- Email: [email protected]
Findings from long-term electrocardiographic monitoring of patients with variant angina in a coronary care unit
Findingsfrom Long-TermElectrocardiographic Monitoring of Patientswith VariantAngina in a CoronaryCare Unit CLARA CARPEGGIANI, MD, CLAUDIO MICHELASSI, ANDREA BIAGINI, MD, ROBERTO TESTA, MD, MARIA GIOVANNA MAZZEI, MD, MICHELE EMDIN, MD, and ANTONIO L’ABBATE, MD
cadian rhythm for HR was observed in 81 of the 70 days (87 % ), consistently in 7 patients; the nadir occurred at 2.4 f 1.5 AM; simultaneous cycling in HR and transient ischemia was found on 32 days. The intrapatient difference between the peak and the nadir of the ischemic and the HR function was, on average, 2.6 f 3.3 hours. Thus, a circadian rhythm of ischemic episodes was present in all patients, although it was not consistently present; simultaneous occurrence of circadian variation in ischemic episodes and HR was observed only in 60% of the days with a sufficiently high number of attacks and when this occurred, a significant phase shift was observed; occasional loss of HR cycling was observed in some patients, without an apparent cause. (Am J Cardiol 1987;60:36-39)
Eleven patients with frequent episodes of variant angina underwent 24-hour electrocardiographic monitoring in a coronary care unit for a total of 70 days to assess circadian variation in ischemic episodes and its correlation with circadian heart rate (HR) rhythm. In each patient a series of 4 to 13 consecutive days, in the absence of therapy, with 8 or more ischemic episodes per day were analyzed. Harmonic regression models were fitted to the hourly number of ischemic episodes and the hourly values of HR. Out of 54 days, with 8 or more episodes per day for a total of 1,357 episodes, a circadian rhythm was observed for 34 days (84 % ), in at least 1 day in all patients and during the entire period of observation in only 3. Its presence was independent of the number of episodes; the peak of periodic functions occurred at 2.9 f 2.7 AM. A
A
unique daily distribution of ischemic attacks has been described in patients with variant angina, with attacks being worse late at night or in the early morning.1-6 However, the frequency of this phenomenon remains controversia14t5 and no information is available on its persistence in the same patient during consecutive days. On the basis of the apparent specularity between the rhythm of ischemia in Prinzmetal’s angina and the heart rate (HR) rhythm reported for the normal population,’ a pathogenetic role of the parasympathetic nervous system in coronary vasospasm has been postulated.s However, in no instance has the time correlation between occurrence of ischemic attacks and circadian changes in HR been analyzed in the same patient and during the same period of time. The present study was undertaken to answer 3 questions: how frequently a circadian rhythm of ischemia is
observed in patients with Prinzmetal’s angina; whether the circadian rhythm of ischemia persists in individual patients; and whether a phase correlation exists between cyclic variations in ischemic attacks and HR.
Methods Patient selection: Eleven consecutive men, aged 38 to 68 years (mean 51 f 9), with variant angina admitted to the coronary care unit because of an exacerbation of angina were selected. The following criteria were required for entering the study: angina at rest, characterized by ST-segment elevation; 8 or more ischemic attacks per day (the lower limit required for periodic analysis] for at least 4 consecutive days of electrocardiographic (ECG) Holter monitoring, which required 4 to 13 days of Holter monitoring in each patient; and absence of treatment, except for sublingual nitrates when needed. The clinical characteristics of the patients are listed in Table I. Continuous electrocardiographic recording: All patients underwent ECG Holter monitoring while in the coronary care unit. In this environment patients rested approximately between 10 and 11 PM to 6 and 6:30 AM and had their meals at 7 to 7:30 AM, 12 to 1230 PM and 6 to 6:30 PM. A bipolar lead system recorder was used (ICR 7200). The most appropriate lead showing
From the C.N.R. Institute of Clinical Physiology and Istituto di Patologia Medica, University of Pisa, Pisa, Italy. This study was supported in part by Grant 850068404 from the C.N.R. Cardiopulmonary Project, Pisa, Italy. Manuscript received November 17,1986; revised manusript received February 18, 1987, accepted February 19,1987. Address for reprints: Clara Carpeggiani, MD, C.N.R. Institute of Clinical Physiology, via Savi, 8, 56100 Pisa, Italy. 36
the highest ST-segment elevation during an angina1 attack was chosen, and in all cases and inferior and a precordial lead were recorded. Postural changes (supine, sitting and standing) of QRS and ST segment were obtained for reference. Diary and therapy of each patient were accurately timed. Tapes were analyzed using a full disclosure system. Episodes of transient ST-segment elevation of 0.1 mV or greater, lasting for at least 1 minute, were considered indicative of ischemia.g Data analysis: Identified ischemic episodes were pooled for 1 hour periods. HR was sampled for a lminute period every 10 minutes, averaged per hour, and 24 values per day were derived. To quantify the periodic structure of the frequency of ischemic attacks and of HR, the hourly data were fitted by a harmonic regression model. For the single-day periodic analysis of ischemic attacks the lower limit of 8 attacks per day was arbitrarily chosen. The period of oscillation was taken to be 24 hours. Model “goodness of fit”lO was evaluated by the modified curvilinear correlation index (Rz]: [Z(yi - y12/Z(yi - yz], where vi = fitted theoretical values, y = fourier transformat first coefficient (a’), and yi = experimental data.
