Hypoxemia alone does not explain blood pressure elevations after obstructive apneas JACK RINGLER, ROBERT C. BASNER, RICHARD SHANNON, RICHARD SCHWARTZSTEIN, HAROLD MANNING, STEVEN E. WEINBERGER, AND J. WOODROW WEISS Charles A. Dana Institute and Harvard-Thorndike Laboratory of Beth Israel Hospital, Beth Israel Hospital Sleep Disorders Center, and Departments of Medicine, Beth Israel Hospital and Harvard Medical School, Boston, Mussachusetts 02115
RINGLER,JACK, ROBERT C. BASNER,RICHARD SHANNON, RICHARD SCHWARTZSTEIN,HAROLD MANNING, STEVEN E. WEINBERGER,ANDJ. W~~~~~~W~~~~.Hypoxemiaalonedoes not explain blood pressure elevations after obstructive apneas. J. Appl. Physiol. 69(6): 2143-2148, 1990.-In patients with obstructive sleep apnea (OSA), substantial elevations of systemic blood pressure (BP) and depressions of oxyhemoglobin saturation (Sao,) accompany apnea termination. The causes of the BP elevations, which contribute significantly to nocturnal hypertension in OSA, have not been defined precisely. To assess the relative contribution of arterial hypoxemia, we observed mean arterial pressure (MAP) changes following obstructive apneas in 11 OSA patients during non-rapid-eye-movement (NREM) sleep and then under three experimental conditions: 1) apnea with 0, supplementation; 2) hypoxemia (Sao, 80%) without apnea; and 3) arousal from sleep with neither hypoxemia nor apnea. We found that apneas recorded during O2 supplementation (Sao, nadir 93.6% t 2.4; mean t SD) in six subjects were associated with equivalent postapneic MAP elevations compared with unsupplemented apneas ( Sao.,nadir 7982%): 18.8 -t- 7.1 vs. 21.3 t 9.2 mmHg (mean change MAP k SD); in the absence of respiratory and sleep disruption in eight subjects, hypoxemia was not associated with the BP elevations observed following apneas: -5.4 t 19 vs. 19.1 rfi 7.8 mmHg (P c 0.01); and in five subjects, auditory arousal alone was associated with MAP elevation similar to that observed following apneas: 24.0 t 8.1 vs. 22.0 k 6.9 mmHg. We conclude that in NREM sleep postapneic BP elevations are not primarily attributable to arterial hypoxemia. Other factors associated with apnea termination, including arousal from sleep, reinflation of the lungs, and changes of intrathoracic pressure, may be responsible for these elevations. obstructive sleep apnea; hemodynamics; arousal; human; nonrapid-eye-movement sleep OBSTRUCTIVE APNEAS during sleep are associated with oscillations of systemic blood pressure (BP) that occur in phase with disruptions of respiration and sleep. Substantial BP elevations coincide with termination of apneas. Even in patients who are normotensive during the day, postapneic systolic BP can exceed 200 mmHg (20). Several factors have been reported to be responsible for the BP changes associated with obstructive sleep apnea (OSA). These include arterial hypoxemia, intrathoracic pressure changes, interruptions of ventilation, and disruption of sleep architecture (11, 13, 15-18, 20). The 0161-7567/90
$1.50
Copyright
relative contribution of each of these factors to the integrated BP response remains unclear, however. By analyzing apneas associated with similar degrees of hypoxemia and then by manipulating independently our subjects’ oxygenation, ventilation, and sleep architecture, we hoped to clarify whether hypoxemia has an independent role in effecting BP elevations following non-rapid-eye-movement (NREM) obstructive apneas. METHODS Subjects We enrolled 16 subjects (13 men, 3 women) with clinical features of OSA: 11 of them successfully completed the experimental protocol (see Data Analysis, below). Each was referred to the Beth Israel Hospital Sleep Disorders Center for diagnostic polysomnography (PSG). Informed consent was obtained in accordance with guidelines of the Committee on Clinical Investigations at Beth Israel Hospital. Vasoactive medications were withheld 48 h before the study night. Measurements After undergoing history and physical examination, subjects were prepared for overnight PSG. A standard PSG montage, including three channels of electroencephalogram (EEG), left and right electrooculograms, chin and intercostal electromyogram (EMG), electrocardiogram (ECG), nasal and oral airflow by thermistors, respiratory effort by inductance plethysmography belts (Respitrace, Ardsley, NY), and ear oximetry (Biox IIA, Biox Technology, Boulder, CO), was utilized (14). Systemic arterial pressure was monitored by means of a 20gauge radial artery catheter. All data were recorded on a ‘I$-channel EEG (Grass model 8-24D, Grass Instruments, Quincy, MA). Pro toe01 Each subject was studied on a single night between 1090 P.M. and 6:00 A.M. For the first 2-4 h of sleep we recorded BP continuously without intervention. During the remainder of the night we performed a series of experiments. Since the subjects were patients undergoing
0 1990 the American
Physiological
Society
2143
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 12, 2018.
2144
BLOOD
PRESSURE
ELEVATIONS
FOLLOWING
clinical PSG, the complete protocol could not be completed in each case. Experiment 1. In six subjects, we began by observing BP and heart rate (HR) changes during administration of supplemental O2 at a rate of 3-11 l/min via nasal cannula. We adjusted the flow rate to maintain oxyhemoglobin saturation (Sao,) > 90% during and after apneas. Experiment 2. In eight subjects, we assessed the BP and HR responses to hypoxemia without apnea. &neas were eliminated by means of nasal continuous positive airway pressure (CPAP; Sleepeasy III, Respironics, Monroeville, PA). Episodic hypoxemia was achieved by connecting the CPAP tubing to a second CPAP compressor equipped with a balloon reservoir for hypoxic gas (8-12% 02, 88-92% N2). We manually adjusted the composition of the gas in the reservoir to allow Sao, to fall to 80% over a period of L-4 min. Experiment 3. We aroused from sleep five subjects undergoing nasal CPAP therapy by providing intermittent auditory stimulation with a portable audiometer (model MA-20, Maico Hearing Instruments, Minneapolis, MN). This allowed us to examine the BP response to arousal accompanied by neither apnea nor hypoxemia. We defined arousal polysomnographically as an increase in EMG amplitude accompanied by a change in pattern on at least one additional channel (14). We adjusted the amplitude and frequency of the auditory stimulus to induce PSG changes qualitatively similar to those associated with apneas. Data Analysis
Data obtained from a given subject were included in the analysis if the following criteria were met: 1) apneahypopnea index > 30 events/h; 2) at least six NREM apneas with associated Sao, nadir of 79-82%; 3) at least six NREM apneas recorded during O2 therapy associated with Sao, nadir > 90% or at least one episode of experimentally induced hypoxemia without apnea ending with Sao, 5 80% without change of sleep stage. Data from 11 of the 16 enrolled subjects were included in the analysis. Clinical characteristics of these subjects are displayed in Table 1. Of the five who did not complete 1. Subject characteristics
TABLE
Sao,% y-u-
sex
A;;;
;;>
fg-)s
sn
*
I 2 3 4 5 6 7 8 9 10 II
MAP9 mm&
F M F F M M M M M M M
24 41 29 45 71 38 47 35 38 46 41
127 125 72 123 70 150 173 118 120 97 102
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
114 108 90 97 103 93 129 107 117 84 82
AI-II, events/h
203 61 106 103 47 57 119 118 120 105 78
Baseline
Nfean lowest
Lowest
97
64 83 88 60 80
56 57
97 95 97 97 93 94 98 97 97 96
EDS, excessive daytime hypersomnolence; Sn, snoring; line mean arterial pressure during quiet wakefulness; hypopnea index; Sao,%, percent oxyhemoglobin saturation.
