Heart rate variability and renal failure
Research Paper
Clinical Autonomic Research 1, 131-133 (1991)
HEART rate variability was measured_ from 24-h electrocardiograms in 61 patients with end stage chronic renal failure. The method used counts the number of times successive RR intervals differ by more than 50 ms over the 24-h period, and is a reliable indicator of cardiac parasympathetic activity. Also analysed were the frequency and type of ectopic beats and other arrhythmias. Twentyone subjects (34%) had varying numbers of ventricular ectopic beats, and twelve (20%) had frequent supraventricular ectopics. Total 24-h count values were abnormal in 30 (76%) of the 41 subjects whose tapes were technically suitable for this analysis. There were no sex differences, but those patients maintained on haemodialysis had significantly lower counts than those on continuous ambulatory peritoneal dialysis. We conclude that about three-quarters of patients with chronic renal failure have abnormal cardiac parasympathetic activity. This may increase susceptibility to cardiac arrhythmias and sudden death and contribute to the high mortality of patients with chronic renal failure.
Heart rate variability and cardiac arrhythmias in patients w i t h chronic renal failure B. J. T h o m s o n 2, MB MRCP, O. M c A r e a v e y 1, MB MRCP, J. M . M . Neilson z, PhD, R. J. W i n n e y a, MB FRCP and D. J. E w i n g 1"cA, MD FRCP 1Departments of Medicine, and 2Medical Physics and Medical Engineering, and the 3Medical Renal Unit, Royal Infirmary, Edinburgh EH3 9YW, UK.
CACorresponding Author
Key words: Chronicrenal failure,Autonomicnervous system, Parasympathetic, Heart rate, Heart rate variability
Introduction Autonomic dysfunction is well recognized in patients with chronic renal failure, although symptoms are often rather vague and non-specific. 1 A variety of methods has been used to define abnormalities, usually using one or more noninvasive cardiovascular reflex tests, and there is general agreement that abnormalities can be detected in up to half the patients investigated. 1'2 While the duration of renal failure and dialysis do not appear to influence results, there are contradictory reports of benefits with different forms of dialysis. 1'2 We have previously described a more sensitive measure of cardiac parasympathetic function, using 24-h electrocardiogram (ECG) tape recording and calculating heart rate variability. >s The aim of this study, therefore, was to apply this new method to patients with chronic renal failure to assess the extent of cardiac parasympathetic involvement. At the same time we have documented the prevalence of cardiac arrhythmias in this group. 6
Patients Sixty-nine patients with chronic renal failure w e r e selected for 24-h ECG tape recording. Subjects with diabetes were not included. Recordings in eight patients were technically unsatisfactory, leaving 61 patients (26 male, 35 female) aged 18-73 years, for study. Sixteen subjects (26%) were being treated with continuous ambulatory peritoneal dialysis (CAPD), while the remainder, 45 © Rapid Communications of Oxford Ltd.
subjects (74%), were on regular haemodialysis (HD). The various causes of renal failure and their distribution were similar to published surveys. Twenty-two patients had glomerulonephritis; nine had pyelonephritis; five had severe hypertension; four had polycystic kidneys; two had renal amyloidosis; seven had miscellaneous causes including analgesic nephropathy, hypokalaemic damage, and renal tuberculosis; the remaining seven patients had chronic renal failure of unknown aetiology. The patients were maintained on a variety of drugs, including 15 on beta-adrenergic blocking drugs, but none known to affect cardiac parasympathetic pathways. Length of dialysis treatment had been from 1 month to 21 (median 2) years (see Fig. 2). During the recording period each subject noted the time of going to bed and getting up. Of the 61 patients, 15 (nine male, six female) had supraventricular ectopics or non-sinus rhythm that rendered the tapes unsuitable for heart rate variability analysis. The remaining 46 subjects (17 male, 29 female) with a mean age of 49 (range 18-65) years had recordings that were able to be fully analysed.
Methods In each patient an ambulatory ECG was recorded over a 24-h period using a miniature portable tape recording device (Tracker, Reynolds Medical Ltd, Hertford, England), while patients underwent their normal daily routine. The HD group were recorded at variable times in relation to their regular haemodialysis sessions. The tapes were later Clinical Autonomic Research.vol 1 • 1991
131
B. J. Thomson et al. The results are presented as geometric mean and range for counts, and arithmetic mean and standard deviation for heart rate.
I00000-
10000O
'~
$
•
•
Oo
Results
looo-
o• ° ,oo
o
o
~_
02" 0
0 10.