Results To obtain at least 4 consecutive 24-hour ECG recordings containing 8 or more ischemic events each, periods of continuous monitoring in each patient varied from 4 to 13 days. A total of 70 days of ECG recording was analyzed in the 11 patients; 1,397 episodes of ST-segment elevation (ranging from 0.1 to 1 mV) were recorded, only 12% of which were symptomatic. Of 70 days, 54 were found to have 8 or more attacks for a total of 1,357 ischemic episodes. Daily distribution of iscbemic episodes: For this analysis only the fifty-four 24-hour ECG recordings containing at least 8 attacks each were considered. Cumulative data: The cumulative daily distribution of the 1,357 ischemic events is shown in Figure 1. Most of the ischemic attacks occurred during the night (593 from 12 PM to 7 AM] with the maximum number occur-
EPISODES
l?am
12pm
FIGURE 1. Cumulative daily distribution of 1,357 ischemic episodes recorded in 54 days with 8 or more episodes per day in 11 patients with variant angina. Most of the ischemic episodes occurred during the night (593 from 12 PM to 7 AM) with the maximum values from 3 to 6 AM (21%). The symptomatic episodes (12% of the total) are evenly distributed during the 24 hours.
I
Summary
linical
Data --___
No. of CAs
Pt 1 2 3 4 5 6 7
a 9 10 11
Age W
History
Site of ST Elevation
54 38 51 46 49 38 68 58 61 55 45
E R+E R+E R+E R R R R R+E R+P R
A I I I A-L A A-L I A I I
Exercise Stress, Test
Narrowed Angio
+ + + 0 -
3 3 3(5) 2 -
0 f 0 0 0 +
1 2
by -
l(S) l(S) -
A = anterior; Angio = angiography: E = exertional angina: I = inferior; L = lateral: P = previous myocardial infarction; R = rest angina; (S) = coronary spasm; + = positive test: 0 = negative test; - = not done.
ring from 3 to 6 AM (21%). In contrast, symptomatic episodes were evenly distributed throughout the 24 hours. Daily analysis: When the hourly data of single days were fitted by the harmonic-regression model a significant circadian rhythm was found on 34 of 54 days (64%) The presence of a significant daily rhythm was independent of the number of ischemic attacks recorded. Peak value of the periodic function was at 2.9 f 2.7 AM (& standard deviation] (range 8.18 PM to 9.11 AM].
Intrapatient analysis: A significant circadian rhythm of ischemic events was found in all patients in at least 1 day. A circadian distribution was consistently found during the entire period of observation in 3 patients and not in the other 8. Similar to circadian distribution, the variability in peak time was independent of the daily frequency of events. Heart rate: Daily analysis: The harmonic-regression model was applied to the 70 days of Holter monitoring. Significant daily rhythm of HR was observed in all days but 9 187%). The nadir of the periodic function was at 2.4 f 1.5 AM (range 10.32 PM to 5.02 AM). Intrapatient analysis: Seven patients (64%) showed consistently a daily distribution of HR periodicity during the entire period of observation (4 days in 6 patients and 9 days in 1 patient); in the remaining 4 patients it was absent in 1 of lo,4 of 10,3 of 13 and 1 of 4 days. Possible clinical and environmental causes for such loss of HR periodicity were searched for without success. Time correlation between ischemic attacks an heart rate: Coherence between cycling in HR and ischemia was found for 32 days showing both rhythms. In this subset the intrapatient difference between peak of the ischemic function and the nadir of the function was on average 2.6 f 3.3 hours. An example of such coherence is shown in Figure 2. Lack of coherence was found in the following combinations: in 14 instances, a circadian rhythm of HR but not of ischemia was present despite a sufficient daily number of episodes; in contrast, in 2 cases a circadian rhythm of ischemia was not accompanied by a circadian rhythm
38
ISCHEMIA
AND HEART
RATE RHYTHMS
IN VARIANT
ANGINA
of HR. In 6 instancesrhythms were absent both for ischemia and HR. An example of lack of coherenceis shown in Figure 3.