80 88 83 84 72 81
79 51 70 67 71 79 71 71 77
MAP, baseAHI, apnea-
OBSTRUCTIVE
APNEAS
the protocol, four did not develop oxyhemoglobin desaturations below 83% during NREM sleep, thus failing to meet criterion 2 above; the fifth failed to meet criterion 3 by arousing repeatedly during attempts to induce hypoxemia through the nasal CPAP system. Sleep recordings were scored using standard criteria (14). We estimated baseline BP during quiet wakefulness and during NREM sleep on nasal CPAP therapy in each subject by averaging mean arterial pressure (MAP) values tabulated at 2-s intervals from a typical 30-s epoch of the PSG fulfilling the desired criteria. We considered MAP equivalent to one-third systolic BP plus two-thirds diastolic BP. Between 6 and 15 apneas associated with Sao, nadir of 79-82% were analyzed for each subject. If more than 15 apneas qualified, as was the case in six of the subjects, we randomly selected 15 for analysis. For each selected apnea, we tabulated MAP values over the three to six cardiac cycles coinciding with the Sao, nadir. These values were without exception within 10 mmHg of the peak postapneic MAP and included the peak in the majority of cases. For each subject, data from all analyzed apneas were combined, yielding an average MAP at the Sao, nadir of 79-82%. We then analyzed similarly the 08-supplemented apneas in the six subjects for experiment 1. Again, 6-15 apneas per subject were analyzed, and as observed in unsupplemented apneas Sao,, nadir correlated very closely with peak postapneic BP. By comparing MAP following apneas with nadirs of 79-82% to MAP following apneas with nadirs > 90% in each subject, we were able to examine the relationship between postapneic BP elevation and Sao, nadir. Episodes of hypoxemia without apnea were analyzed in eight subjects by tabulating MAP values over the first 10 cardiac cycles after Sao, reached 80%. Because of our concern for subject safety, hypoxemia was achieved more gradually during these episodes than during the naturally occurring apneas. Data from one to three hypoxemia trials in each subject were combined yielding an average MAP at Sao, of 80%. We were thus able to compare the average MAP associated with similar levels of Sao, in apneic and nonapneic conditions. In the five subjects for experiment 3, we assessed the effect on BP of auditory arousal accompanied by neither apnea nor hypoxemia by calculating average MAP over the first five cardiac cycles following the 3-s stimulus. Because of the constraints imposed by the clinical testing, the number of arousal trials was necessarily limited (3 trials in subj 2 and 10, 2 trials in subj 6 and 8, and 1 trial in subj 3). The responses were quite stereotypical, however (see RESULTS). We compiled HR data in the same manner as the BP analysis described above. HR in beats per minute was derived as 60/RR interval in seconds. For the statistical analysis, we employed two-tailed paired t tests and/or Wilcoxon’s signed rank tests, each with Bonferroni correction for multiple comparisons (24). We compared in this way MAP responses following naturally occurring apneas, 02-supplemented apneas, episodes of induced hypoxemia without apnea, and episodes of auditory arousal each to the NREM baseline MAP
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 12, 2018.
BLOOD
PRESSURE
ELEVATIONS
FOLLOWING
during nasal CPAP therapy. These calculations were performed at the Core Laboratory Computer Facilities of Beth Israel Hospital. MAP values referable to extrasystoles or to motion artifact were excluded from analysis in all phases of the study. RESULTS
Experiment 1. Postapneic BP Response with 02 Supplementation
Experiment 2. BP Response to Hypoxemia Without Apnea
EEG EOG
between MAP and Sao, to of apneas with nasal CPAP
WFF ’
140 F E E $
m
-~~’
-a
120loo809
g
60’
6
7
8
9
10
11
SUBJECT FIG. 2. Mean arterial pressure (MAP) following apneas associated with oxyhemoglobin saturation (Sao,) nadir 7942% (solid bars) and apneas associated with Sao, nadir 2 90% (hatched bars) in 6 subjects. NREM baseline MAP represented by open bars. For group, postapneic conditions both differed significantly from baseline but not from each other (see text). Subjects are identified by number (see Table 1). All data in Figs. 2-5 displayed as means t SD.
160
Both apneas with Sao, nadirs > 90% (mean 94 t 2.4%) and apneas with Sao, nadirs of 79-82% were associated with significant elevations of MAP compared with NREM baseline (Fig. 2). In the six subjects studied, we found the mean elevation over NREM baseline for 02supplemented apneas to be 18.8 t 7.1 mmHg (mean change in MAP t SD; P = 0.0026). For unsupplemented apneas, the mean elevation was 21.3 t 9.2 mmHg (P c 0.0001). We did not detect a significant difference between the two conditions (supplemented vs. unsupplemented) with regard to the magnitude of the postapneic BP elevation. Supplemental O2 was associated with I- to 10-s increases in mean apnea duration in subjects 6, 7, 9, 10, and I1 and with a decrease of 1 s in subject 8.
We found the relationship be different after elimination
2145
APNEAS
160~
=
Figure 1 displays the PSG recorded during a typical NREM apnea. All subjects had substantial and repetitive BP elevations in the immediate postapneic period. This period in all cases encompassed the lowest Sao, and the highest BP of the apnea-recovery cycle. BP responses during apneas were characterized by intersubject and intrasubject variability. For this reason all comparisons were made to the NREM baseline MAP.
OBSTRUCTIVE
I-
. 10 SECON;
FIG. 1. Polysomnographic and radial arterial pressure recording during and after a typical NREM obstructive apnea (su6j 8). Blood pressure (BP) and heart rate (HR) fall during most of apnea, and cyclic BP depressions coincide with inspiratory efforts toward end of event. Sharp rises of BP and HR coincide with arousal, resumption of ventilation, and oxyhemoglobin saturation nadir. EEG, electroencephalogram; EOG, left and right electrooculograms; EMG,S, submental electromyogram; EMG-I, intercostal electromyogram; V, oronasal airflow (thermistor signals); ABD, respiratory effort (by abdominal inductance plethysmography); AP, radial arterial pressure; Sao,, percent oxyhernoglobin saturation (by ear oximetry).
1 -
t40-
0, 2 E
120-
; d 2
loo80606 1
2
3
4
5
6
7
8
SUBJECT FIG. 3. Mean arterial pressure (MAP) after induction of hypoxemia [oxyhemoglobin saturation (Sao,) 80%] without apnea (hatched bars) in 8 subjects compared with NREM baseline (open bars) and with postapneic MAP at Sa 0, nadir 7942% (solid bars). In absence of respiratory and sleep disruption, hypoxemia was associated with no significant change in MAP compared with baseline (see text). Subjects are identified by number (see Table 1).
(Fig. 3). In the eight subjects studied, postapneic hypoxemia was associated with marked BP elevations as in experiment 1. The mean elevation for these subjects was 19.1 t 7.8 mmHg (P c 0.0001). By contrast, without respiratory and sleep disruption, the same degree of hypoxemia in the same eight subjects was not associated with substantial BP elevations. In this condition, the mean change was -5.4 t 18.9 mmHg, not significantly different from baseline. We calculated the difference in MAP response between the apneic and nonapneic conditions to be statistically significant (P < 0.01). This difference persists even if subject 1, in whom BP declined precipitously under the conditions of experiment 2, is treated as a statistical outlier. Experiment 3. BP Response to Arousal With Neither Apnea Nor Hypoxemia
Five subjects were aroused from NREM sleep during treatment with nasal CPAP (Fig. 4). Auditory arousal in the absence of apnea and hypoxemia was associated with significant BP elevation compared with NREM baseline (mean MAP elevation 24.0 t 8.1 mmHg; P = 0.0026). For this group of subjects, postapneic elevations averaged 22.0 t 6.9 mmHg (P < 0.001). As in experiment 1, no significantly between the two * v ” different MAP response
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 12, 2018.
2146
BLOOD
PRESSURE
ELEVATIONS
FOLLOWING
;
a z
120loo-
80-
DISCUSSION
600 2
3
6
8
10
SUBJECT 4. Mean
arterial pressure (MAP) changes associated with auditory arousal from NREM sleep (hatched bars) in 5 OSA patients receiving nasal continuous positive airway pressure. Stimuli (3 s) were administered by audiometer to each subject. NREM baseline and postapneic MAP represented by open and solid bars, respectively. Subjects are identified by number (see Table 1). FIG.
A
C
-
1
I
a,
a 2
-J W
m
a c3
Z
a I t)
m
-
2 W u
APNEAS
min (P = 0.0009). We calculated the difference in the magnitude of postapneic tachycardia between the two conditions to be significantly different (P c 0.02). The mean change in HR from NREM baseline after auditory arousal was -2.8 t 7.9 beats/min. Nonapneic hypoxemia (expt 2) was also associated with insignificant HR changes (6.1 t 11.3 beats/min).