0 ,
20
o~^
0
0
0 ~
30
,
~0
0 , 0
50
el
60
,
70
Age (y¢ars)
FIG. 1. Individual total 24-h RR counts in 46 subjects with chronic renal failure plotted against age. The lines indicate the 95% tolerance limits for normal subjects (O, males; O, females).
replayed and analysed at 120 times real speed using a Pathfinder arrhythmia analyser (Reynolds Medical Ltd). The methodology has been described in detail elsewhere.>s In brief, the computer measures each RR interval, and calculates successive RR interval differences. Where these exceed 50 ms, a 'count' is generated. Counts are then accumulated over the whole 24-h period, and the results presented as 'total 24-h counts', mean hourly waking and mean hourly sleeping counts; and mean waking and mean sleeping heart rates. The method is able to deal with even large numbers of ventricular ectopic beats/ We have previously published in detail count values for normal subjects. These are based on results from 57 normal subjects of varying ages, for which age related normal tolerance limits have been derived, s'v
Statistical Analysis As the count values are not normally distributed, logarithmic transformation of the count results was undertaken. Comparisons were then made using unpaired t-tests, and linear regression calculations.
Heart rate variability: Figure 1 shows individual total 24-h counts for the chronic renal failure patients plotted against age, and comparison with our previously published 95% normal tolerance limits. Only eleven subjects (24%) had values within the normal range. Table I shows the group mean values for counts, with waking and sleeping counts showing the same reduced numbers as the total counts. Mean heart rate was significantly higher in females; this has been reported in other studies of normal subjects. There were no significant differences in counts between males and females. There were, however, slight, but significantly reduced counts in subjects on haemodialysis compared with CAPD, but no significant differences between the HD and CAPD groups in length of sleep during the 24-h recordings. Prior duration of dialysis was similarly distributed in the two groups. There was no significant correlation between duration of dialysis and the count values (Fig. 2). The two patients with amyloidosis had 24-h counts within the normal range. Ectopic beats and arrhythmias: Of the 61 patients with technically satisfactory tapes, 25 subjects (41%) had normal sinus rhythm with no ectopic beats. Another 21 subjects (34%) had varying numbers of ventricular ectopic beats, of whom eleven (18%) had only occasional extrasystoles (less than ten per hour), five (8%) had more frequent ectopics, but only one had more than 100 per hour, and a further five (8%) had occasional bursts of ectopy, usually consisting of showers of individual extrasystoles over periods of up to 2 h, with no ectopics for the
Table 1. 24-h counts and mean heart rate in 46 patients with and type of dialysis (arithmetic mean 4- SD or geometric mean and range) Age
chronic
renal
failure,
Counts
divided
according
to
sex
Heart rate (beats/min)
Total 24 h
Mean hourly waking
Mean hourly sleeping
Waking
Heart rate
445 (9-7822) 215 (5-5082) NS
17 (1-369) 10 (1-166) NS
14 (1-441) 9 (1-289) NS
77 _+ 10
71 +_ 10
87 4- 13 p = 0.005
79 _+ 13 p = 0.022
638 (156-7822) 217 (5-5082) p = 0.005
19 (2-369) 8 (1-343) NS
26 (4-441) 7 (1-289) p = 0.024
80 4- 8
74 4- 12
85 ___14
76 4- 13
NS
NS
Sex
Males (n = 19)
51 _4-11
Females (n = 29)
48 _+ 12
Significance
NS
Type of dialysis CAPD (n = 11 )
52 _-t-11
HD = ( n = 35) Significance 132
48 +_ 11 NS
Clinical Autonomic Research.vol 1 - 1991
Heart rate variability and rena]failure
%O0OO]
•
o
O--~21yrs o
~ lOOO-~
o
8
• .
$
1
¢"
-I
2
3
~
5
5
7 8
9
DURATION OF DIALYSIS[ YEARS)
FIG. 2. Individual total 24-h RR counts in 42 subjects (in the other four the duration of dialysis was not known) with chronic renal failure plotted against duration of dialysis (O, patients on haemodialysis, O, patients on CAPD).
remainder of the tape. Of the 15 subjects (25%) excluded because of non-sinus rhythm, twelve had frequent superventricular ectopic beats, two had atrial fibrillation, and one had bigeminy.