Discussion Several investigators1*6J0 have reported that in Prinzmetal’s angina, ischemic attacks tend to occur during certain hours of the day, with the highest frequency in the early morning. This demonstrationhas been basedeither on the description of the daily distribution of pooled ischemic episodesrecordedin sever-
al patients1T3 or on the comparisonbetween the number of attacksin two l&hour periods of single days.5 To our knowledge no study has used chronobiologic mathematical models. The main task in chronobiology is to derive the most significant periodic functions from the frequencyanalysisof the data.The simple description of the %-hour distribution of pooled data from different days and different subjects(Fig. l] doesnot allow the definition of single periodic functions or the assessmentof the actual prevalence,incidence and reproducibility of the phenomenonin the
TIME TIME FIGURE 2. Single-day periodic analysis of ischemic episodes (left) and heart rate (right). Both hourly number of ischemic episodes ues of heart rate are fitted by a periodic function. A coherence between cycling in heart rate and ischemia is present. Time in hours 24 as 12 AM; 12 as 12 PM. The central hortzontal lines represent average values durtng 24 hours.
TIME FIGURE rate.
3. In this case
the hourly
experimental
and val= 0 and
TIME data
of the tschemic
episodes
are not fitted
by the periodic
function,
as is the case
for heart
diy
populations studied. On the other hand the applicability of frequency analysis to 24-hour data is conditioned by the lower limit in the number of observations; we arbitrarily choose the lower limit of 8 attacks per day to enhance statistical significance. Most days analyzed yielded many more attacks [55% had more than 20 attacks/day). Circadian rhythm of ischemic attacks: A circadian rhythm was documented in all patients for at least 1 of the &day minimum period of observation. In 27% it was consistently found during the entire period of observation (patients 2,5 and 6); in all but 2 it was present in at least 50% of the days. This underlines the high prevalence of circadian variation of ischemic attacks in this population as well as the high incidence of the phenomenon within the same patient. Both conclusions are dependent upon the duration of observation and independent of the number of attacks per day, according to Araki3, on the condition that 24-hour periods with a sufficient number of attacks are considered. Symptomatic ischemic episodes: The low proportion of symptomatic episodes did not allow a separate analysis of this subset. Pooled symptomatic ischemic episodes show a uniform distribution over the 24hour period; this is in accordance with previous observations.1,3Jj Circadian rhythm of heart rate: A long-term, reproducible, 24-hour analysis of rhythm of HR has been reported in normal subjects, with the maximal reduction (nadir) during the night.7 This pattern is apparently abolished by autonomic dysfunction, as in orthostatic hypotension ,I1 diabetic peripheral neuropathy and after cardiac transplantation.12 In the present study a circadian rhythm was documented in all patients for at least 3 days and in 87% of the analyzed days. The nadir of the calculated periodic function occurred from 10.32 PM to 5.02 AM [mean 2.4 f 1.5 AM). This time precedes that described for normal persons (nadir at 4 AM)~ and for patients with chronic ischemic heart disease in whom it is reported to be shifted toward midday.13 The relevance of these differences is difficult to establish due to possible methodologic, environmental and ethnic differences. A circadian rhythm of HR was consistently found in 64% of patients. Although in the remaining patients its absence was infrequent, the reason for the loss of rhythmicity remains obscure. Pathogenetic implications: No solid data are available to explain the unique daily distribution of ischemic attacks in Prinzmetal’s angina. Its relation to REM sleep14 has not been confirmed.15 Although daily rhythms have been ascertained for numerous endogenous substances, their phase shift relative to peak time of ischemia is an argument against their role in the genesis of coronary vasospasm. HR is believed to reflect the autonomic nervous activity, whose role in the genesis of coronary spasm is still under consideration.16-18 The task of the present study was to document a possible phase coherence in the daily frequency of HR and ischemic attacks which might support a hypothesis for a role of the autonomic nervous system in triggering coronary smooth muscle contraction. Beside the fre-
1,
quent occurrence 3f iack of rhythmicity in ischemia relative to HR and vice versa, in pleriods in which both rhythms were simultaneously present (32 days) the shift of philse averaged 2.7 f 3.4’hours. This single-day inconsistency of phase between HR and ischemia would not have been appreciated by the analysis of the pooled data. In conclusion, a circadian rhythm of ischemia is a frequent event in patients with variant angina. Its occurrence can be documented in all patients on the condition that the observation period is long enough and the number of events sufficiently great. HR circadian rhythm is more consistent and reproducible phenomenon, although occasional loss of HR rhythm can be observed in sequential 24-hour periods, without apparent reason. Bn the basis of a single day analysis a time correlation between HR and ischemia appears inconclusive. Acknowledgment: We are particularly indebted to Prof. Gerald A. Klassen for his useful and friendly criticism and to Dr. Felicita Andreotti for reviewing the manuscript for English language.
