Heart Rate Variability during Sleep in Snorers with and without Obstructive Sleep Apnea* Luigi Ferini-Strambi, M.D.;t Marco Zucconi, M.D.; Alessandro Oldani, M.D.; and Salvatore Smirne, M.D.+
Changes in sympathetic and vagal tone may be the substrate for the development of cardiac arrhythmias in patients with obstructive sleep apnea (OSA). The cardiovascular responses in the traditional autonomic tests show great interindividual and intraindividual variations. During sleep there are repetitive modifications of heart rate (DR) that are not in8uenced by psychologic factors or the patient's cooperation. For this reason, we evaluated DR modifications in relation to spontaneous body movements (BM) and sleep apneas during nonrapid eye movement (NREM) and rapid eye movement (REM) sleep in habitual snorers with normal and pathologic respiratory disturbance index (RDI). From 132 consecutive patients referred to our sleep center for habitual snoring and/or daytime somnolence, we selected 35 male patients younger than 60 years without clinical evidence of autonomic dysfunction. They were divided into three groups: group A (RDI 10 and 20). No significant
difference was found among the three groups in the DR variability related to BM. In the evaluation of bradytachyarrhythmias related to apneic events of 20 to 30 s, we found a significant difference between group A and the other two groups. In patients with RDI> 10, a reduced DR variability related to a reduced sympathetic tone in the post-apnea phase was observed. Some authors suggested that an DR increase during the post-apnea period can he used as an index of "brainstem arousal." Our results seem to indicate a reduced apnea-related "arousability" in patients with RDI> 10. This finding might be one of the factors contributing to the worsening of OSA. (Chest 1992; 102:1023-27)
Cardiac arrhythmias noted in association with obstructive sleep apnea (OSA) syndrome l -3 probably cause the increased incidence of sudden death that is present in this illness. 4-6 In patients with OSA, usually the heart rate (HR) slows during the apnea and then increases markedly as the subject takes a breath. 7 In these patients with normal autonomic nervous function, atropine sulfate blocks the pattern by eliminating the bradycardia component, while 100 percent oxygen, even at high rates of administration, causes only moderate blunting ofHR change. 7 Therefore, electrocardiographic variations observed in OSA are mediated by the autonomic nervous system (ANS). Changes in sympathetic and vagal tone may be the necessary substrate for the development ofcardiac arrhythmias. R The cardiovascular responses utilized in the traditional autonomic tests (ie, Valsalva maneuver, respiratory sinus arrhythmia, handgrip test) show great interindividual and intraindividual variations. 9 A number of factors may contribute to this finding and they
are greatly influenced by psychologic factors, such as the attitude of the patient to the test. During sleep there are repetitive modifications of ANS that are very constant and not influenced by the patient's emotional status and the degree of cooperation. For this reason, we previously evaluated tonic (vagal activity) HR modifications in relation to the deepening of nonrapid eye movement (NREM) sleep, as well as phasic (sympathetic activity) HR modifications in relation to spontaneous body movements (BM) during NREM and rapid eye movement (REM) sleep, in healthy subjects and in patients with some neurologic diseases. 10,11 Our intention in this study was to apply the same methodology for evaluating cardiac autonomic function in snorers with and without significant OSA. In the same patients, we evaluated the HR changes associated with apneic events.
*From the Sleep Disorders Center, Department of Neurology, State University and Istituto Scientifico H San Raffaele, Milan, Italy. tVice-Chief, Department of Neurology. tProfessor of Neurology, and Chief, Department of Neurol()~ and Sleep Disorders Center. Manuscript received November 20, 1991; revision accepted February 20. Reprint requests: Dr. Ferini-Strambi, Department of Neurology, H San Raffaele, Vuz Primetti 27, Milan, Italy 20127
=
= =
ANS autonomic nervous system; 8M body movements; NREM non-rapid eye movement; OSA obstructive sleep apnea; Rbm = body movement ratio; RDI = respiratory disturbance index; REM rapid eye movement; Rs/w sleep/wakefulness ratio
=
=
=
METHODS AND MATERIALS
Patients Thirty-five patients were studied. They \\'ere selected froIll 132 consecutive subjects referred to our sleep center for habitual snorin~ and/or daytime somnolence. They met the f(}lIowin~ inclusion criteria: (1) male; (2) aged younger than 60 years; (3) no history of hypertension, cardiac disease, or pulmonary diseases; (4) normal electrocardiograms recorded during wakefulness; (5) normal results of pulmonary function tests; (6) no medication or alcohol; and (7) no clinical evidence of autonomic dysfunction (orthostatic hypotenCHEST I 102 I 4 I OCTOBER, 1992
1023
sion, intestinal motilit y ahnornlalities, bladder dysfunction, sexual impotence, and sweating ahnormalities).
