- Email: [email protected]
Circadian rhythms in dynamic electrocardiography
J. ELECTROCARDIOLOGY 16 (4), 1983, 351-354
Circadian Rhythms in Dynamic Electrocardiography BY VITTORIO DE LEONARDIS, M.D.*, PIERO CINELLI, M.D.,I"F i n l O CAPACCI, MAURIZlO DE SCALZI, M . D . * AND SILIANO CITI, M.D/f
SUMMARY A group of ten healthy subjects aged 32-58 years was studied by dynamic electrocardiography. An electrocardiogram (ECG) was recorded on CM 5 lead for 14 seconds every 15 minutes during the 24 hours. All the subjects were under a similar nychthemeral schedule, sleeping between 23.30 and 06.45, and all followed their spontaneous diet without any restriction. "Mean Cosinor" analysis of the data revealed statistically significant circadian rhythms of heart rate (HR), R and T-wave voltages, duration of the QT interval and ST segment displacement, but not of the duration of the corrected QT (QTc).
tion and were under a similar nychthemeral schedule, sleeping between 23.30 and 06.45. They were considered healthy by means of routine clinical and laboratory examinations and ECGs recorded at rest and after exercise. An ECG was recorded for 14 seconds every 15 minutes during the 24 hours by magnetic tape recorder Cardiocassette II (Cardiodyne, Cupertino, California). The exploring electrode was placed in the conventional V5 position, the ground electrode was attached to the manubrium sterni and the negative electrode was placed at the right sternal border in the fourth intercostal space. Control tracings were recorded with the subject standing, sitting, in supine and prone positions, and lying on his left and right side. The magnetic tapes were transferred to standard electrocardiographic paper for manual measurement of the following parameters: HR (beat/min), duration of the QT interval (msec), amplitude of R and T-waves (mm) (10 mm = 1 mV), and ST segment displacement (mm). In both the control tracings and the dynamic ECG, ItR and QT were computed averaging the measurements performed on five heart beats, while the voltage of R and T-waves and the ST segment displacement were computed where the QRS complex showed the maximal amplitude. QT interval was corrected for HR by means of the following formula: QTr = QTN RR. The 5,231 measurements obtained were analyzed by the "Mean Cosinor" method for both detection and characterization of circadian rhythm7. A statistically significant circadian rhythm is demonstrated by: Mesor (a rhythm-determined average), Amplitude (measure of one half the extent of rhythmic change in a cycle estimated by the function used to approximate the rhythm; e.g. half of the peaktrough difference) and Acrophase (measure of timing; the lag from a defined reference time point of the crest of the cosine curve used to approximate a rhythm) along with 95% confidence limitss.
Circadian r h y t h m s of heart rate (HR) in healthy subjects1-4 and in subjects who in the pas t had myocardial infarction3 are known. A statistically s i g n i f i c a n t circadian r h y t h m has also been observed for the voltage of the T-wave of the E C G in healthy subjects and in patients with ischemic heart diseases. F u r th er mo r e, circadian variations for myocardial ischemic atta cks a t rest have been reported6. We are not aware of previous works concerning group studies about circadian variations for the duration and/or the amplitude of other electrocardiographic parameters, such as QT interval and R-wave. T h e aim of this work is to s t u d y the HR, the voltage of R and T-waves, the duration of QT interval non-corrected and corrected for H R (QTc) and the possible S T s e gm ent displacement, in order to determine their circadian variability. M A T E R I A L S AND M E T H O D S We studied a group of ten presumably healthy males aged 32-58 (mean age 44 ___8) (Table I), while they performed their normal everyday activities. All the subjects followed their spontaneous diet without restricFrom the Istituto di Clinica Medica II, Cattedra di Semeiotica Medica, Universit~ degli Studi di Firenze. *Ricercatore, Clinica Medica II. ~Assistente, Clinica Medica II. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. w1734 solely to indicate this fact. Reprint requests to: Vittorio de Leonardis, Istituto di Clinica Medica II, Viale Pieraccini 18, 50139, Firenze-Italia.
351
352
DE LEONARDIS
ET A L
TABLE I Age of the ten subjects studied. Subject 1
G.P.
2 3 4 5 6
V.P. B.L. F.G. B.G. S.L.
7
B.R.