eferenrce 1. Biagini A, Carpeggiani C, Mazzei MG, Testa R, Michelassi C, Antonelli 19. L’Abbate A, Maseri A. Distribuzione oraria degli episodi di angina a riposo. Effetto della terapia medica. G Ital Cardiol 1981;11:4-11. 2. Salerno J, De Marco R, Medici A, Previtali M, Chimienti M, Ray M, De Servi S, Tavazzi L, Bobba P. Clinical and physiopathologic aspects derived from continuous eJectrocardiographic recording in angina pectoris at rest. In: Bertrand ME, ed. Coronary Arterial Spasm. Paris: Imprimee Laboratoires Dausse, 1981:126-138. 3. Araki H, Koiwaya Y, Nakagaki 0, Nakamura M. Diurnal distribution of STsegment elevation and related arrhythmias in patients with variant angina: a study by ambulatory ECG monitoring. Circulation 1983;67:995-1000. 4. Kimura E. Variant form of angina pectoris: its clinical features and treatment. Jpn Heart [ 1980;21:131-139. 5. Waters DD, Miller DD, Bouchard A, Bosch X, ThQoux P. Circadian variation in variant angina. Am r Cardiol 1984;54:61-64. 6. Von Arnim T, Hofling B, Schreiber M. Characteristics of episodes of ST elevation or ST depression during ambulatorv monitoring in natients subsequently undergoing coronary anaography. E& Heart J 1;85;;4:484-488. 7. Millar-Craig MW, Bishop CN, Raftery EB. Circadian variation of blood pressure. Lancet 1978;1:795-797. 8. Endo M, Hirosawa K, Kaneko N. Hase K. Inoue Y. Konno S. Coronarv arteriogram and left ventriculogram during angina attack induced by Methacholine. N Engl 1 Med 1976;294:252-255. 9. Bragg-Remschel DA, Anderson CH, Winkle RA. Frequency response characteristics of ambuIatory ECG monitoring system and iheirimpiications for ST segment analysis. Am Heart 1 1982;103:20-31. 10. Bliss CI. Statistics in Biology. New York: McGraw-Hill, 1970:219. 11. Mann S, Altman DC, Raftery EB, Bannister RR. Circadian variations of blood pressure in autonomic failure. Circulation 1983;68:477-483. 12. Reeves RA, Shapiro A, Thompson ME, Johnsen AM. Loss of nocturnal decline in blood pressure after cardiac transplantation. Circulation 1986; 73:401-408.
13. Aslanian NL, Adamian KG, Grigorian SV, Bagdassarian RA, Assatrian DG. Circadian rhythms of ECG T-wave, arterial pressure and heart rate in patients with ischemic heart disease. Chronobiologia 1980;7:481-492. 14. King MJ, Zir LM, Kaltman A], Fox AC. Variant angina associated with angiographically demonstrated coronary artery spasm and REM sleep. Am Med Sci 1973;265:419-422. 15. Maggini G, Guazzelli M, Mauri M, Chierchia, Cassano GB, Maseri A. Relation of transient myocardial ischemia to the sleep pattern in patients with primary angina. In: Maseri A, Klassen GA, Lesch M, eds. Primary and Secondary Angina Pectoris. New York: Grune 8 Stratton, 1978:157-167. 16. Yasue H, Touyama M, Kate H, Tanaka S, Akiyama F. Prinzmetal’s variant form of angina as a manifestation of alpha-adrenergic receptor-mediated coronary artery spasm. Documentation by coronary arteriography. Am Heart T 1976:91:148-155. 17. Robertson D, Robertson RM, Nies AS, Oates ]A. Friesinger GC. Variant angina pectoris: investigation of indexes of sympathetic nervous system. Am J Cardiol 1979;43:1080-1085. 18. Robertson RM, Bernard Y, Robertson D. Arterial and coronary sinus catecholamines in the course of spontaneous coronary artery spasm. Am Heart r 1983;105:901-908.