20 s
Quiet wakefulness
R-R mean
Methods All patients had nocturnal polygraphic monitoring. The following variables were monitored: electroencephalogram (C3/A2-C4IAl), electro-oculogram, and chin and leg electromyograms. Thoracic and abdominal strain-gauges were used to record respiratory excursions. Nasal and oral thermistors were used to detect airflow. Oxygen saturation (Sa0 2) was continuously recorded with a calibrated ear oximeter (Ohmeda Biox II or III). Lead 2 of the surface ECG was continuously recorded on the polygraph.
Data Analysis Sleep was s(,'Ored following the criteria of Rechstschaffen and Kales. 12 Apneas and hypopneas were scored according to standard definitions based on respiratory, airflo~ and oximetric findings. 13 Respiratory disturbance index (RDI) was calculated based on the following formula: apnea + hypopnea/total sleep time (min) x 60. Tonic and phasic llR variability was evaluated in relation to spontaneous BM during sleep, according to the procedure and the method published elsewhere. lO . 1I Apnea- or hypopnea-related BM were obviously excluded. During 20 s of quiet wakefulness, we measured the mean R-R interval before sleep onset. During sleep, we measured the shortest R-R interval during the 20 s after the onset of BM and the longest and the mean R-R intervals between 30 and 10 s before BM (Fig 1). The 10 s immediately preceding BM were excluded as the BM-related tachycardia has been shown to commence 8 s before BM. 14 The following indices were calculated as average ofat least three measurements randomly selected, either during NREM or REM sleep: (I) the ratio of the mean R-R interval before BM to the mean R-R interval during wakefulness (sleep! wakefulness ratio = Rs/w); (2) the ratio of the longest R-R interval before BM to the shortest one after BM (body movement ratio=Rbm). Rs/w was considered an index of tonic HR decrease, induced by sleep (mainly vagal activity). Rbm was considered an index of phasic HR increase, induced by BM (mainly sympathetic activity). Moreover, we considered fIR changes induced by OSA (Fig 2). The following index was calculated as average of at least three measurements in relation to apneic events, randomly selected, but all between 20 and 30 s, during NREM or REM sleep: the ratio of the longest R-R interval during the apneic event to the shortest RR interval after the end of the apneic event (R apnea). Statistical analysis was perfoTlned by means of analysis of variance (ANOVA). RESULTS
On the basis of polysomnographic results, the patients were divided into three groups (Table 1): group A (12 subjects \vith RDI 20).
20 s
'0 s
1excluded :
sleep R-R mean and the longest R-R (R-R max) R-R mean during sleep R s/w= --------------------------------R-R mean during wakefulness
body movement
;.
20 s
the shortest R-R (R-R min)
R-R max R bm= ------------R-R min
FIGURE 1. Parameters measured during quiet wakefulness and in relation to spontaneous body movements during sleep.
None of the patients showed periodic leg movements during sleep. The Rs/w during NREM sleep was similar in the three groups: group A = 1.09 ± 0.13 (mean±SD); group B=I.07±0.08; and group C= 1.14±0.08. The Rs/w during REM sleep was as follows: group A=I.11±0.16; group B=I.10±0.12; and group C= 1.16±0.09. Here too, no significant difference was found among the three groups. The values of Rbm in NREM resulted in the following: group A = 1.43 ± 0.12; group B = 1.44 ± 0.12; and group C=I.41±0.14. Rbm in REM sleep was as follows: group A= 1.48±0.13; group B= 1.51±0.18; and group C=I.51±0.17. No significant difference was found in Rbm during NREM and REM sleep among the three groups. The values of Rs/w and Rbm in the three groups were not different from those of age-matched controls (15 subjects, mean age=49.1±5.3 years; Rs/w in NREM sleep = 1.15±0.13, in REM sleep = 1.17± 0.14; Rbm in NREM sleep = 1.46±0.11, in REM sleep = 1.53±0.13). As reported in the "Methods" section, we also considered HR changes induced by obstructive apneic events. Apneas were all between 20 and 30 s. No significant difference was found among the mean lowest Sa0 2 during the selected apnea (group A = 82.4 ± 6.1; group B = 79.6 ± 4.9; group C=77.9±8.5). Sleep body position was verified visually in each patient by a technologist through a lowlight camera. In all the selected apneas, the patient was in the supine position. The index calculated in relation to apneic events (R apnea) during NREM sleep resulted in the following
Table 1- Patient Characteristics* Group
n
age, yr
Height, cm
Weight, kg
RDI
Mean Lowest Sa02 during Apnea
A: RDI20
12 10 13
5O.8±5.1 49.5±4.2 48.0±8.5
170.75± 10.75 168.75± 14 167.25± 12
94.095 ± 12.735 96.39 ± 14.13 102.915± 17.91
7.3±1.6t 16.2±2.5 39.4± 13.7
85.7±5.8+ 79.9±6.3 75.1 ± 12.2
*Values are expressed as mean ± SD. RDI = respiratory disturbance index. tp