Age 32 36 38 39 41 45 46 50 50 58
8 C.O. 9 P.G. 10 B.A.
RESULTS Table II shows that there are not statistically significant differences between mean values of the control tracings recorded with the subject lying down and those recorded with the subject standing and sitting. Statistically significant circadian r h y t h m s were documented for all the variables except QTr Table III summarizes the rhythmometric characteristics of HR, QT interval, QTr ST segment and R and T-waves. Figs. 1 and 2 show that the Acrophase of H R circadian rhythm occurs at the midday hours while the Acrophases of the other circadian r h y t h m s detected occur in the early morning hours or at night. Fig. 3 shows the lack of a circadian t h y t h m for QTr The number of the measurements of the ST segment displacement (which was always < 1.5 mm) is smaller if compared with that of the other parameters (Table III); in fact only five of the ten
Fig. 1 Cosinor representation of amplitude and Aerophase and their 95% confidence limits for QT interval (A) and heart rate {B). The Acrophase for the circadian rhythm in the QT interval occurs about 12 hours out of phase with the rhythm in heart rate. There is a 95% probability that the parameters lie within the region of the parameter space enclosed within the confidence ellipses. Amplitude shown as % of Mesor.
subjects studied were found to have ST elevation with an appearance of upward concavity and only one was found to have upsloping ST depression. In detail: subjects 1, 4, 6, 8, 9 (Table I) had increased nocturnal values for ST indicating an increase of a positive value, while subject 7 had a decrease of a negative value.
T A B L E II Mean values of heart rate (HR), QT interval, R and T-wave voltages and ST segment displacement found in the control tracings and their variations between lying down position (A) and standing and sitting positions (B). Lying down position mean values and standing and sitting position mean values were obtained averaging the measurements performed on the highest QRS complex respectively recorded in each of the positions described in the text. Ao. Variable Heart Rate QT R T ST
S~% (Mean _ S.D.)
b/min msec mm mm mm
68 378 18.6 5.1 0.2
+ 11 + 36 + 7.5 + 2.9 • 0.4
P
74 364 17.5 5.2 0.3
+ 15 + 32 + 7.3 • 2.8 • 0.4
N.S. N.S. N.S. N.S. N.S.
A = Lying down position; B = Standing and sitting positions N.S. = Not significant Number of observations for each variable: 99 = 4 0 ; ,'~ = 20.
J. E L E C T R O C A R D I O L O G Y
16 (4), 1983
ECG AND CIRCADIAN RHYTHMS
353
a circadian rhythm for QTr demonstrates that this interval depends only on HR variations and that it is not influenced by other vegetative rhythms. The voltages of R and T-waves show statistically significant circadian rhythms of moderate amplitude with Acrophases at 02.55 and 04.57 respectively. These results differ from those reported by Smolensky9 who, re-evaluating earlier data from Bergesl0, showed that the Acrophases for the circadian rhythms of the voltages of R and T-waves occur in phase with the rhythm in HR. These data are obtained from only one healthy subject while ours refer to a group of ten subjects. Our results differ also from those obtained by Aslanlan5 who showed that the Acrophase of the voltage of T-wave in a group of 25 healthy subjects occurs during the midday hours. In both these studies data were not collected by dynamic electrocardiography, and this makes comparison meaningless. In normal subjects the ST segment may show a displacement even if always below 1.5 mm. Of interest is the fact that the Acrophase for the rhythm of ST segment displacement occurs at 03.56, that particular time of the day when myocardial ischemic attacks at rest seem to be more frequent6. Since the Acrophases of circadian rhythm in R and T-wave voltages and in ST segment displacement occur during the night, together with the Acrophase of the QT interval, the question of a relationship between these circadian rhythms and the HR arises. Our unpublished data, however, exclude this relationship, suggesting the possible existence of endogenous circadian rhythms in cardiac activity which are unrelated to the rhythm in HR. The data here reported, documenting circadian
Fig. 2 Cosinor representation of amplitude and Acrophase and their 95% confidence limits for voltages of R (C) and T-waves (D), and ST segment displacement (E). This figure shows that the acrophases for the circadian rhythm in the three parameters occur between 02.55 and 04.57. Amplitude in nun.
DISCUSSION The data obtained for HR in this study show almost no difference from that reported by other authors1-4. The Acrophase for the rhythm of the QT interval differs by about 12 hours from that of the HR. This may be explained by the relationship between QT interval and HR. In fact, longer durations of the QT interval occur when HR is slower. Furthermore, when we correct the QT interval for HR, eliminating its influence, we do not find a significant circadian rhythm for QT c. The lack of
TABLE III Mean Cosinor Summary of the Rhythm Parameters for the ECG Variables Studied
Variable Heart Rate QT QTc R ST T
b/min msec msec mm mm mm
N.O.