cadian rhythm for HR was observed in 81 of the 70 days (87 % ), consistently in 7 patients; the nadir occurred at 2.4 f 1.5 AM; simultaneous cycling in HR and transient ischemia was found on 32 days. The intrapatient difference between the peak and the nadir of the ischemic and the HR function was, on average, 2.6 f 3.3 hours. Thus, a circadian rhythm of ischemic episodes was present in all patients, although it was not consistently present; simultaneous occurrence of circadian variation in ischemic episodes and HR was observed only in 60% of the days with a sufficiently high number of attacks and when this occurred, a significant phase shift was observed; occasional loss of HR cycling was observed in some patients, without an apparent cause. (Am J Cardiol 1987;60:36-39)
Eleven patients with frequent episodes of variant angina underwent 24-hour electrocardiographic monitoring in a coronary care unit for a total of 70 days to assess circadian variation in ischemic episodes and its correlation with circadian heart rate (HR) rhythm. In each patient a series of 4 to 13 consecutive days, in the absence of therapy, with 8 or more ischemic episodes per day were analyzed. Harmonic regression models were fitted to the hourly number of ischemic episodes and the hourly values of HR. Out of 54 days, with 8 or more episodes per day for a total of 1,357 episodes, a circadian rhythm was observed for 34 days (84 % ), in at least 1 day in all patients and during the entire period of observation in only 3. Its presence was independent of the number of episodes; the peak of periodic functions occurred at 2.9 f 2.7 AM. A
A
unique daily distribution of ischemic attacks has been described in patients with variant angina, with attacks being worse late at night or in the early morning.1-6 However, the frequency of this phenomenon remains controversia14t5 and no information is available on its persistence in the same patient during consecutive days. On the basis of the apparent specularity between the rhythm of ischemia in Prinzmetal’s angina and the heart rate (HR) rhythm reported for the normal population,’ a pathogenetic role of the parasympathetic nervous system in coronary vasospasm has been postulated.s However, in no instance has the time correlation between occurrence of ischemic attacks and circadian changes in HR been analyzed in the same patient and during the same period of time. The present study was undertaken to answer 3 questions: how frequently a circadian rhythm of ischemia is
observed in patients with Prinzmetal’s angina; whether the circadian rhythm of ischemia persists in individual patients; and whether a phase correlation exists between cyclic variations in ischemic attacks and HR.
Methods Patient selection: Eleven consecutive men, aged 38 to 68 years (mean 51 f 9), with variant angina admitted to the coronary care unit because of an exacerbation of angina were selected. The following criteria were required for entering the study: angina at rest, characterized by ST-segment elevation; 8 or more ischemic attacks per day (the lower limit required for periodic analysis] for at least 4 consecutive days of electrocardiographic (ECG) Holter monitoring, which required 4 to 13 days of Holter monitoring in each patient; and absence of treatment, except for sublingual nitrates when needed. The clinical characteristics of the patients are listed in Table I. Continuous electrocardiographic recording: All patients underwent ECG Holter monitoring while in the coronary care unit. In this environment patients rested approximately between 10 and 11 PM to 6 and 6:30 AM and had their meals at 7 to 7:30 AM, 12 to 1230 PM and 6 to 6:30 PM. A bipolar lead system recorder was used (ICR 7200). The most appropriate lead showing
From the C.N.R. Institute of Clinical Physiology and Istituto di Patologia Medica, University of Pisa, Pisa, Italy. This study was supported in part by Grant 850068404 from the C.N.R. Cardiopulmonary Project, Pisa, Italy. Manuscript received November 17,1986; revised manusript received February 18, 1987, accepted February 19,1987. Address for reprints: Clara Carpeggiani, MD, C.N.R. Institute of Clinical Physiology, via Savi, 8, 56100 Pisa, Italy. 36
the highest ST-segment elevation during an angina1 attack was chosen, and in all cases and inferior and a precordial lead were recorded. Postural changes (supine, sitting and standing) of QRS and ST segment were obtained for reference. Diary and therapy of each patient were accurately timed. Tapes were analyzed using a full disclosure system. Episodes of transient ST-segment elevation of 0.1 mV or greater, lasting for at least 1 minute, were considered indicative of ischemia.g Data analysis: Identified ischemic episodes were pooled for 1 hour periods. HR was sampled for a lminute period every 10 minutes, averaged per hour, and 24 values per day were derived. To quantify the periodic structure of the frequency of ischemic attacks and of HR, the hourly data were fitted by a harmonic regression model. For the single-day periodic analysis of ischemic attacks the lower limit of 8 attacks per day was arbitrarily chosen. The period of oscillation was taken to be 24 hours. Model “goodness of fit”lO was evaluated by the modified curvilinear correlation index (Rz]: [Z(yi - y12/Z(yi - yz], where vi = fitted theoretical values, y = fourier transformat first coefficient (a’), and yi = experimental data.