16Cb
E E
OBSTRUCTIVE
I
Z
FIG. 5. Changes in mean arterial pressure (MAP) from NREM baseline in 4 conditions: A, postapneic at oxyhemoglobin saturation (Sao,) nadir 7942% (11 subj); B, postapneic at Sao, nadir L 90% (6 subj); C, hypoxemia (Sa o, 80%) without apnea (8 subj); D, arousal without apnea and without hypoxemia (5 subj). Responses in group C differ significantly from those in groups A, B, and D (P < 0.01).
conditions was detected. Figure 5 summarizes the results of the three experiments. Mean change in MAP from NREM baseline is
displayed for four conditions. BP elevations associated with postapneic Sa o, nadir of 79--82%, postapneic Sao, nadir of > 90%, and arousal with neither hypoxemia nor apnea were similar. The BP response to nonapneic hypoxemia did not differ significantly from NREM baseline, HR responses differed from BP responses in two notable ways. Neither 02-supplemented apneas (e3cpt 1) nor auditory arousals (expt 3) were associated with significant increases in HR. In experiment
1, the mean increase
for O2 supplemented apneas was 2.7 t 9.1 beats/min, not significantly different from baseline. For unsupplemented apneas, the mean increase was 10.5 2 6.6 beats/
Our studies of BP in OSA patients during NREM sleep yielded three notable findings. First, postapneic BP elevations were not significantly affected by amelioration of hypoxemia with supplemental OZ. Second, hypoxemia without apnea was not associated with significant BP elevations. Finally, auditory arousal from sleep was associated with BP elevations even in the absence of apnea and hypoxemia. Our inability to demonstrate an independent contribution from hypoxemia in effecting postapneic BP changes in OSA is intriguing. A search of the literature revealed one report (16) of blunting of the cyclic BP rise in response to supplemental O2 in OSA patients in NREM sleep. In that study, Sao, was restored to 100% and the duration of apneas is reported to have increased significantly. Statistical analysis does not accompany the data. In our subjects supplemented with 02, Sao, was allowed to fall as low as 90% and mean apnea duration increased by (10 s. Thus differences in experimental protocol may account for discrepant results in the two studies. In the induced hypoxemia experiments, our study design required matching levels of hypoxemia in the apneic and nonapneic conditions. Since we did not feel it was safe to induce Sao, below 79% in subjects undergoing nasal CPAP therapy, it was not possible to analyze potential independent hemodynamic effects of Sao,, below that level. There is evidence from studies in dogs that profound hypoxemia elicits a more powerful BP response to airway occlusion than does moderate hypoxemia (9). Because we did not analyze apneas associated with Sao, c 79%, we cannot exclude the possibility that hypoxemia becomes an important contributor to hemodynamic control in OSA at more severe levels of oxyhemoglobin desaturation than those studied here. Nevertheless, only the most severely afflicted OSA patients develop repetitive nocturnal desaturations significantly ~80%. Finally, it is possible that a sample size larger than was practical to study in this way might demonstrate
a statistically
significant
difference between
the responses of the two groups in experimerzt 1. Such a difference, if real, must be very small under any circumstances and would not detract from the validity of the contention that hypoxemia is not the controlling influence on BP in this setting. Since the 1970’s there have been several published reports of continuous BP measurements in OSA patients (3, 4, 12, 13, 16, 18, 20, 21). All of these accounts document BP oscillations in association with respiratory events. The postapneic BP elevation has been noted to coincide with the Sa9, nadir (18, 20). Variable intraapneic BP changes have been reported by two groups (11, 13), although others have described more reproduc-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 12, 2018.
BLOOD
PRESSURE
ELEVATIONS
FOLLOWING
ible patterns (4, 16, 18, 20). Mechanisms responsible for BP responses in OSA have been inferred in several ways. Shepard (18), using a linear regression analysis of the relationship of BP to Sao,, calculated that one-third of the BP elevation is attributable to hypoxemia. Employing impedance cardiography and measurements of esophageal pressure, Tolle et al. (21) argued that intra-apneic falls in pleural pressure are associated with decreased cardiac output and BP. Qualitatively similar intra-apneic falls in cardiac output were reported by Guilleminault and colleagues (7) using the thermodilution technique. Schroeder and colleagues (16), reporting BP changes in two OSA patients with Shy-Drager autonomic dysfunction and in several others with experimentally induced sympathetic or parasympathetic blockade, concluded that the BP elevations are mediated by the sympathetic nervous system. The same group constructed a speculative mechanism for BP elevations that included contributions from carotid chemoreceptor and pulmonary inflation reflexes as well as possible physical heart-lung interactions and central nervous system influences. Several investigators (2, 5, 10) have demonstrated correction of cardiovascular abnormalities in OSA patients after therapy with nasal CPAP or tracheostomy. Recent reviews of OSA hemodynamics (11, 15, 17, 18) have reiterated the notion that several factors combine to effect the cyclic hypertension observed in these patients. Other than the studies cited above, however, few investigations have sought to corroborate the model of Schroeder et al. (16) or to assign relative importance to each of the proposed contributing factors. Specifically, none to our knowledge has investigated separately the hemodynamic response to apnea termination, an event that is followed by the highest BP and the lowest Sao, of the apnea-recovery cycle. The concurrence of postapneic BP elevations and maximal oxyhemoglobin desaturation raises the possibility that the hypoxemia plays a central etiologic role in effecting the nocturnal hypertension described in patients with OSA. The results of our experiments, however, argue strongly to the contrary; i.e., postapneic BP elevations were preserved even after amelioration of hypoxemia. Thus other factors associated with apnea termination must be responsible for these elevations in NREM sleep. Possibilities include resolution of highly negative intrathoracic pressures, reinflation of the lungs, and abrupt transient arousal from sleep. Before these possibilities are explored, it should be noted that the signal from an ear oximeter does not reflect exactly the instantaneous increases in pulmonary venous Sao, resulting from postapneic ventilation. These increases are recorded only after a brief circulation delay between the lung and the ear. This delay has been estimated at 5 s (18). It is likely that the oximeter more precisely reflects the Sao, of blood perfusing carotid chemoreceptors at any given moment, and from a standpoint of reflex control of blood pressure, estimation of instantaneous carotid Sao, may be more important than pulmonary venous Sao,. For this reason, we did not correct the oximeter signal for circulatory delay in these studies. Scharf (15) has reviewed the potential hemodynamic influence of physical heart-lung interactions in the set-
OBSTRUCTIVE
2147
APNEAS
ting of the marked shifts of intrathoracic pressure observed during and following obstructive apneas. As we did not employ measurements of esophageal (pleural) pressure in these studies, we cannot assess the relative importance of intrathoracic pressure changes in the pathogenesis of nocturnal hypertension in OSA. The presence or absence of pulmonary inflation may affect BP responses in OSA in several ways. When we induced hypoxemia through the nasal CPAP mask (e@ Z), BP did not change significantly relative to NREM baseline (Fig. 3). One possible explanation for this finding is that the duration of the nonapneic hypoxemia trials (2-3 times the mean apnea duration) altered the hemodynamic response. Studies in dogs of the ventilatory response to chemoreceptor stimulation demonstrate that the temporal profile of the stimulus has a substantial effect on the response (6). Although we did not analyze critically the dynamics of hypoxemia induction in the present study, the primary finding of experiment 2, absence of BP elevation, was preserved in the several subjects whose Saq, reached 80% within 1 min. We therefore consider it unlikely that the timing of the hypoxic stimulus confounded our results. An alternative explanation for the finding in experiment 2 is that hyperventilation, which accompanied nonapneic hypoxemia, altered the cardiovascular effects of peripheral chemoreceptor stimulation. Respiratory-circulatory interactions of this type were first described in dogs by Daly and colleagues (l), who demonstrated that the bradycardia and systemic vasoconstriction observed during apneic hypoxemia could be reversed to tachycardia and vasodilation by allowing the animal to ventilate freely. Our data suggest that this relationship may be present in humans in NREM sleep. A similar pulmonary inflation reflex has been invoked to explain the tachycardia that follows obstructive apneas (16). If such a chronotropic effect precedes or supersedes any vasodilatory effect, reinflation of the lungs could also account for postapneic BP elevations. Our data do not support this hypothesis, however, in that postapneic BP elevations occurred in the absence of significant tachycardia during 02 supplementation. The disparities in our HR and BP results may, alternatively, reflect the need for larger sample size to appreciate relatively small HR changes. The attenuation of postapneic tachycardia during 02 supplementation is nevertheless intriguing in light of the experiments of Zwillich et al. (25) demonstrating that hypoxemia is necessary for the development dai .
of intra-apneic
bradycar-
Our auditory arousal experiments confirm the ability of sleep disruption alone to elevate BP. Repetitive abrupt arousals from sleep may be more important contributors to the hemodynamic response and hence to potential cardiovascular implications of OSA than previously appreciated. We have found no references in the literature regarding the acute cardiovascular effects of sudden sleep disruption as they may relate to OSA. Specific patterns of blood flow redistribution have been demonstrated in many species in response to defense or anger reactions. Hilton (8) has reviewed these responses, which, in summary, appear qualitatively similar to those observed in
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 12, 2018.