Changes in sympathetic and vagal tone may be the substrate for the development of cardiac arrhythmias in patients with obstructive sleep apnea (OSA). The cardiovascular responses in the traditional autonomic tests show great interindividual and intraindividual variations. During sleep there are repetitive modifications of heart rate (DR) that are not in8uenced by psychologic factors or the patient's cooperation. For this reason, we evaluated DR modifications in relation to spontaneous body movements (BM) and sleep apneas during nonrapid eye movement (NREM) and rapid eye movement (REM) sleep in habitual snorers with normal and pathologic respiratory disturbance index (RDI). From 132 consecutive patients referred to our sleep center for habitual snoring and/or daytime somnolence, we selected 35 male patients younger than 60 years without clinical evidence of autonomic dysfunction. They were divided into three groups: group A (RDI 10 and 20). No significant
difference was found among the three groups in the DR variability related to BM. In the evaluation of bradytachyarrhythmias related to apneic events of 20 to 30 s, we found a significant difference between group A and the other two groups. In patients with RDI> 10, a reduced DR variability related to a reduced sympathetic tone in the post-apnea phase was observed. Some authors suggested that an DR increase during the post-apnea period can he used as an index of "brainstem arousal." Our results seem to indicate a reduced apnea-related "arousability" in patients with RDI> 10. This finding might be one of the factors contributing to the worsening of OSA. (Chest 1992; 102:1023-27)
Cardiac arrhythmias noted in association with obstructive sleep apnea (OSA) syndrome l -3 probably cause the increased incidence of sudden death that is present in this illness. 4-6 In patients with OSA, usually the heart rate (HR) slows during the apnea and then increases markedly as the subject takes a breath. 7 In these patients with normal autonomic nervous function, atropine sulfate blocks the pattern by eliminating the bradycardia component, while 100 percent oxygen, even at high rates of administration, causes only moderate blunting ofHR change. 7 Therefore, electrocardiographic variations observed in OSA are mediated by the autonomic nervous system (ANS). Changes in sympathetic and vagal tone may be the necessary substrate for the development ofcardiac arrhythmias. R The cardiovascular responses utilized in the traditional autonomic tests (ie, Valsalva maneuver, respiratory sinus arrhythmia, handgrip test) show great interindividual and intraindividual variations. 9 A number of factors may contribute to this finding and they
are greatly influenced by psychologic factors, such as the attitude of the patient to the test. During sleep there are repetitive modifications of ANS that are very constant and not influenced by the patient's emotional status and the degree of cooperation. For this reason, we previously evaluated tonic (vagal activity) HR modifications in relation to the deepening of nonrapid eye movement (NREM) sleep, as well as phasic (sympathetic activity) HR modifications in relation to spontaneous body movements (BM) during NREM and rapid eye movement (REM) sleep, in healthy subjects and in patients with some neurologic diseases. 10,11 Our intention in this study was to apply the same methodology for evaluating cardiac autonomic function in snorers with and without significant OSA. In the same patients, we evaluated the HR changes associated with apneic events.
*From the Sleep Disorders Center, Department of Neurology, State University and Istituto Scientifico H San Raffaele, Milan, Italy. tVice-Chief, Department of Neurology. tProfessor of Neurology, and Chief, Department of Neurol()~ and Sleep Disorders Center. Manuscript received November 20, 1991; revision accepted February 20. Reprint requests: Dr. Ferini-Strambi, Department of Neurology, H San Raffaele, Vuz Primetti 27, Milan, Italy 20127
=
= =
ANS autonomic nervous system; 8M body movements; NREM non-rapid eye movement; OSA obstructive sleep apnea; Rbm = body movement ratio; RDI = respiratory disturbance index; REM rapid eye movement; Rs/w sleep/wakefulness ratio
=
=
=
METHODS AND MATERIALS
Patients Thirty-five patients were studied. They \\'ere selected froIll 132 consecutive subjects referred to our sleep center for habitual snorin~ and/or daytime somnolence. They met the f(}lIowin~ inclusion criteria: (1) male; (2) aged younger than 60 years; (3) no history of hypertension, cardiac disease, or pulmonary diseases; (4) normal electrocardiograms recorded during wakefulness; (5) normal results of pulmonary function tests; (6) no medication or alcohol; and (7) no clinical evidence of autonomic dysfunction (orthostatic hypotenCHEST I 102 I 4 I OCTOBER, 1992
1023
sion, intestinal motilit y ahnornlalities, bladder dysfunction, sexual impotence, and sweating ahnormalities).