Results To obtain at least 4 consecutive 24-hour ECG recordings containing 8 or more ischemic events each, periods of continuous monitoring in each patient varied from 4 to 13 days. A total of 70 days of ECG recording was analyzed in the 11 patients; 1,397 episodes of ST-segment elevation (ranging from 0.1 to 1 mV) were recorded, only 12% of which were symptomatic. Of 70 days, 54 were found to have 8 or more attacks for a total of 1,357 ischemic episodes. Daily distribution of iscbemic episodes: For this analysis only the fifty-four 24-hour ECG recordings containing at least 8 attacks each were considered. Cumulative data: The cumulative daily distribution of the 1,357 ischemic events is shown in Figure 1. Most of the ischemic attacks occurred during the night (593 from 12 PM to 7 AM] with the maximum number occur-
EPISODES
l?am
12pm
FIGURE 1. Cumulative daily distribution of 1,357 ischemic episodes recorded in 54 days with 8 or more episodes per day in 11 patients with variant angina. Most of the ischemic episodes occurred during the night (593 from 12 PM to 7 AM) with the maximum values from 3 to 6 AM (21%). The symptomatic episodes (12% of the total) are evenly distributed during the 24 hours.
I
Summary
linical
Data --___
No. of CAs
Pt 1 2 3 4 5 6 7
a 9 10 11
Age W
History
Site of ST Elevation
54 38 51 46 49 38 68 58 61 55 45
E R+E R+E R+E R R R R R+E R+P R
A I I I A-L A A-L I A I I
Exercise Stress, Test
Narrowed Angio
+ + + 0 -
3 3 3(5) 2 -
0 f 0 0 0 +
1 2
by -
l(S) l(S) -
A = anterior; Angio = angiography: E = exertional angina: I = inferior; L = lateral: P = previous myocardial infarction; R = rest angina; (S) = coronary spasm; + = positive test: 0 = negative test; - = not done.
ring from 3 to 6 AM (21%). In contrast, symptomatic episodes were evenly distributed throughout the 24 hours. Daily analysis: When the hourly data of single days were fitted by the harmonic-regression model a significant circadian rhythm was found on 34 of 54 days (64%) The presence of a significant daily rhythm was independent of the number of ischemic attacks recorded. Peak value of the periodic function was at 2.9 f 2.7 AM (& standard deviation] (range 8.18 PM to 9.11 AM].
Intrapatient analysis: A significant circadian rhythm of ischemic events was found in all patients in at least 1 day. A circadian distribution was consistently found during the entire period of observation in 3 patients and not in the other 8. Similar to circadian distribution, the variability in peak time was independent of the daily frequency of events. Heart rate: Daily analysis: The harmonic-regression model was applied to the 70 days of Holter monitoring. Significant daily rhythm of HR was observed in all days but 9 187%). The nadir of the periodic function was at 2.4 f 1.5 AM (range 10.32 PM to 5.02 AM). Intrapatient analysis: Seven patients (64%) showed consistently a daily distribution of HR periodicity during the entire period of observation (4 days in 6 patients and 9 days in 1 patient); in the remaining 4 patients it was absent in 1 of lo,4 of 10,3 of 13 and 1 of 4 days. Possible clinical and environmental causes for such loss of HR periodicity were searched for without success. Time correlation between ischemic attacks an heart rate: Coherence between cycling in HR and ischemia was found for 32 days showing both rhythms. In this subset the intrapatient difference between peak of the ischemic function and the nadir of the function was on average 2.6 f 3.3 hours. An example of such coherence is shown in Figure 2. Lack of coherence was found in the following combinations: in 14 instances, a circadian rhythm of HR but not of ischemia was present despite a sufficient daily number of episodes; in contrast, in 2 cases a circadian rhythm of ischemia was not accompanied by a circadian rhythm
38
ISCHEMIA
AND HEART
RATE RHYTHMS
IN VARIANT
ANGINA
of HR. In 6 instancesrhythms were absent both for ischemia and HR. An example of lack of coherenceis shown in Figure 3.