2148
BLOOD
PRESSURE
ELEVATIONS
association with the pulmonary inflation reflex (tachycardia, vasodilation of limb muscle beds). Data from studies of cats suggest that central effects of arousal may be able to override inhibitory chemoreceptor afferents (19). This mechanism provides a theoretical explanation for the relative unimportance of hypoxemia in our experiments. Specifically, peripheral chemoreceptor afferents may be overwhelmed by central effects of arousal. The hemodynamic potency of arousal may have practical cardiovascular implications as well. Ventricular irritability has been demonstrated to increase in dogs after provocation of an anger reaction (22), and life-threatening ventricular tachyarrhythmias have been observed in humans in response to abrupt auditory arousal from sleep (23). In summary, we have provided evidence that the BP elevations that follow obstructive apneas in NREM sleep are attributable primarily to factors other than arterial hypoxemia. Studies of nocturnal hypertension and other hemodynamic implications of OSA should not focus solely on severity of hypoxemia but, rath .er, should inelude consideration of reflex autonomic responses to arousal and resumption of ventilation. The authors thank David Lacasse, Nancy Logowitz, Sharon Berkowitz, and Edwyna von Gal for technical assistance; Dr. Bernard Ransil for assistance with statistical analysis; and Dr. Vladimir Fencl for reviewing the manuscript. This study was supported by Training Grant HL-07633 and Pulmonary Specialized Center of Research Grant HL-19170 from the National Heart, Lung, and Blood Institute and by a grant from the Jack and Pauline Freeman Foundation. This work was presented in part at the Annual Meeting of the American Thoracic Society, Cincinnati, OH, May 1989. Address for reprint requests: J. W. Weiss, Pulmonary Div., Beth Israel Hospital, 330 Brookline Ave., Boston, MA 02215. Received 20 November 1989; accepted in final form 20 July 1990. REFERENCES
J. E., AND M. DEBURGH DALY. Cardiovascular responses in apneic asphyxia: role of arterial chemoreceptors and the modification of their effects by a pulmonary vagal inflation reflex. J. Physiol. Land. 201: 87-104, 1969. 2. BECKER, H., M. FAUST, U. KOHLER, AND J. H. PETER. Reversibility of severe cardiac arrhythmias in sleep apnea under nasal continuous positive airway pressure therapy. In: Sleep Related Disorders and Internal Diseases, edited by J. H. Peter, T. Podszus, and P. von Wichert. Berlin: Springer-Verlag, 1987, p. 380-387. 3. BUDA, A. J., J. S. SCHROEDER, AND C. GUILLEMINAULT. Abnormalities of pulmonary artery wedge pressure in sleep-induced apnea. Int. J. Cardiol. 1: 67-74, 1981. 4. COCCAGNA, G., M. MANTOVANI, F. BRIGNANI, C. PARCHI, AND E. LUGARESI. Continuous recording of the pulmonary and systemic arterial pressure during sleep in syndromes of hypersomnia with periodic breathing. Bull. Eur. Physiopathol. Respir. 8: 1159-l 172, 1. ANGELL
JAMES,
1972. 5. COCCAGNA, G., M. MANTOVANI, F. BRIGNANI, C. PARCHI, AND E. LUGARESI. Tracheostomy in hypersomnia with periodic breathing. Bull. Eur. Physiopathol. Respir. 8: 1217-1227, 1972. 6. DUTTON, R. E., E. J. SMITH, P. K. GHATAK, AND D. G. DAVIES.
FOLLOWING
OBSTRUCTIVE
APNEAS
Dynamics of the respiratory controller during carotid body hypoxia. J. Appl. Physiol. 7. GUILLEMINAULT,
35: 844-850, 1973. C., J. MOTTA, F. MIHM,
AND K. MELVIN. Obstructive sleep apnea and cardiac index. Chest 89: 331-334, 1986. 8. HILTON, S. M. The defense-arousal system and its relevance for circulatory and respiratory control. J. Exp. Biol. 100: 159-174, 1982. 9. IWASE, OKABE,
N., Y. KIKUCHI, W. HIDA, M. SATOH, 0. TAGUCHI, S. H. INOUE, AND T. TAKISHIMA. Effects of phentolamine and carotid body resection on pulmonary artery and systemic artery pressure during repetitive airway obstruction (Abstract). Am. Rev.
Respir. Dis. 139: A181, 1989. 10. MARRONE, O., G. FERRARA, C. MACALUSO, BELLA, V. BELLIA, AND G. BONSIGNORE.
F. MILONE, F. CIHemodynamics in obstructive sleep apnea patients treated by continuous positive airway pressure. In: Sleep Related Disorders and Internal Diseases, edited by J. H. Peter, T. Podszus, and P. von Wichert. Berlin: SpringerVerlag, 1987, p. 375-379. 11. MARTIN, R. J. The role of sleep-related breathing disorders in cardiorespiratory disease: hemodynamics. In: Sleep Related Disorders and InternaL Diseases, edited by J. H. Peter, T. Podszus, and P. von Wichert. Berlin: Springer-Verlag, 1987, p. 299-312. 12. MAYER, J., H. GREB, B. HERRES, T. M. KLOSS, T. PENZEL, J. H. PETER, T. PODSZUS, AND P. VON WICHERT. Nocturnal hemodynamics in patients with sleep apnea. In: Sleep Related Disorders and Internal Diseases, edited by J. H. Peter, T. Podszus, and P.
von Wichert. Berlin: Springer-Verlag,
1987, p. 315-320.
13. PODSZUS, T., J. MAYER, T. PENZEL, J. H. PETER, AND P. VON WICHERT. Nocturnal hemodynamics in patients with sleep apnea. Eur. J. Respir. Dis. 69: 435-442, 1986. 14. RECHTSCHAFFEN, A., AND A. KALES. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Sleep Subjects. Washington, DC: US Govt. Printing Office, 1968. 15. SCHARF, S. M. Influence of sleep state and breathing on cardiovascular function. In: In: SZeep and Breathing, edited by N. A. Saunders and C. E. Sullivan. New York: Dekker, 1984, vol. 21, p. 221240. (Lung Biol. Health Dis. Ser.) 16. SCHROEDER, J. S., J. MOTTA, AND C. GUILLEMINAULT. Hemodynamic studies in sleep apnea. In: Sleep Apnea Syndromes, edited by C. Guilleminault and W. C. Dement. New York: Liss, 1982, p. 177-196. 17. SHEPARD, J. W. Gas exchange and hemodynamics during sleep. Med. Clin. N. Am. 69: 1243-1264, 1985. 18. SHEPARD, J. W. Hemodynamics in Obstructive Sleep Apnea. In: Abnormalities of Respiration During Sleep, edited by E. C. Fletcher.
Orlando, FL: Grune & Stratton, 1986, p. 39-62. M. R., AND F. R. CALARESU. Hypothalamic inhibition of chemoreceptor-induced bradycardia in the cat. Am. J. Physiol. 225:
19. THOMAS,
201-208,1973. 20. TILKIAN, LEHRMAN,
A. G., C. GUILLEMINAULT, J. S. SCHROEDER, K. L. F. B. SIMMONS, AND W. DEMENT. Hemodynamics in sleep-induced apnea. Ann. Intern. Med. 85: 714-719,1976. 21. TOLLE, F. A., W. V. JUDY, P. Yu, AND 0. N. MARKAND. Reduced stroke volume related to pleural pressure in obstructive sleep apnea. J. Appl. Physiol. 55: 1718-1724, 1983. 22. VERRIER, R. L. Behavioral state and Clinical Physiology of Sleep, edited by R.
cardiac arrhythmias. In: Lydic and J. F. Biebuyck. Bethesda, MD: Am. Physiol. Sot., 1988, p. 31-51. (Clin. Physiol. Ser.) 23. WELLENS, H. J. J., A. VERMEULEN, AND D. DURRER. Ventricular fibrillation occurring on arousal from sleep by auditory stimuli. Circulation 52: 73-81, 1975. 24. ZAR, J. H. Biostatistical Analysis. Englewood Cliffs, NJ: PrenticeHall, 1974. 25. ZWILLICH, C., T. DEVLIN, D. WHITE, N. DOUGLAS, J. WEIL, AND R. MARTIN. Bradycardia during sleep apnea. J. Clin. Inuest. 69: 1286-1292,1982.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 12, 2018.