P R.D.
933 933 933 933 556 933
.0001 .0000 .3539 .0098 .0268 .0051
Mesor __. S.E. 76 361 398 17 0.3 3.8
• • • • • •
1 2.3 3.2 0.2 0.03 O.1
N.O. = Number of observations; R.D. = Rhythm detection
J. ELECTROCARDIOLOGY 16 (4), 1983
Amplitude Acrophase (Hours) ( 9 5 % Confidence Limits) 11 23 6.5 1.5 0.4 0.8
(6.5; 14.7) ( 1 8 . 5 ; 29) ( -- ) (0.4; 2.8) ( 0 . 0 5 ; 0.8) ( 0 . 2 5 ; 1.2)
16.13 05.06 13.28 02.55 03.56 04.57
(14.56; (03.48; ( ) (21.08; (23.04; (03.36;
18.00) 06.44) 04.28) 06.24) 08.20)
354
DE LEONARDIS ET AL
3.
4.
5.
6. Fig. 3 The cosinor representation shows that the confidence ellipse overlaps the pole of the diagram; therefore QTc does not have a circadian rhythm. 7.
r h y t h m s in ECG parameters, suggest new investigations in this field to s t u d y the chronopharmacologic implications of clinical electrocardiographic findings.
REFERENCES 1. HALBERG,F, REINIIARDT,J, BARTTER,F, DELEA, C, GORDON, R, WOLFF, S, REINBERG, A, GnATA, J, HOFMANN, H, HALIIUMBER, M, GUNTIIER, R, KNAPP, E, PENA, J C AND GARCIA SAINZ, M:Agreement in end points from circadian rhythmometry on healthy human beings living on different continents. Experientia {Basel) 25:107, 1969 2. KANABROKI, S L, SCHEVING, L E, HALBERG, F, BREWER, R L AND BIRD, T J: Circadian variations
8.
9.
I0.
in presumably healthy men under conditions of peace army reserve unit training. Space Life Sci 4:258, 1973 LUCENTE, M, DOMENICIIELLI,B, CoPPoLA, E, DI FOLLA, A L, ALESSANDRI, N, SCHIAVONI, G, CAPANI, F, SENSI, S AND MANZOLL U: L'elettrocardiografia dinamica, un nuovo approccio alia cronobiologia. Min Cardioang 29:399, 1981 REINBERG,A, GIIATA, J, HALBRG,F, API-'ELBAUM, F, GEaVAm, P, BOUDON, P, AnULKER, C AND DUPONT, J: .Treatment schedules modify circadian timing in human adrenocortical insufficiency. In: Chronobiology, L SClmVlNG, F HALnERG ANDJ PAULY,eds. Igaku Shoin Ltd., Tokyo, 1974, p 168 ASLANIAN,N L, ADAMIAN,K G, GRIGORIAN,S V, BAGDASSARIAN,R A AND ASSATRIAN,D G: Circadian rhythms of ECG T-wave, arterial pressure and heart rate in patients with ischemic heart disease. Chronobiologia 7:481, 1980 BIAGINI, A, CARPEGGIANI, C, MAZZEI, M C, TESTA, R, MICIIELASSI, C, ANTONELLI, R, L'ABBATE, A AND MASERS, A: Distribuzione oraria degli episodi di angina a riposo. Effetto della terapia medica. G Ital Cardiol 11:4, 1981 HALBERG, F, TONG, Y L AND JOIINSON, E A: Circadian system phase: An aspect of temporal morphology. Procedures and illustrativeexamples. In: The Cellular Aspects of Biorhythms. Syrup. on Rhythmic Res. 8th Int. Cong. Anat., H MAYERSBACII,ed, Springer-Verlag, Berlin, 1967, p 20 HALBERG, F, CARANDENTE, F, CORNELlSSEN, G, AND KATINAS, G S: Glossary of chronobiology. Chronobiologia, Suppl. 1, p 11, Ponte Ed., Milano, 1977 SMOLENSKY, M H, TATAR, S E, BERGMAN, S A, LOSMAN, J G, BARNARD, C N, DAcso, C C AND KRAF, I A: Circadian rhythmic aspects of human cardiovascular function; A review by chronobiological statistical methods. Chronobiologia 3:337, 1976 BERGES, D: Untersuchungen uber das ausmass von tagesschvankungen in E K G gesunder versuchs-personen. Z Kreisl Forsch 54:35, 1965
J. ELECTROCARDIOLOGY 16 (4), 1983
Circadian Rhythms in Dynamic Electrocardiography BY VITTORIO DE LEONARDIS, M.D.*, PIERO CINELLI, M.D.,I"F i n l O CAPACCI, MAURIZlO DE SCALZI, M . D . * AND SILIANO CITI, M.D/f
SUMMARY A group of ten healthy subjects aged 32-58 years was studied by dynamic electrocardiography. An electrocardiogram (ECG) was recorded on CM 5 lead for 14 seconds every 15 minutes during the 24 hours. All the subjects were under a similar nychthemeral schedule, sleeping between 23.30 and 06.45, and all followed their spontaneous diet without any restriction. "Mean Cosinor" analysis of the data revealed statistically significant circadian rhythms of heart rate (HR), R and T-wave voltages, duration of the QT interval and ST segment displacement, but not of the duration of the corrected QT (QTc).