Discussion We have shown that about three-quarters of patients with chronic renal failure have abnormal cardiac parasympathetic activity. This is higher than most previous studies examining cardiovascular reflexes in such patients, 1'2 and probably reflects the increased sensitivity of this 24-h heart rate variability method. In diabetic patients we have also shown the method to be more sensitive in detecting abnormalities than conventional cardiovascular autonomic' function tests. 3,s We have previously demonstrated that the 'counts' method used here is both reliable as a marker of cardiac parasympathetic activity, and reproducible in normal, diabetic and cardiac subjects. 3-5'v While we noted no differences between males and females in their mean counts, there was less heart rate variability in those maintained on haemodialysis, perhaps reflecting more severe renal failure and greater autonomic damage. This did not appear to relate to length of sleep, nor to previous duration of dialysis. As the subjects on H D were taped at varying times in relation to their concurrent dialysis sessions, we were unable to establish a formal relationship, or lack of it, between counts on and off dialysis. Visual inspection of the data, however, suggested that the dialysis itself had little influence on the counts. While it is always possible that the drugs the patients were taking could have affected our results, we consider this unlikely as we have previously shown that beta-blockers do not affect the counts. 3 The other drugs are not known to affect cardiac parasympathetic pathways.
just over one-third had ventricular ectopic beats, but very few of these were frequent. Additionally a further 20% of our subjects had supraventricular arrhythmias. This contrasts with a previous more detailed study in renal failure where 76% had some degree of ventricular ectopy defined as at least two ectopic beats per hour on average. 6 It is difficult to compare the two studies as the extent of underlying coronary artery disease was not defined in either study. Coronary artery disease is, however, well documented in chronic renal failure, and may account for a number of deaths, some of which are known to be sudden. 6 Whether low heart rate variability, ventricular arrhythmias and sudden deaths in chronic renal failure can be related together is as yet unclear. It is well established that this group of patients is vulnerable to the risk of sudden death perhaps associated with electrolyte disturbances, and the relatively large numbers with some degree of ventricular arrhythmia suggests that some of the deaths could be due to arrhythmias. It has been shown that, using a different method to measure heart rate variability, patients with low heart rate variability have an increased mortality in the months following a myocardial infarction. 8 The suggested mechanism is that this indicates lower cardiac parasympathetic activity, which in turn may be coupled with increased sympathetic activity, or at least some undefined alteration in autonomic 'balance'. Such an imbalance may render the heart more susceptible to potentially fatal arrhythmias, as has been demonstrated in dogs. 9 The high proportion of our subjects with apparently reduced cardiac parasympathetic activity suggests that this could be part of the explanation of the high mortality of patients with chronic renal failure.
References 1. Vita G, Messina C, Savica V, Bellinghieri G. Uraemic autonomic neuropathy. JAul Nerv Sysl 1990; 30: S179-S184. 2. Malik S, Winney R J, Ewing DJ. Chronic renal failure and cardiovascular autonomic function. Nepbron 1986; 43: 191-195. 3. Ewing D J, Neilson JMM, Travis P. New method for assessing cardiac parasympathetic activity using 24 hour electrocardiograms. Br Heart J 1984; 82:396 402. 4. Zuanetti G, Ladni R, Neilson JMM, Schwartz P, Ewing DJ. Heart rate variability in patients with ventricular arrhythmias: effect of antiarrhythmic drugs. J Am Call Cardiol 1991; 17: 604-612. 5. Ewing D J, Neilson JMM, Shapiro CM, Stewart JA, Reid W. 24 hour heart rate variability: effects of posture, sleep and time of day in normal subjects and comparison with 'bedside' autonomic function tests in diabetic patients. Br Heart J 1991; 68: 239-244. 6. Gruppo Emodialise E Patologie Cardiovascolari. Muhicentre, cross-sectinnai study of ventricular arrhythmias in chronically haemodialysed patients. Lancet 1988; ii: 305-309. 7. McAreavey D, Neilson JMM, Ewing D J, Russell DC. Cardiac parasympathetic activity during the early hours of acute myocardial infarction. Br Heart J 1989; 62: 165-170. 8. Kleiger RE, Miller JP, Bigger JT, Moss AJ. Decreased heart rate variability and its association with increased mortality after acute myocardial infarction. A m J Cardiol 1987; ~i9: 256-262. 9. Schwartz PJ. Sympathetic imbalance and cardiac arrhythmias. In: Randall WC, ed. Nemous Control of Cardioua;cular Function. Oxford: Oxford University Press, 1984; 225-252.
