapeutic implications. Further studies will be needed to provide prospective validation of our findings and to help to elucidate the mechanisms involved.
- Received November 16, 1995.
- Revision received March 13, 1996.
- Accepted March 26, 1996.
- Copyright © 1996 by American Heart Association
Muller JE, Ludmer PL, Willich SN, Tofler GH, Aylmer G, Klangos I, Stone PH. Circadian variation in the frequency of sudden cardiac death. Circulation. 1987;75:131-138.
Willich SN, Lowel H, Lewis M, Hormann A, Arntz H-R, Keil U. Weekly variation of acute myocardial infarction: increased Monday risk in the working population. Circulation. 1994;90:87-93.
Gnecchi-Ruscone T, Piccaluga P, Guzzetti S, Contini M, Montano N, Nicolis E. Morning and Monday: critical periods for the onset of acute myocardial infarction: the GISSI 2 Study experience. Eur Heart J. 1994;15:882-887.
Peters RW, Brooks MM, Zoble RG, Liebson PR, Seals AA. Chronobiology of acute myocardial infarction (CAST experience). Am J Cardiol. In press. Identification of a secondary peak in myocardial infarction onset 11 to 12 hours after awakening: the Cardiac Arrhythmia Suppression Trial (CAST) experience. J Am Coll Cardiol. 1993;22:998-1003.