tion and were under a similar nychthemeral schedule, sleeping between 23.30 and 06.45. They were considered healthy by means of routine clinical and laboratory examinations and ECGs recorded at rest and after exercise. An ECG was recorded for 14 seconds every 15 minutes during the 24 hours by magnetic tape recorder Cardiocassette II (Cardiodyne, Cupertino, California). The exploring electrode was placed in the conventional V5 position, the ground electrode was attached to the manubrium sterni and the negative electrode was placed at the right sternal border in the fourth intercostal space. Control tracings were recorded with the subject standing, sitting, in supine and prone positions, and lying on his left and right side. The magnetic tapes were transferred to standard electrocardiographic paper for manual measurement of the following parameters: HR (beat/min), duration of the QT interval (msec), amplitude of R and T-waves (mm) (10 mm = 1 mV), and ST segment displacement (mm). In both the control tracings and the dynamic ECG, ItR and QT were computed averaging the measurements performed on five heart beats, while the voltage of R and T-waves and the ST segment displacement were computed where the QRS complex showed the maximal amplitude. QT interval was corrected for HR by means of the following formula: QTr = QTN RR. The 5,231 measurements obtained were analyzed by the "Mean Cosinor" method for both detection and characterization of circadian rhythm7. A statistically significant circadian rhythm is demonstrated by: Mesor (a rhythm-determined average), Amplitude (measure of one half the extent of rhythmic change in a cycle estimated by the function used to approximate the rhythm; e.g. half of the peaktrough difference) and Acrophase (measure of timing; the lag from a defined reference time point of the crest of the cosine curve used to approximate a rhythm) along with 95% confidence limitss.
Circadian r h y t h m s of heart rate (HR) in healthy subjects1-4 and in subjects who in the pas t had myocardial infarction3 are known. A statistically s i g n i f i c a n t circadian r h y t h m has also been observed for the voltage of the T-wave of the E C G in healthy subjects and in patients with ischemic heart diseases. F u r th er mo r e, circadian variations for myocardial ischemic atta cks a t rest have been reported6. We are not aware of previous works concerning group studies about circadian variations for the duration and/or the amplitude of other electrocardiographic parameters, such as QT interval and R-wave. T h e aim of this work is to s t u d y the HR, the voltage of R and T-waves, the duration of QT interval non-corrected and corrected for H R (QTc) and the possible S T s e gm ent displacement, in order to determine their circadian variability. M A T E R I A L S AND M E T H O D S We studied a group of ten presumably healthy males aged 32-58 (mean age 44 ___8) (Table I), while they performed their normal everyday activities. All the subjects followed their spontaneous diet without restricFrom the Istituto di Clinica Medica II, Cattedra di Semeiotica Medica, Universit~ degli Studi di Firenze. *Ricercatore, Clinica Medica II. ~Assistente, Clinica Medica II. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. w1734 solely to indicate this fact. Reprint requests to: Vittorio de Leonardis, Istituto di Clinica Medica II, Viale Pieraccini 18, 50139, Firenze-Italia.
351
352
DE LEONARDIS
ET A L
TABLE I Age of the ten subjects studied. Subject 1
G.P.
2 3 4 5 6
V.P. B.L. F.G. B.G. S.L.
7
B.R.