Received 19 January 1991 ; accepted with revision 24 March 1991. Clinical Autonomic Research. vol 1 • 1991
133
Research Paper
Clinical Autonomic Research 1, 131-133 (1991)
HEART rate variability was measured_ from 24-h electrocardiograms in 61 patients with end stage chronic renal failure. The method used counts the number of times successive RR intervals differ by more than 50 ms over the 24-h period, and is a reliable indicator of cardiac parasympathetic activity. Also analysed were the frequency and type of ectopic beats and other arrhythmias. Twentyone subjects (34%) had varying numbers of ventricular ectopic beats, and twelve (20%) had frequent supraventricular ectopics. Total 24-h count values were abnormal in 30 (76%) of the 41 subjects whose tapes were technically suitable for this analysis. There were no sex differences, but those patients maintained on haemodialysis had significantly lower counts than those on continuous ambulatory peritoneal dialysis. We conclude that about three-quarters of patients with chronic renal failure have abnormal cardiac parasympathetic activity. This may increase susceptibility to cardiac arrhythmias and sudden death and contribute to the high mortality of patients with chronic renal failure.
Heart rate variability and cardiac arrhythmias in patients w i t h chronic renal failure B. J. T h o m s o n 2, MB MRCP, O. M c A r e a v e y 1, MB MRCP, J. M . M . Neilson z, PhD, R. J. W i n n e y a, MB FRCP and D. J. E w i n g 1"cA, MD FRCP 1Departments of Medicine, and 2Medical Physics and Medical Engineering, and the 3Medical Renal Unit, Royal Infirmary, Edinburgh EH3 9YW, UK.
CACorresponding Author
Key words: Chronicrenal failure,Autonomicnervous system, Parasympathetic, Heart rate, Heart rate variability
Introduction Autonomic dysfunction is well recognized in patients with chronic renal failure, although symptoms are often rather vague and non-specific. 1 A variety of methods has been used to define abnormalities, usually using one or more noninvasive cardiovascular reflex tests, and there is general agreement that abnormalities can be detected in up to half the patients investigated. 1'2 While the duration of renal failure and dialysis do not appear to influence results, there are contradictory reports of benefits with different forms of dialysis. 1'2 We have previously described a more sensitive measure of cardiac parasympathetic function, using 24-h electrocardiogram (ECG) tape recording and calculating heart rate variability. >s The aim of this study, therefore, was to apply this new method to patients with chronic renal failure to assess the extent of cardiac parasympathetic involvement. At the same time we have documented the prevalence of cardiac arrhythmias in this group. 6
Patients Sixty-nine patients with chronic renal failure w e r e selected for 24-h ECG tape recording. Subjects with diabetes were not included. Recordings in eight patients were technically unsatisfactory, leaving 61 patients (26 male, 35 female) aged 18-73 years, for study. Sixteen subjects (26%) were being treated with continuous ambulatory peritoneal dialysis (CAPD), while the remainder, 45 © Rapid Communications of Oxford Ltd.
subjects (74%), were on regular haemodialysis (HD). The various causes of renal failure and their distribution were similar to published surveys. Twenty-two patients had glomerulonephritis; nine had pyelonephritis; five had severe hypertension; four had polycystic kidneys; two had renal amyloidosis; seven had miscellaneous causes including analgesic nephropathy, hypokalaemic damage, and renal tuberculosis; the remaining seven patients had chronic renal failure of unknown aetiology. The patients were maintained on a variety of drugs, including 15 on beta-adrenergic blocking drugs, but none known to affect cardiac parasympathetic pathways. Length of dialysis treatment had been from 1 month to 21 (median 2) years (see Fig. 2). During the recording period each subject noted the time of going to bed and getting up. Of the 61 patients, 15 (nine male, six female) had supraventricular ectopics or non-sinus rhythm that rendered the tapes unsuitable for heart rate variability analysis. The remaining 46 subjects (17 male, 29 female) with a mean age of 49 (range 18-65) years had recordings that were able to be fully analysed.
Methods In each patient an ambulatory ECG was recorded over a 24-h period using a miniature portable tape recording device (Tracker, Reynolds Medical Ltd, Hertford, England), while patients underwent their normal daily routine. The HD group were recorded at variable times in relation to their regular haemodialysis sessions. The tapes were later Clinical Autonomic Research.vol 1 • 1991
131
B. J. Thomson et al. The results are presented as geometric mean and range for counts, and arithmetic mean and standard deviation for heart rate.
I00000-
10000O
'~
$
•
•
Oo
Results
looo-
o• ° ,oo
o
o
~_
02" 0
0 10.