Discussion Several investigators1*6J0 have reported that in Prinzmetal’s angina, ischemic attacks tend to occur during certain hours of the day, with the highest frequency in the early morning. This demonstrationhas been basedeither on the description of the daily distribution of pooled ischemic episodesrecordedin sever-
al patients1T3 or on the comparisonbetween the number of attacksin two l&hour periods of single days.5 To our knowledge no study has used chronobiologic mathematical models. The main task in chronobiology is to derive the most significant periodic functions from the frequencyanalysisof the data.The simple description of the %-hour distribution of pooled data from different days and different subjects(Fig. l] doesnot allow the definition of single periodic functions or the assessmentof the actual prevalence,incidence and reproducibility of the phenomenonin the
TIME TIME FIGURE 2. Single-day periodic analysis of ischemic episodes (left) and heart rate (right). Both hourly number of ischemic episodes ues of heart rate are fitted by a periodic function. A coherence between cycling in heart rate and ischemia is present. Time in hours 24 as 12 AM; 12 as 12 PM. The central hortzontal lines represent average values durtng 24 hours.
TIME FIGURE rate.
3. In this case
the hourly
experimental
and val= 0 and
TIME data
of the tschemic
episodes
are not fitted
by the periodic
function,
as is the case
for heart
diy
populations studied. On the other hand the applicability of frequency analysis to 24-hour data is conditioned by the lower limit in the number of observations; we arbitrarily choose the lower limit of 8 attacks per day to enhance statistical significance. Most days analyzed yielded many more attacks [55% had more than 20 attacks/day). Circadian rhythm of ischemic attacks: A circadian rhythm was documented in all patients for at least 1 of the &day minimum period of observation. In 27% it was consistently found during the entire period of observation (patients 2,5 and 6); in all but 2 it was present in at least 50% of the days. This underlines the high prevalence of circadian variation of ischemic attacks in this population as well as the high incidence of the phenomenon within the same patient. Both conclusions are dependent upon the duration of observation and independent of the number of attacks per day, according to Araki3, on the condition that 24-hour periods with a sufficient number of attacks are considered. Symptomatic ischemic episodes: The low proportion of symptomatic episodes did not allow a separate analysis of this subset. Pooled symptomatic ischemic episodes show a uniform distribution over the 24hour period; this is in accordance with previous observations.1,3Jj Circadian rhythm of heart rate: A long-term, reproducible, 24-hour analysis of rhythm of HR has been reported in normal subjects, with the maximal reduction (nadir) during the night.7 This pattern is apparently abolished by autonomic dysfunction, as in orthostatic hypotension ,I1 diabetic peripheral neuropathy and after cardiac transplantation.12 In the present study a circadian rhythm was documented in all patients for at least 3 days and in 87% of the analyzed days. The nadir of the calculated periodic function occurred from 10.32 PM to 5.02 AM [mean 2.4 f 1.5 AM). This time precedes that described for normal persons (nadir at 4 AM)~ and for patients with chronic ischemic heart disease in whom it is reported to be shifted toward midday.13 The relevance of these differences is difficult to establish due to possible methodologic, environmental and ethnic differences. A circadian rhythm of HR was consistently found in 64% of patients. Although in the remaining patients its absence was infrequent, the reason for the loss of rhythmicity remains obscure. Pathogenetic implications: No solid data are available to explain the unique daily distribution of ischemic attacks in Prinzmetal’s angina. Its relation to REM sleep14 has not been confirmed.15 Although daily rhythms have been ascertained for numerous endogenous substances, their phase shift relative to peak time of ischemia is an argument against their role in the genesis of coronary vasospasm. HR is believed to reflect the autonomic nervous activity, whose role in the genesis of coronary spasm is still under consideration.16-18 The task of the present study was to document a possible phase coherence in the daily frequency of HR and ischemic attacks which might support a hypothesis for a role of the autonomic nervous system in triggering coronary smooth muscle contraction. Beside the fre-
1,
quent occurrence 3f iack of rhythmicity in ischemia relative to HR and vice versa, in pleriods in which both rhythms were simultaneously present (32 days) the shift of philse averaged 2.7 f 3.4’hours. This single-day inconsistency of phase between HR and ischemia would not have been appreciated by the analysis of the pooled data. In conclusion, a circadian rhythm of ischemia is a frequent event in patients with variant angina. Its occurrence can be documented in all patients on the condition that the observation period is long enough and the number of events sufficiently great. HR circadian rhythm is more consistent and reproducible phenomenon, although occasional loss of HR rhythm can be observed in sequential 24-hour periods, without apparent reason. Bn the basis of a single day analysis a time correlation between HR and ischemia appears inconclusive. Acknowledgment: We are particularly indebted to Prof. Gerald A. Klassen for his useful and friendly criticism and to Dr. Felicita Andreotti for reviewing the manuscript for English language.