Age 32 36 38 39 41 45 46 50 50 58
8 C.O. 9 P.G. 10 B.A.
RESULTS Table II shows that there are not statistically significant differences between mean values of the control tracings recorded with the subject lying down and those recorded with the subject standing and sitting. Statistically significant circadian r h y t h m s were documented for all the variables except QTr Table III summarizes the rhythmometric characteristics of HR, QT interval, QTr ST segment and R and T-waves. Figs. 1 and 2 show that the Acrophase of H R circadian rhythm occurs at the midday hours while the Acrophases of the other circadian r h y t h m s detected occur in the early morning hours or at night. Fig. 3 shows the lack of a circadian t h y t h m for QTr The number of the measurements of the ST segment displacement (which was always < 1.5 mm) is smaller if compared with that of the other parameters (Table III); in fact only five of the ten
Fig. 1 Cosinor representation of amplitude and Aerophase and their 95% confidence limits for QT interval (A) and heart rate {B). The Acrophase for the circadian rhythm in the QT interval occurs about 12 hours out of phase with the rhythm in heart rate. There is a 95% probability that the parameters lie within the region of the parameter space enclosed within the confidence ellipses. Amplitude shown as % of Mesor.
subjects studied were found to have ST elevation with an appearance of upward concavity and only one was found to have upsloping ST depression. In detail: subjects 1, 4, 6, 8, 9 (Table I) had increased nocturnal values for ST indicating an increase of a positive value, while subject 7 had a decrease of a negative value.
T A B L E II Mean values of heart rate (HR), QT interval, R and T-wave voltages and ST segment displacement found in the control tracings and their variations between lying down position (A) and standing and sitting positions (B). Lying down position mean values and standing and sitting position mean values were obtained averaging the measurements performed on the highest QRS complex respectively recorded in each of the positions described in the text. Ao. Variable Heart Rate QT R T ST
S~% (Mean _ S.D.)
b/min msec mm mm mm
68 378 18.6 5.1 0.2
+ 11 + 36 + 7.5 + 2.9 • 0.4
P
74 364 17.5 5.2 0.3
+ 15 + 32 + 7.3 • 2.8 • 0.4
N.S. N.S. N.S. N.S. N.S.
A = Lying down position; B = Standing and sitting positions N.S. = Not significant Number of observations for each variable: 99 = 4 0 ; ,'~ = 20.
J. E L E C T R O C A R D I O L O G Y
16 (4), 1983
ECG AND CIRCADIAN RHYTHMS
353
a circadian rhythm for QTr demonstrates that this interval depends only on HR variations and that it is not influenced by other vegetative rhythms. The voltages of R and T-waves show statistically significant circadian rhythms of moderate amplitude with Acrophases at 02.55 and 04.57 respectively. These results differ from those reported by Smolensky9 who, re-evaluating earlier data from Bergesl0, showed that the Acrophases for the circadian rhythms of the voltages of R and T-waves occur in phase with the rhythm in HR. These data are obtained from only one healthy subject while ours refer to a group of ten subjects. Our results differ also from those obtained by Aslanlan5 who showed that the Acrophase of the voltage of T-wave in a group of 25 healthy subjects occurs during the midday hours. In both these studies data were not collected by dynamic electrocardiography, and this makes comparison meaningless. In normal subjects the ST segment may show a displacement even if always below 1.5 mm. Of interest is the fact that the Acrophase for the rhythm of ST segment displacement occurs at 03.56, that particular time of the day when myocardial ischemic attacks at rest seem to be more frequent6. Since the Acrophases of circadian rhythm in R and T-wave voltages and in ST segment displacement occur during the night, together with the Acrophase of the QT interval, the question of a relationship between these circadian rhythms and the HR arises. Our unpublished data, however, exclude this relationship, suggesting the possible existence of endogenous circadian rhythms in cardiac activity which are unrelated to the rhythm in HR. The data here reported, documenting circadian
Fig. 2 Cosinor representation of amplitude and Acrophase and their 95% confidence limits for voltages of R (C) and T-waves (D), and ST segment displacement (E). This figure shows that the acrophases for the circadian rhythm in the three parameters occur between 02.55 and 04.57. Amplitude in nun.
DISCUSSION The data obtained for HR in this study show almost no difference from that reported by other authors1-4. The Acrophase for the rhythm of the QT interval differs by about 12 hours from that of the HR. This may be explained by the relationship between QT interval and HR. In fact, longer durations of the QT interval occur when HR is slower. Furthermore, when we correct the QT interval for HR, eliminating its influence, we do not find a significant circadian rhythm for QT c. The lack of
TABLE III Mean Cosinor Summary of the Rhythm Parameters for the ECG Variables Studied
Variable Heart Rate QT QTc R ST T
b/min msec msec mm mm mm
N.O.