RINGLER,JACK, ROBERT C. BASNER,RICHARD SHANNON, RICHARD SCHWARTZSTEIN,HAROLD MANNING, STEVEN E. WEINBERGER,ANDJ. W~~~~~~W~~~~.Hypoxemiaalonedoes not explain blood pressure elevations after obstructive apneas. J. Appl. Physiol. 69(6): 2143-2148, 1990.-In patients with obstructive sleep apnea (OSA), substantial elevations of systemic blood pressure (BP) and depressions of oxyhemoglobin saturation (Sao,) accompany apnea termination. The causes of the BP elevations, which contribute significantly to nocturnal hypertension in OSA, have not been defined precisely. To assess the relative contribution of arterial hypoxemia, we observed mean arterial pressure (MAP) changes following obstructive apneas in 11 OSA patients during non-rapid-eye-movement (NREM) sleep and then under three experimental conditions: 1) apnea with 0, supplementation; 2) hypoxemia (Sao, 80%) without apnea; and 3) arousal from sleep with neither hypoxemia nor apnea. We found that apneas recorded during O2 supplementation (Sao, nadir 93.6% t 2.4; mean t SD) in six subjects were associated with equivalent postapneic MAP elevations compared with unsupplemented apneas ( Sao.,nadir 7982%): 18.8 -t- 7.1 vs. 21.3 t 9.2 mmHg (mean change MAP k SD); in the absence of respiratory and sleep disruption in eight subjects, hypoxemia was not associated with the BP elevations observed following apneas: -5.4 t 19 vs. 19.1 rfi 7.8 mmHg (P c 0.01); and in five subjects, auditory arousal alone was associated with MAP elevation similar to that observed following apneas: 24.0 t 8.1 vs. 22.0 k 6.9 mmHg. We conclude that in NREM sleep postapneic BP elevations are not primarily attributable to arterial hypoxemia. Other factors associated with apnea termination, including arousal from sleep, reinflation of the lungs, and changes of intrathoracic pressure, may be responsible for these elevations. obstructive sleep apnea; hemodynamics; arousal; human; nonrapid-eye-movement sleep OBSTRUCTIVE APNEAS during sleep are associated with oscillations of systemic blood pressure (BP) that occur in phase with disruptions of respiration and sleep. Substantial BP elevations coincide with termination of apneas. Even in patients who are normotensive during the day, postapneic systolic BP can exceed 200 mmHg (20). Several factors have been reported to be responsible for the BP changes associated with obstructive sleep apnea (OSA). These include arterial hypoxemia, intrathoracic pressure changes, interruptions of ventilation, and disruption of sleep architecture (11, 13, 15-18, 20). The 0161-7567/90
$1.50
Copyright
relative contribution of each of these factors to the integrated BP response remains unclear, however. By analyzing apneas associated with similar degrees of hypoxemia and then by manipulating independently our subjects’ oxygenation, ventilation, and sleep architecture, we hoped to clarify whether hypoxemia has an independent role in effecting BP elevations following non-rapid-eye-movement (NREM) obstructive apneas. METHODS Subjects We enrolled 16 subjects (13 men, 3 women) with clinical features of OSA: 11 of them successfully completed the experimental protocol (see Data Analysis, below). Each was referred to the Beth Israel Hospital Sleep Disorders Center for diagnostic polysomnography (PSG). Informed consent was obtained in accordance with guidelines of the Committee on Clinical Investigations at Beth Israel Hospital. Vasoactive medications were withheld 48 h before the study night. Measurements After undergoing history and physical examination, subjects were prepared for overnight PSG. A standard PSG montage, including three channels of electroencephalogram (EEG), left and right electrooculograms, chin and intercostal electromyogram (EMG), electrocardiogram (ECG), nasal and oral airflow by thermistors, respiratory effort by inductance plethysmography belts (Respitrace, Ardsley, NY), and ear oximetry (Biox IIA, Biox Technology, Boulder, CO), was utilized (14). Systemic arterial pressure was monitored by means of a 20gauge radial artery catheter. All data were recorded on a ‘I$-channel EEG (Grass model 8-24D, Grass Instruments, Quincy, MA). Pro toe01 Each subject was studied on a single night between 1090 P.M. and 6:00 A.M. For the first 2-4 h of sleep we recorded BP continuously without intervention. During the remainder of the night we performed a series of experiments. Since the subjects were patients undergoing
0 1990 the American
Physiological
Society
2143
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 12, 2018.
2144
BLOOD
PRESSURE
ELEVATIONS
FOLLOWING
clinical PSG, the complete protocol could not be completed in each case. Experiment 1. In six subjects, we began by observing BP and heart rate (HR) changes during administration of supplemental O2 at a rate of 3-11 l/min via nasal cannula. We adjusted the flow rate to maintain oxyhemoglobin saturation (Sao,) > 90% during and after apneas. Experiment 2. In eight subjects, we assessed the BP and HR responses to hypoxemia without apnea. &neas were eliminated by means of nasal continuous positive airway pressure (CPAP; Sleepeasy III, Respironics, Monroeville, PA). Episodic hypoxemia was achieved by connecting the CPAP tubing to a second CPAP compressor equipped with a balloon reservoir for hypoxic gas (8-12% 02, 88-92% N2). We manually adjusted the composition of the gas in the reservoir to allow Sao, to fall to 80% over a period of L-4 min. Experiment 3. We aroused from sleep five subjects undergoing nasal CPAP therapy by providing intermittent auditory stimulation with a portable audiometer (model MA-20, Maico Hearing Instruments, Minneapolis, MN). This allowed us to examine the BP response to arousal accompanied by neither apnea nor hypoxemia. We defined arousal polysomnographically as an increase in EMG amplitude accompanied by a change in pattern on at least one additional channel (14). We adjusted the amplitude and frequency of the auditory stimulus to induce PSG changes qualitatively similar to those associated with apneas. Data Analysis
Data obtained from a given subject were included in the analysis if the following criteria were met: 1) apneahypopnea index > 30 events/h; 2) at least six NREM apneas with associated Sao, nadir of 79-82%; 3) at least six NREM apneas recorded during O2 therapy associated with Sao, nadir > 90% or at least one episode of experimentally induced hypoxemia without apnea ending with Sao, 5 80% without change of sleep stage. Data from 11 of the 16 enrolled subjects were included in the analysis. Clinical characteristics of these subjects are displayed in Table 1. Of the five who did not complete 1. Subject characteristics
TABLE
Sao,% y-u-
sex
A;;;
;;>
fg-)s
sn
*
I 2 3 4 5 6 7 8 9 10 II
MAP9 mm&
F M F F M M M M M M M
24 41 29 45 71 38 47 35 38 46 41
127 125 72 123 70 150 173 118 120 97 102
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
114 108 90 97 103 93 129 107 117 84 82
AI-II, events/h
203 61 106 103 47 57 119 118 120 105 78
Baseline
Nfean lowest
Lowest
97
64 83 88 60 80
56 57
97 95 97 97 93 94 98 97 97 96
EDS, excessive daytime hypersomnolence; Sn, snoring; line mean arterial pressure during quiet wakefulness; hypopnea index; Sao,%, percent oxyhemoglobin saturation.