eferenrce 1. Biagini A, Carpeggiani C, Mazzei MG, Testa R, Michelassi C, Antonelli 19. L’Abbate A, Maseri A. Distribuzione oraria degli episodi di angina a riposo. Effetto della terapia medica. G Ital Cardiol 1981;11:4-11. 2. Salerno J, De Marco R, Medici A, Previtali M, Chimienti M, Ray M, De Servi S, Tavazzi L, Bobba P. Clinical and physiopathologic aspects derived from continuous eJectrocardiographic recording in angina pectoris at rest. In: Bertrand ME, ed. Coronary Arterial Spasm. Paris: Imprimee Laboratoires Dausse, 1981:126-138. 3. Araki H, Koiwaya Y, Nakagaki 0, Nakamura M. Diurnal distribution of STsegment elevation and related arrhythmias in patients with variant angina: a study by ambulatory ECG monitoring. Circulation 1983;67:995-1000. 4. Kimura E. Variant form of angina pectoris: its clinical features and treatment. Jpn Heart [ 1980;21:131-139. 5. Waters DD, Miller DD, Bouchard A, Bosch X, ThQoux P. Circadian variation in variant angina. Am r Cardiol 1984;54:61-64. 6. Von Arnim T, Hofling B, Schreiber M. Characteristics of episodes of ST elevation or ST depression during ambulatorv monitoring in natients subsequently undergoing coronary anaography. E& Heart J 1;85;;4:484-488. 7. Millar-Craig MW, Bishop CN, Raftery EB. Circadian variation of blood pressure. Lancet 1978;1:795-797. 8. Endo M, Hirosawa K, Kaneko N. Hase K. Inoue Y. Konno S. Coronarv arteriogram and left ventriculogram during angina attack induced by Methacholine. N Engl 1 Med 1976;294:252-255. 9. Bragg-Remschel DA, Anderson CH, Winkle RA. Frequency response characteristics of ambuIatory ECG monitoring system and iheirimpiications for ST segment analysis. Am Heart 1 1982;103:20-31. 10. Bliss CI. Statistics in Biology. New York: McGraw-Hill, 1970:219. 11. Mann S, Altman DC, Raftery EB, Bannister RR. Circadian variations of blood pressure in autonomic failure. Circulation 1983;68:477-483. 12. Reeves RA, Shapiro A, Thompson ME, Johnsen AM. Loss of nocturnal decline in blood pressure after cardiac transplantation. Circulation 1986; 73:401-408.
13. Aslanian NL, Adamian KG, Grigorian SV, Bagdassarian RA, Assatrian DG. Circadian rhythms of ECG T-wave, arterial pressure and heart rate in patients with ischemic heart disease. Chronobiologia 1980;7:481-492. 14. King MJ, Zir LM, Kaltman A], Fox AC. Variant angina associated with angiographically demonstrated coronary artery spasm and REM sleep. Am Med Sci 1973;265:419-422. 15. Maggini G, Guazzelli M, Mauri M, Chierchia, Cassano GB, Maseri A. Relation of transient myocardial ischemia to the sleep pattern in patients with primary angina. In: Maseri A, Klassen GA, Lesch M, eds. Primary and Secondary Angina Pectoris. New York: Grune 8 Stratton, 1978:157-167. 16. Yasue H, Touyama M, Kate H, Tanaka S, Akiyama F. Prinzmetal’s variant form of angina as a manifestation of alpha-adrenergic receptor-mediated coronary artery spasm. Documentation by coronary arteriography. Am Heart T 1976:91:148-155. 17. Robertson D, Robertson RM, Nies AS, Oates ]A. Friesinger GC. Variant angina pectoris: investigation of indexes of sympathetic nervous system. Am J Cardiol 1979;43:1080-1085. 18. Robertson RM, Bernard Y, Robertson D. Arterial and coronary sinus catecholamines in the course of spontaneous coronary artery spasm. Am Heart r 1983;105:901-908.