0 ,
20
o~^
0
0
0 ~
30
,
~0
0 , 0
50
el
60
,
70
Age (y¢ars)
FIG. 1. Individual total 24-h RR counts in 46 subjects with chronic renal failure plotted against age. The lines indicate the 95% tolerance limits for normal subjects (O, males; O, females).
replayed and analysed at 120 times real speed using a Pathfinder arrhythmia analyser (Reynolds Medical Ltd). The methodology has been described in detail elsewhere.>s In brief, the computer measures each RR interval, and calculates successive RR interval differences. Where these exceed 50 ms, a 'count' is generated. Counts are then accumulated over the whole 24-h period, and the results presented as 'total 24-h counts', mean hourly waking and mean hourly sleeping counts; and mean waking and mean sleeping heart rates. The method is able to deal with even large numbers of ventricular ectopic beats/ We have previously published in detail count values for normal subjects. These are based on results from 57 normal subjects of varying ages, for which age related normal tolerance limits have been derived, s'v
Statistical Analysis As the count values are not normally distributed, logarithmic transformation of the count results was undertaken. Comparisons were then made using unpaired t-tests, and linear regression calculations.
Heart rate variability: Figure 1 shows individual total 24-h counts for the chronic renal failure patients plotted against age, and comparison with our previously published 95% normal tolerance limits. Only eleven subjects (24%) had values within the normal range. Table I shows the group mean values for counts, with waking and sleeping counts showing the same reduced numbers as the total counts. Mean heart rate was significantly higher in females; this has been reported in other studies of normal subjects. There were no significant differences in counts between males and females. There were, however, slight, but significantly reduced counts in subjects on haemodialysis compared with CAPD, but no significant differences between the HD and CAPD groups in length of sleep during the 24-h recordings. Prior duration of dialysis was similarly distributed in the two groups. There was no significant correlation between duration of dialysis and the count values (Fig. 2). The two patients with amyloidosis had 24-h counts within the normal range. Ectopic beats and arrhythmias: Of the 61 patients with technically satisfactory tapes, 25 subjects (41%) had normal sinus rhythm with no ectopic beats. Another 21 subjects (34%) had varying numbers of ventricular ectopic beats, of whom eleven (18%) had only occasional extrasystoles (less than ten per hour), five (8%) had more frequent ectopics, but only one had more than 100 per hour, and a further five (8%) had occasional bursts of ectopy, usually consisting of showers of individual extrasystoles over periods of up to 2 h, with no ectopics for the
Table 1. 24-h counts and mean heart rate in 46 patients with and type of dialysis (arithmetic mean 4- SD or geometric mean and range) Age
chronic
renal
failure,
Counts
divided
according
to
sex
Heart rate (beats/min)
Total 24 h
Mean hourly waking
Mean hourly sleeping
Waking
Heart rate
445 (9-7822) 215 (5-5082) NS
17 (1-369) 10 (1-166) NS
14 (1-441) 9 (1-289) NS
77 _+ 10
71 +_ 10
87 4- 13 p = 0.005
79 _+ 13 p = 0.022
638 (156-7822) 217 (5-5082) p = 0.005
19 (2-369) 8 (1-343) NS
26 (4-441) 7 (1-289) p = 0.024
80 4- 8
74 4- 12
85 ___14
76 4- 13
NS
NS
Sex
Males (n = 19)
51 _4-11
Females (n = 29)
48 _+ 12
Significance
NS
Type of dialysis CAPD (n = 11 )
52 _-t-11
HD = ( n = 35) Significance 132
48 +_ 11 NS
Clinical Autonomic Research.vol 1 - 1991
Heart rate variability and rena]failure
%O0OO]
•
o
O--~21yrs o
~ lOOO-~
o
8
• .
$
1
¢"
-I
2
3
~
5
5
7 8
9
DURATION OF DIALYSIS[ YEARS)
FIG. 2. Individual total 24-h RR counts in 42 subjects (in the other four the duration of dialysis was not known) with chronic renal failure plotted against duration of dialysis (O, patients on haemodialysis, O, patients on CAPD).
remainder of the tape. Of the 15 subjects (25%) excluded because of non-sinus rhythm, twelve had frequent superventricular ectopic beats, two had atrial fibrillation, and one had bigeminy.