80 88 83 84 72 81
79 51 70 67 71 79 71 71 77
MAP, baseAHI, apnea-
OBSTRUCTIVE
APNEAS
the protocol, four did not develop oxyhemoglobin desaturations below 83% during NREM sleep, thus failing to meet criterion 2 above; the fifth failed to meet criterion 3 by arousing repeatedly during attempts to induce hypoxemia through the nasal CPAP system. Sleep recordings were scored using standard criteria (14). We estimated baseline BP during quiet wakefulness and during NREM sleep on nasal CPAP therapy in each subject by averaging mean arterial pressure (MAP) values tabulated at 2-s intervals from a typical 30-s epoch of the PSG fulfilling the desired criteria. We considered MAP equivalent to one-third systolic BP plus two-thirds diastolic BP. Between 6 and 15 apneas associated with Sao, nadir of 79-82% were analyzed for each subject. If more than 15 apneas qualified, as was the case in six of the subjects, we randomly selected 15 for analysis. For each selected apnea, we tabulated MAP values over the three to six cardiac cycles coinciding with the Sao, nadir. These values were without exception within 10 mmHg of the peak postapneic MAP and included the peak in the majority of cases. For each subject, data from all analyzed apneas were combined, yielding an average MAP at the Sao, nadir of 79-82%. We then analyzed similarly the 08-supplemented apneas in the six subjects for experiment 1. Again, 6-15 apneas per subject were analyzed, and as observed in unsupplemented apneas Sao,, nadir correlated very closely with peak postapneic BP. By comparing MAP following apneas with nadirs of 79-82% to MAP following apneas with nadirs > 90% in each subject, we were able to examine the relationship between postapneic BP elevation and Sao, nadir. Episodes of hypoxemia without apnea were analyzed in eight subjects by tabulating MAP values over the first 10 cardiac cycles after Sao, reached 80%. Because of our concern for subject safety, hypoxemia was achieved more gradually during these episodes than during the naturally occurring apneas. Data from one to three hypoxemia trials in each subject were combined yielding an average MAP at Sao, of 80%. We were thus able to compare the average MAP associated with similar levels of Sao, in apneic and nonapneic conditions. In the five subjects for experiment 3, we assessed the effect on BP of auditory arousal accompanied by neither apnea nor hypoxemia by calculating average MAP over the first five cardiac cycles following the 3-s stimulus. Because of the constraints imposed by the clinical testing, the number of arousal trials was necessarily limited (3 trials in subj 2 and 10, 2 trials in subj 6 and 8, and 1 trial in subj 3). The responses were quite stereotypical, however (see RESULTS). We compiled HR data in the same manner as the BP analysis described above. HR in beats per minute was derived as 60/RR interval in seconds. For the statistical analysis, we employed two-tailed paired t tests and/or Wilcoxon’s signed rank tests, each with Bonferroni correction for multiple comparisons (24). We compared in this way MAP responses following naturally occurring apneas, 02-supplemented apneas, episodes of induced hypoxemia without apnea, and episodes of auditory arousal each to the NREM baseline MAP
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 12, 2018.
BLOOD
PRESSURE
ELEVATIONS
FOLLOWING
during nasal CPAP therapy. These calculations were performed at the Core Laboratory Computer Facilities of Beth Israel Hospital. MAP values referable to extrasystoles or to motion artifact were excluded from analysis in all phases of the study. RESULTS
Experiment 1. Postapneic BP Response with 02 Supplementation
Experiment 2. BP Response to Hypoxemia Without Apnea
EEG EOG
between MAP and Sao, to of apneas with nasal CPAP
WFF ’
140 F E E $
m
-~~’
-a
120loo809
g
60’
6
7
8
9
10
11
SUBJECT FIG. 2. Mean arterial pressure (MAP) following apneas associated with oxyhemoglobin saturation (Sao,) nadir 7942% (solid bars) and apneas associated with Sao, nadir 2 90% (hatched bars) in 6 subjects. NREM baseline MAP represented by open bars. For group, postapneic conditions both differed significantly from baseline but not from each other (see text). Subjects are identified by number (see Table 1). All data in Figs. 2-5 displayed as means t SD.
160
Both apneas with Sao, nadirs > 90% (mean 94 t 2.4%) and apneas with Sao, nadirs of 79-82% were associated with significant elevations of MAP compared with NREM baseline (Fig. 2). In the six subjects studied, we found the mean elevation over NREM baseline for 02supplemented apneas to be 18.8 t 7.1 mmHg (mean change in MAP t SD; P = 0.0026). For unsupplemented apneas, the mean elevation was 21.3 t 9.2 mmHg (P c 0.0001). We did not detect a significant difference between the two conditions (supplemented vs. unsupplemented) with regard to the magnitude of the postapneic BP elevation. Supplemental O2 was associated with I- to 10-s increases in mean apnea duration in subjects 6, 7, 9, 10, and I1 and with a decrease of 1 s in subject 8.
We found the relationship be different after elimination
2145
APNEAS
160~
=
Figure 1 displays the PSG recorded during a typical NREM apnea. All subjects had substantial and repetitive BP elevations in the immediate postapneic period. This period in all cases encompassed the lowest Sao, and the highest BP of the apnea-recovery cycle. BP responses during apneas were characterized by intersubject and intrasubject variability. For this reason all comparisons were made to the NREM baseline MAP.
OBSTRUCTIVE
I-
. 10 SECON;
FIG. 1. Polysomnographic and radial arterial pressure recording during and after a typical NREM obstructive apnea (su6j 8). Blood pressure (BP) and heart rate (HR) fall during most of apnea, and cyclic BP depressions coincide with inspiratory efforts toward end of event. Sharp rises of BP and HR coincide with arousal, resumption of ventilation, and oxyhemoglobin saturation nadir. EEG, electroencephalogram; EOG, left and right electrooculograms; EMG,S, submental electromyogram; EMG-I, intercostal electromyogram; V, oronasal airflow (thermistor signals); ABD, respiratory effort (by abdominal inductance plethysmography); AP, radial arterial pressure; Sao,, percent oxyhernoglobin saturation (by ear oximetry).
1 -
t40-
0, 2 E
120-
; d 2
loo80606 1
2
3
4
5
6
7
8
SUBJECT FIG. 3. Mean arterial pressure (MAP) after induction of hypoxemia [oxyhemoglobin saturation (Sao,) 80%] without apnea (hatched bars) in 8 subjects compared with NREM baseline (open bars) and with postapneic MAP at Sa 0, nadir 7942% (solid bars). In absence of respiratory and sleep disruption, hypoxemia was associated with no significant change in MAP compared with baseline (see text). Subjects are identified by number (see Table 1).
(Fig. 3). In the eight subjects studied, postapneic hypoxemia was associated with marked BP elevations as in experiment 1. The mean elevation for these subjects was 19.1 t 7.8 mmHg (P c 0.0001). By contrast, without respiratory and sleep disruption, the same degree of hypoxemia in the same eight subjects was not associated with substantial BP elevations. In this condition, the mean change was -5.4 t 18.9 mmHg, not significantly different from baseline. We calculated the difference in MAP response between the apneic and nonapneic conditions to be statistically significant (P < 0.01). This difference persists even if subject 1, in whom BP declined precipitously under the conditions of experiment 2, is treated as a statistical outlier. Experiment 3. BP Response to Arousal With Neither Apnea Nor Hypoxemia
Five subjects were aroused from NREM sleep during treatment with nasal CPAP (Fig. 4). Auditory arousal in the absence of apnea and hypoxemia was associated with significant BP elevation compared with NREM baseline (mean MAP elevation 24.0 t 8.1 mmHg; P = 0.0026). For this group of subjects, postapneic elevations averaged 22.0 t 6.9 mmHg (P < 0.001). As in experiment 1, no significantly between the two * v ” different MAP response
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 12, 2018.
2146
BLOOD
PRESSURE
ELEVATIONS
FOLLOWING
;
a z
120loo-
80-
DISCUSSION
600 2
3
6
8
10
SUBJECT 4. Mean
arterial pressure (MAP) changes associated with auditory arousal from NREM sleep (hatched bars) in 5 OSA patients receiving nasal continuous positive airway pressure. Stimuli (3 s) were administered by audiometer to each subject. NREM baseline and postapneic MAP represented by open and solid bars, respectively. Subjects are identified by number (see Table 1). FIG.
A
C
-
1
I
a,
a 2
-J W
m
a c3
Z
a I t)
m
-
2 W u
APNEAS
min (P = 0.0009). We calculated the difference in the magnitude of postapneic tachycardia between the two conditions to be significantly different (P c 0.02). The mean change in HR from NREM baseline after auditory arousal was -2.8 t 7.9 beats/min. Nonapneic hypoxemia (expt 2) was also associated with insignificant HR changes (6.1 t 11.3 beats/min).