P R.D.
933 933 933 933 556 933
.0001 .0000 .3539 .0098 .0268 .0051
Mesor __. S.E. 76 361 398 17 0.3 3.8
• • • • • •
1 2.3 3.2 0.2 0.03 O.1
N.O. = Number of observations; R.D. = Rhythm detection
J. ELECTROCARDIOLOGY 16 (4), 1983
Amplitude Acrophase (Hours) ( 9 5 % Confidence Limits) 11 23 6.5 1.5 0.4 0.8
(6.5; 14.7) ( 1 8 . 5 ; 29) ( -- ) (0.4; 2.8) ( 0 . 0 5 ; 0.8) ( 0 . 2 5 ; 1.2)
16.13 05.06 13.28 02.55 03.56 04.57
(14.56; (03.48; ( ) (21.08; (23.04; (03.36;
18.00) 06.44) 04.28) 06.24) 08.20)
354
DE LEONARDIS ET AL
3.
4.
5.
6. Fig. 3 The cosinor representation shows that the confidence ellipse overlaps the pole of the diagram; therefore QTc does not have a circadian rhythm. 7.
r h y t h m s in ECG parameters, suggest new investigations in this field to s t u d y the chronopharmacologic implications of clinical electrocardiographic findings.
REFERENCES 1. HALBERG,F, REINIIARDT,J, BARTTER,F, DELEA, C, GORDON, R, WOLFF, S, REINBERG, A, GnATA, J, HOFMANN, H, HALIIUMBER, M, GUNTIIER, R, KNAPP, E, PENA, J C AND GARCIA SAINZ, M:Agreement in end points from circadian rhythmometry on healthy human beings living on different continents. Experientia {Basel) 25:107, 1969 2. KANABROKI, S L, SCHEVING, L E, HALBERG, F, BREWER, R L AND BIRD, T J: Circadian variations
8.
9.
I0.
in presumably healthy men under conditions of peace army reserve unit training. Space Life Sci 4:258, 1973 LUCENTE, M, DOMENICIIELLI,B, CoPPoLA, E, DI FOLLA, A L, ALESSANDRI, N, SCHIAVONI, G, CAPANI, F, SENSI, S AND MANZOLL U: L'elettrocardiografia dinamica, un nuovo approccio alia cronobiologia. Min Cardioang 29:399, 1981 REINBERG,A, GIIATA, J, HALBRG,F, API-'ELBAUM, F, GEaVAm, P, BOUDON, P, AnULKER, C AND DUPONT, J: .Treatment schedules modify circadian timing in human adrenocortical insufficiency. In: Chronobiology, L SClmVlNG, F HALnERG ANDJ PAULY,eds. Igaku Shoin Ltd., Tokyo, 1974, p 168 ASLANIAN,N L, ADAMIAN,K G, GRIGORIAN,S V, BAGDASSARIAN,R A AND ASSATRIAN,D G: Circadian rhythms of ECG T-wave, arterial pressure and heart rate in patients with ischemic heart disease. Chronobiologia 7:481, 1980 BIAGINI, A, CARPEGGIANI, C, MAZZEI, M C, TESTA, R, MICIIELASSI, C, ANTONELLI, R, L'ABBATE, A AND MASERS, A: Distribuzione oraria degli episodi di angina a riposo. Effetto della terapia medica. G Ital Cardiol 11:4, 1981 HALBERG, F, TONG, Y L AND JOIINSON, E A: Circadian system phase: An aspect of temporal morphology. Procedures and illustrativeexamples. In: The Cellular Aspects of Biorhythms. Syrup. on Rhythmic Res. 8th Int. Cong. Anat., H MAYERSBACII,ed, Springer-Verlag, Berlin, 1967, p 20 HALBERG, F, CARANDENTE, F, CORNELlSSEN, G, AND KATINAS, G S: Glossary of chronobiology. Chronobiologia, Suppl. 1, p 11, Ponte Ed., Milano, 1977 SMOLENSKY, M H, TATAR, S E, BERGMAN, S A, LOSMAN, J G, BARNARD, C N, DAcso, C C AND KRAF, I A: Circadian rhythmic aspects of human cardiovascular function; A review by chronobiological statistical methods. Chronobiologia 3:337, 1976 BERGES, D: Untersuchungen uber das ausmass von tagesschvankungen in E K G gesunder versuchs-personen. Z Kreisl Forsch 54:35, 1965
J. ELECTROCARDIOLOGY 16 (4), 1983