Discussion We have shown that about three-quarters of patients with chronic renal failure have abnormal cardiac parasympathetic activity. This is higher than most previous studies examining cardiovascular reflexes in such patients, 1'2 and probably reflects the increased sensitivity of this 24-h heart rate variability method. In diabetic patients we have also shown the method to be more sensitive in detecting abnormalities than conventional cardiovascular autonomic' function tests. 3,s We have previously demonstrated that the 'counts' method used here is both reliable as a marker of cardiac parasympathetic activity, and reproducible in normal, diabetic and cardiac subjects. 3-5'v While we noted no differences between males and females in their mean counts, there was less heart rate variability in those maintained on haemodialysis, perhaps reflecting more severe renal failure and greater autonomic damage. This did not appear to relate to length of sleep, nor to previous duration of dialysis. As the subjects on H D were taped at varying times in relation to their concurrent dialysis sessions, we were unable to establish a formal relationship, or lack of it, between counts on and off dialysis. Visual inspection of the data, however, suggested that the dialysis itself had little influence on the counts. While it is always possible that the drugs the patients were taking could have affected our results, we consider this unlikely as we have previously shown that beta-blockers do not affect the counts. 3 The other drugs are not known to affect cardiac parasympathetic pathways.
just over one-third had ventricular ectopic beats, but very few of these were frequent. Additionally a further 20% of our subjects had supraventricular arrhythmias. This contrasts with a previous more detailed study in renal failure where 76% had some degree of ventricular ectopy defined as at least two ectopic beats per hour on average. 6 It is difficult to compare the two studies as the extent of underlying coronary artery disease was not defined in either study. Coronary artery disease is, however, well documented in chronic renal failure, and may account for a number of deaths, some of which are known to be sudden. 6 Whether low heart rate variability, ventricular arrhythmias and sudden deaths in chronic renal failure can be related together is as yet unclear. It is well established that this group of patients is vulnerable to the risk of sudden death perhaps associated with electrolyte disturbances, and the relatively large numbers with some degree of ventricular arrhythmia suggests that some of the deaths could be due to arrhythmias. It has been shown that, using a different method to measure heart rate variability, patients with low heart rate variability have an increased mortality in the months following a myocardial infarction. 8 The suggested mechanism is that this indicates lower cardiac parasympathetic activity, which in turn may be coupled with increased sympathetic activity, or at least some undefined alteration in autonomic 'balance'. Such an imbalance may render the heart more susceptible to potentially fatal arrhythmias, as has been demonstrated in dogs. 9 The high proportion of our subjects with apparently reduced cardiac parasympathetic activity suggests that this could be part of the explanation of the high mortality of patients with chronic renal failure.
References 1. Vita G, Messina C, Savica V, Bellinghieri G. Uraemic autonomic neuropathy. JAul Nerv Sysl 1990; 30: S179-S184. 2. Malik S, Winney R J, Ewing DJ. Chronic renal failure and cardiovascular autonomic function. Nepbron 1986; 43: 191-195. 3. Ewing D J, Neilson JMM, Travis P. New method for assessing cardiac parasympathetic activity using 24 hour electrocardiograms. Br Heart J 1984; 82:396 402. 4. Zuanetti G, Ladni R, Neilson JMM, Schwartz P, Ewing DJ. Heart rate variability in patients with ventricular arrhythmias: effect of antiarrhythmic drugs. J Am Call Cardiol 1991; 17: 604-612. 5. Ewing D J, Neilson JMM, Shapiro CM, Stewart JA, Reid W. 24 hour heart rate variability: effects of posture, sleep and time of day in normal subjects and comparison with 'bedside' autonomic function tests in diabetic patients. Br Heart J 1991; 68: 239-244. 6. Gruppo Emodialise E Patologie Cardiovascolari. Muhicentre, cross-sectinnai study of ventricular arrhythmias in chronically haemodialysed patients. Lancet 1988; ii: 305-309. 7. McAreavey D, Neilson JMM, Ewing D J, Russell DC. Cardiac parasympathetic activity during the early hours of acute myocardial infarction. Br Heart J 1989; 62: 165-170. 8. Kleiger RE, Miller JP, Bigger JT, Moss AJ. Decreased heart rate variability and its association with increased mortality after acute myocardial infarction. A m J Cardiol 1987; ~i9: 256-262. 9. Schwartz PJ. Sympathetic imbalance and cardiac arrhythmias. In: Randall WC, ed. Nemous Control of Cardioua;cular Function. Oxford: Oxford University Press, 1984; 225-252.
Received 19 January 1991 ; accepted with revision 24 March 1991. Clinical Autonomic Research. vol 1 • 1991
133