16Cb
E E
OBSTRUCTIVE
I
Z
FIG. 5. Changes in mean arterial pressure (MAP) from NREM baseline in 4 conditions: A, postapneic at oxyhemoglobin saturation (Sao,) nadir 7942% (11 subj); B, postapneic at Sao, nadir L 90% (6 subj); C, hypoxemia (Sa o, 80%) without apnea (8 subj); D, arousal without apnea and without hypoxemia (5 subj). Responses in group C differ significantly from those in groups A, B, and D (P < 0.01).
conditions was detected. Figure 5 summarizes the results of the three experiments. Mean change in MAP from NREM baseline is
displayed for four conditions. BP elevations associated with postapneic Sa o, nadir of 79--82%, postapneic Sao, nadir of > 90%, and arousal with neither hypoxemia nor apnea were similar. The BP response to nonapneic hypoxemia did not differ significantly from NREM baseline, HR responses differed from BP responses in two notable ways. Neither 02-supplemented apneas (e3cpt 1) nor auditory arousals (expt 3) were associated with significant increases in HR. In experiment
1, the mean increase
for O2 supplemented apneas was 2.7 t 9.1 beats/min, not significantly different from baseline. For unsupplemented apneas, the mean increase was 10.5 2 6.6 beats/
Our studies of BP in OSA patients during NREM sleep yielded three notable findings. First, postapneic BP elevations were not significantly affected by amelioration of hypoxemia with supplemental OZ. Second, hypoxemia without apnea was not associated with significant BP elevations. Finally, auditory arousal from sleep was associated with BP elevations even in the absence of apnea and hypoxemia. Our inability to demonstrate an independent contribution from hypoxemia in effecting postapneic BP changes in OSA is intriguing. A search of the literature revealed one report (16) of blunting of the cyclic BP rise in response to supplemental O2 in OSA patients in NREM sleep. In that study, Sao, was restored to 100% and the duration of apneas is reported to have increased significantly. Statistical analysis does not accompany the data. In our subjects supplemented with 02, Sao, was allowed to fall as low as 90% and mean apnea duration increased by (10 s. Thus differences in experimental protocol may account for discrepant results in the two studies. In the induced hypoxemia experiments, our study design required matching levels of hypoxemia in the apneic and nonapneic conditions. Since we did not feel it was safe to induce Sao, below 79% in subjects undergoing nasal CPAP therapy, it was not possible to analyze potential independent hemodynamic effects of Sao,, below that level. There is evidence from studies in dogs that profound hypoxemia elicits a more powerful BP response to airway occlusion than does moderate hypoxemia (9). Because we did not analyze apneas associated with Sao, c 79%, we cannot exclude the possibility that hypoxemia becomes an important contributor to hemodynamic control in OSA at more severe levels of oxyhemoglobin desaturation than those studied here. Nevertheless, only the most severely afflicted OSA patients develop repetitive nocturnal desaturations significantly ~80%. Finally, it is possible that a sample size larger than was practical to study in this way might demonstrate
a statistically
significant
difference between
the responses of the two groups in experimerzt 1. Such a difference, if real, must be very small under any circumstances and would not detract from the validity of the contention that hypoxemia is not the controlling influence on BP in this setting. Since the 1970’s there have been several published reports of continuous BP measurements in OSA patients (3, 4, 12, 13, 16, 18, 20, 21). All of these accounts document BP oscillations in association with respiratory events. The postapneic BP elevation has been noted to coincide with the Sa9, nadir (18, 20). Variable intraapneic BP changes have been reported by two groups (11, 13), although others have described more reproduc-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 12, 2018.
BLOOD
PRESSURE
ELEVATIONS
FOLLOWING
ible patterns (4, 16, 18, 20). Mechanisms responsible for BP responses in OSA have been inferred in several ways. Shepard (18), using a linear regression analysis of the relationship of BP to Sao,, calculated that one-third of the BP elevation is attributable to hypoxemia. Employing impedance cardiography and measurements of esophageal pressure, Tolle et al. (21) argued that intra-apneic falls in pleural pressure are associated with decreased cardiac output and BP. Qualitatively similar intra-apneic falls in cardiac output were reported by Guilleminault and colleagues (7) using the thermodilution technique. Schroeder and colleagues (16), reporting BP changes in two OSA patients with Shy-Drager autonomic dysfunction and in several others with experimentally induced sympathetic or parasympathetic blockade, concluded that the BP elevations are mediated by the sympathetic nervous system. The same group constructed a speculative mechanism for BP elevations that included contributions from carotid chemoreceptor and pulmonary inflation reflexes as well as possible physical heart-lung interactions and central nervous system influences. Several investigators (2, 5, 10) have demonstrated correction of cardiovascular abnormalities in OSA patients after therapy with nasal CPAP or tracheostomy. Recent reviews of OSA hemodynamics (11, 15, 17, 18) have reiterated the notion that several factors combine to effect the cyclic hypertension observed in these patients. Other than the studies cited above, however, few investigations have sought to corroborate the model of Schroeder et al. (16) or to assign relative importance to each of the proposed contributing factors. Specifically, none to our knowledge has investigated separately the hemodynamic response to apnea termination, an event that is followed by the highest BP and the lowest Sao, of the apnea-recovery cycle. The concurrence of postapneic BP elevations and maximal oxyhemoglobin desaturation raises the possibility that the hypoxemia plays a central etiologic role in effecting the nocturnal hypertension described in patients with OSA. The results of our experiments, however, argue strongly to the contrary; i.e., postapneic BP elevations were preserved even after amelioration of hypoxemia. Thus other factors associated with apnea termination must be responsible for these elevations in NREM sleep. Possibilities include resolution of highly negative intrathoracic pressures, reinflation of the lungs, and abrupt transient arousal from sleep. Before these possibilities are explored, it should be noted that the signal from an ear oximeter does not reflect exactly the instantaneous increases in pulmonary venous Sao, resulting from postapneic ventilation. These increases are recorded only after a brief circulation delay between the lung and the ear. This delay has been estimated at 5 s (18). It is likely that the oximeter more precisely reflects the Sao, of blood perfusing carotid chemoreceptors at any given moment, and from a standpoint of reflex control of blood pressure, estimation of instantaneous carotid Sao, may be more important than pulmonary venous Sao,. For this reason, we did not correct the oximeter signal for circulatory delay in these studies. Scharf (15) has reviewed the potential hemodynamic influence of physical heart-lung interactions in the set-
OBSTRUCTIVE
2147
APNEAS
ting of the marked shifts of intrathoracic pressure observed during and following obstructive apneas. As we did not employ measurements of esophageal (pleural) pressure in these studies, we cannot assess the relative importance of intrathoracic pressure changes in the pathogenesis of nocturnal hypertension in OSA. The presence or absence of pulmonary inflation may affect BP responses in OSA in several ways. When we induced hypoxemia through the nasal CPAP mask (e@ Z), BP did not change significantly relative to NREM baseline (Fig. 3). One possible explanation for this finding is that the duration of the nonapneic hypoxemia trials (2-3 times the mean apnea duration) altered the hemodynamic response. Studies in dogs of the ventilatory response to chemoreceptor stimulation demonstrate that the temporal profile of the stimulus has a substantial effect on the response (6). Although we did not analyze critically the dynamics of hypoxemia induction in the present study, the primary finding of experiment 2, absence of BP elevation, was preserved in the several subjects whose Saq, reached 80% within 1 min. We therefore consider it unlikely that the timing of the hypoxic stimulus confounded our results. An alternative explanation for the finding in experiment 2 is that hyperventilation, which accompanied nonapneic hypoxemia, altered the cardiovascular effects of peripheral chemoreceptor stimulation. Respiratory-circulatory interactions of this type were first described in dogs by Daly and colleagues (l), who demonstrated that the bradycardia and systemic vasoconstriction observed during apneic hypoxemia could be reversed to tachycardia and vasodilation by allowing the animal to ventilate freely. Our data suggest that this relationship may be present in humans in NREM sleep. A similar pulmonary inflation reflex has been invoked to explain the tachycardia that follows obstructive apneas (16). If such a chronotropic effect precedes or supersedes any vasodilatory effect, reinflation of the lungs could also account for postapneic BP elevations. Our data do not support this hypothesis, however, in that postapneic BP elevations occurred in the absence of significant tachycardia during 02 supplementation. The disparities in our HR and BP results may, alternatively, reflect the need for larger sample size to appreciate relatively small HR changes. The attenuation of postapneic tachycardia during 02 supplementation is nevertheless intriguing in light of the experiments of Zwillich et al. (25) demonstrating that hypoxemia is necessary for the development dai .
of intra-apneic
bradycar-
Our auditory arousal experiments confirm the ability of sleep disruption alone to elevate BP. Repetitive abrupt arousals from sleep may be more important contributors to the hemodynamic response and hence to potential cardiovascular implications of OSA than previously appreciated. We have found no references in the literature regarding the acute cardiovascular effects of sudden sleep disruption as they may relate to OSA. Specific patterns of blood flow redistribution have been demonstrated in many species in response to defense or anger reactions. Hilton (8) has reviewed these responses, which, in summary, appear qualitatively similar to those observed in
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 12, 2018.