20 s
Quiet wakefulness
R-R mean
Methods All patients had nocturnal polygraphic monitoring. The following variables were monitored: electroencephalogram (C3/A2-C4IAl), electro-oculogram, and chin and leg electromyograms. Thoracic and abdominal strain-gauges were used to record respiratory excursions. Nasal and oral thermistors were used to detect airflow. Oxygen saturation (Sa0 2) was continuously recorded with a calibrated ear oximeter (Ohmeda Biox II or III). Lead 2 of the surface ECG was continuously recorded on the polygraph.
Data Analysis Sleep was s(,'Ored following the criteria of Rechstschaffen and Kales. 12 Apneas and hypopneas were scored according to standard definitions based on respiratory, airflo~ and oximetric findings. 13 Respiratory disturbance index (RDI) was calculated based on the following formula: apnea + hypopnea/total sleep time (min) x 60. Tonic and phasic llR variability was evaluated in relation to spontaneous BM during sleep, according to the procedure and the method published elsewhere. lO . 1I Apnea- or hypopnea-related BM were obviously excluded. During 20 s of quiet wakefulness, we measured the mean R-R interval before sleep onset. During sleep, we measured the shortest R-R interval during the 20 s after the onset of BM and the longest and the mean R-R intervals between 30 and 10 s before BM (Fig 1). The 10 s immediately preceding BM were excluded as the BM-related tachycardia has been shown to commence 8 s before BM. 14 The following indices were calculated as average ofat least three measurements randomly selected, either during NREM or REM sleep: (I) the ratio of the mean R-R interval before BM to the mean R-R interval during wakefulness (sleep! wakefulness ratio = Rs/w); (2) the ratio of the longest R-R interval before BM to the shortest one after BM (body movement ratio=Rbm). Rs/w was considered an index of tonic HR decrease, induced by sleep (mainly vagal activity). Rbm was considered an index of phasic HR increase, induced by BM (mainly sympathetic activity). Moreover, we considered fIR changes induced by OSA (Fig 2). The following index was calculated as average of at least three measurements in relation to apneic events, randomly selected, but all between 20 and 30 s, during NREM or REM sleep: the ratio of the longest R-R interval during the apneic event to the shortest RR interval after the end of the apneic event (R apnea). Statistical analysis was perfoTlned by means of analysis of variance (ANOVA). RESULTS
On the basis of polysomnographic results, the patients were divided into three groups (Table 1): group A (12 subjects \vith RDI 20).
20 s
'0 s
1excluded :
sleep R-R mean and the longest R-R (R-R max) R-R mean during sleep R s/w= --------------------------------R-R mean during wakefulness
body movement
;.
20 s
the shortest R-R (R-R min)
R-R max R bm= ------------R-R min
FIGURE 1. Parameters measured during quiet wakefulness and in relation to spontaneous body movements during sleep.
None of the patients showed periodic leg movements during sleep. The Rs/w during NREM sleep was similar in the three groups: group A = 1.09 ± 0.13 (mean±SD); group B=I.07±0.08; and group C= 1.14±0.08. The Rs/w during REM sleep was as follows: group A=I.11±0.16; group B=I.10±0.12; and group C= 1.16±0.09. Here too, no significant difference was found among the three groups. The values of Rbm in NREM resulted in the following: group A = 1.43 ± 0.12; group B = 1.44 ± 0.12; and group C=I.41±0.14. Rbm in REM sleep was as follows: group A= 1.48±0.13; group B= 1.51±0.18; and group C=I.51±0.17. No significant difference was found in Rbm during NREM and REM sleep among the three groups. The values of Rs/w and Rbm in the three groups were not different from those of age-matched controls (15 subjects, mean age=49.1±5.3 years; Rs/w in NREM sleep = 1.15±0.13, in REM sleep = 1.17± 0.14; Rbm in NREM sleep = 1.46±0.11, in REM sleep = 1.53±0.13). As reported in the "Methods" section, we also considered HR changes induced by obstructive apneic events. Apneas were all between 20 and 30 s. No significant difference was found among the mean lowest Sa0 2 during the selected apnea (group A = 82.4 ± 6.1; group B = 79.6 ± 4.9; group C=77.9±8.5). Sleep body position was verified visually in each patient by a technologist through a lowlight camera. In all the selected apneas, the patient was in the supine position. The index calculated in relation to apneic events (R apnea) during NREM sleep resulted in the following
Table 1- Patient Characteristics* Group
n
age, yr
Height, cm
Weight, kg
RDI
Mean Lowest Sa02 during Apnea
A: RDI20
12 10 13
5O.8±5.1 49.5±4.2 48.0±8.5
170.75± 10.75 168.75± 14 167.25± 12
94.095 ± 12.735 96.39 ± 14.13 102.915± 17.91
7.3±1.6t 16.2±2.5 39.4± 13.7
85.7±5.8+ 79.9±6.3 75.1 ± 12.2
*Values are expressed as mean ± SD. RDI = respiratory disturbance index. tp