2148
BLOOD
PRESSURE
ELEVATIONS
association with the pulmonary inflation reflex (tachycardia, vasodilation of limb muscle beds). Data from studies of cats suggest that central effects of arousal may be able to override inhibitory chemoreceptor afferents (19). This mechanism provides a theoretical explanation for the relative unimportance of hypoxemia in our experiments. Specifically, peripheral chemoreceptor afferents may be overwhelmed by central effects of arousal. The hemodynamic potency of arousal may have practical cardiovascular implications as well. Ventricular irritability has been demonstrated to increase in dogs after provocation of an anger reaction (22), and life-threatening ventricular tachyarrhythmias have been observed in humans in response to abrupt auditory arousal from sleep (23). In summary, we have provided evidence that the BP elevations that follow obstructive apneas in NREM sleep are attributable primarily to factors other than arterial hypoxemia. Studies of nocturnal hypertension and other hemodynamic implications of OSA should not focus solely on severity of hypoxemia but, rath .er, should inelude consideration of reflex autonomic responses to arousal and resumption of ventilation. The authors thank David Lacasse, Nancy Logowitz, Sharon Berkowitz, and Edwyna von Gal for technical assistance; Dr. Bernard Ransil for assistance with statistical analysis; and Dr. Vladimir Fencl for reviewing the manuscript. This study was supported by Training Grant HL-07633 and Pulmonary Specialized Center of Research Grant HL-19170 from the National Heart, Lung, and Blood Institute and by a grant from the Jack and Pauline Freeman Foundation. This work was presented in part at the Annual Meeting of the American Thoracic Society, Cincinnati, OH, May 1989. Address for reprint requests: J. W. Weiss, Pulmonary Div., Beth Israel Hospital, 330 Brookline Ave., Boston, MA 02215. Received 20 November 1989; accepted in final form 20 July 1990. REFERENCES
J. E., AND M. DEBURGH DALY. Cardiovascular responses in apneic asphyxia: role of arterial chemoreceptors and the modification of their effects by a pulmonary vagal inflation reflex. J. Physiol. Land. 201: 87-104, 1969. 2. BECKER, H., M. FAUST, U. KOHLER, AND J. H. PETER. Reversibility of severe cardiac arrhythmias in sleep apnea under nasal continuous positive airway pressure therapy. In: Sleep Related Disorders and Internal Diseases, edited by J. H. Peter, T. Podszus, and P. von Wichert. Berlin: Springer-Verlag, 1987, p. 380-387. 3. BUDA, A. J., J. S. SCHROEDER, AND C. GUILLEMINAULT. Abnormalities of pulmonary artery wedge pressure in sleep-induced apnea. Int. J. Cardiol. 1: 67-74, 1981. 4. COCCAGNA, G., M. MANTOVANI, F. BRIGNANI, C. PARCHI, AND E. LUGARESI. Continuous recording of the pulmonary and systemic arterial pressure during sleep in syndromes of hypersomnia with periodic breathing. Bull. Eur. Physiopathol. Respir. 8: 1159-l 172, 1. ANGELL
JAMES,
1972. 5. COCCAGNA, G., M. MANTOVANI, F. BRIGNANI, C. PARCHI, AND E. LUGARESI. Tracheostomy in hypersomnia with periodic breathing. Bull. Eur. Physiopathol. Respir. 8: 1217-1227, 1972. 6. DUTTON, R. E., E. J. SMITH, P. K. GHATAK, AND D. G. DAVIES.
FOLLOWING
OBSTRUCTIVE
APNEAS
Dynamics of the respiratory controller during carotid body hypoxia. J. Appl. Physiol. 7. GUILLEMINAULT,
35: 844-850, 1973. C., J. MOTTA, F. MIHM,
AND K. MELVIN. Obstructive sleep apnea and cardiac index. Chest 89: 331-334, 1986. 8. HILTON, S. M. The defense-arousal system and its relevance for circulatory and respiratory control. J. Exp. Biol. 100: 159-174, 1982. 9. IWASE, OKABE,
N., Y. KIKUCHI, W. HIDA, M. SATOH, 0. TAGUCHI, S. H. INOUE, AND T. TAKISHIMA. Effects of phentolamine and carotid body resection on pulmonary artery and systemic artery pressure during repetitive airway obstruction (Abstract). Am. Rev.
Respir. Dis. 139: A181, 1989. 10. MARRONE, O., G. FERRARA, C. MACALUSO, BELLA, V. BELLIA, AND G. BONSIGNORE.
F. MILONE, F. CIHemodynamics in obstructive sleep apnea patients treated by continuous positive airway pressure. In: Sleep Related Disorders and Internal Diseases, edited by J. H. Peter, T. Podszus, and P. von Wichert. Berlin: SpringerVerlag, 1987, p. 375-379. 11. MARTIN, R. J. The role of sleep-related breathing disorders in cardiorespiratory disease: hemodynamics. In: Sleep Related Disorders and InternaL Diseases, edited by J. H. Peter, T. Podszus, and P. von Wichert. Berlin: Springer-Verlag, 1987, p. 299-312. 12. MAYER, J., H. GREB, B. HERRES, T. M. KLOSS, T. PENZEL, J. H. PETER, T. PODSZUS, AND P. VON WICHERT. Nocturnal hemodynamics in patients with sleep apnea. In: Sleep Related Disorders and Internal Diseases, edited by J. H. Peter, T. Podszus, and P.
von Wichert. Berlin: Springer-Verlag,
1987, p. 315-320.
13. PODSZUS, T., J. MAYER, T. PENZEL, J. H. PETER, AND P. VON WICHERT. Nocturnal hemodynamics in patients with sleep apnea. Eur. J. Respir. Dis. 69: 435-442, 1986. 14. RECHTSCHAFFEN, A., AND A. KALES. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Sleep Subjects. Washington, DC: US Govt. Printing Office, 1968. 15. SCHARF, S. M. Influence of sleep state and breathing on cardiovascular function. In: In: SZeep and Breathing, edited by N. A. Saunders and C. E. Sullivan. New York: Dekker, 1984, vol. 21, p. 221240. (Lung Biol. Health Dis. Ser.) 16. SCHROEDER, J. S., J. MOTTA, AND C. GUILLEMINAULT. Hemodynamic studies in sleep apnea. In: Sleep Apnea Syndromes, edited by C. Guilleminault and W. C. Dement. New York: Liss, 1982, p. 177-196. 17. SHEPARD, J. W. Gas exchange and hemodynamics during sleep. Med. Clin. N. Am. 69: 1243-1264, 1985. 18. SHEPARD, J. W. Hemodynamics in Obstructive Sleep Apnea. In: Abnormalities of Respiration During Sleep, edited by E. C. Fletcher.
Orlando, FL: Grune & Stratton, 1986, p. 39-62. M. R., AND F. R. CALARESU. Hypothalamic inhibition of chemoreceptor-induced bradycardia in the cat. Am. J. Physiol. 225:
19. THOMAS,
201-208,1973. 20. TILKIAN, LEHRMAN,
A. G., C. GUILLEMINAULT, J. S. SCHROEDER, K. L. F. B. SIMMONS, AND W. DEMENT. Hemodynamics in sleep-induced apnea. Ann. Intern. Med. 85: 714-719,1976. 21. TOLLE, F. A., W. V. JUDY, P. Yu, AND 0. N. MARKAND. Reduced stroke volume related to pleural pressure in obstructive sleep apnea. J. Appl. Physiol. 55: 1718-1724, 1983. 22. VERRIER, R. L. Behavioral state and Clinical Physiology of Sleep, edited by R.
cardiac arrhythmias. In: Lydic and J. F. Biebuyck. Bethesda, MD: Am. Physiol. Sot., 1988, p. 31-51. (Clin. Physiol. Ser.) 23. WELLENS, H. J. J., A. VERMEULEN, AND D. DURRER. Ventricular fibrillation occurring on arousal from sleep by auditory stimuli. Circulation 52: 73-81, 1975. 24. ZAR, J. H. Biostatistical Analysis. Englewood Cliffs, NJ: PrenticeHall, 1974. 25. ZWILLICH, C., T. DEVLIN, D. WHITE, N. DOUGLAS, J. WEIL, AND R. MARTIN. Bradycardia during sleep apnea. J. Clin. Inuest. 69: 1286-1292,1982.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 12, 2018.
