Blood Flow of the Middle Cerebral Artery With Sleep-Disordered Breathing
Correlation With Obstructive Hypopneas
Background and Purpose—Epidemiological data link heavy snoring to an increased risk for stroke, an association often ascribed to hypertension and/or sleep apnea. The aim of this study was to determine whether obstructive hypopneas, central apneas, or obstructive apneas during sleep alter blood flow of the middle cerebral artery (MCA).
Methods—Doppler sonography of the MCA was performed in conjunction with nightly polysomnography in 11 men and one woman.
Results—A significant decline in blood flow occurred in 76% (169/223) of obstructive hypopneas and in 80% (98/123) of obstructive apneas, compared with only 14% (13/96) of central apneas (P≤.0001). While duration of events was not significantly different, MCA blood flow reductions were associated only with the duration of the obstructive hypopneas (P≤.01) and not with the duration of central (P=.17) or obstructive (P=.07) apneas. The magnitude of fall in arterial oxygen saturation from baseline correlated with a reduced blood flow with obstructive hypopneas but not with obstructive or central apneas.
Conclusions—With obstructive hypopneas and obstructive apneas, MCA blood flow is more often decreased in comparison to central apneas. MCA blood flow reductions occur with longer obstructive hypopneas and with those hypopneas with greater falls in oxygen saturation. These observations indicate pathophysiology relevant to an increased risk for stroke in heavy snorers and patients with obstructive hypopneas and apneas.
The medical consequences of snoring and sleep apnea syndrome are gaining clinical recognition for a possible association with systemic hypertension1 2 3 4 5 and for risk of stroke.6 7 8 9 The theory is that the recurrent periods of asphyxia are the link to these adverse health events.
The earlier observations by Partinen and Palomaki8 and Koskenvuo et al9 have been proved lately by a study of Dyken et al10 (1996), who showed a significantly higher incidence of sleep apnea in patients with stroke than in subjects without stroke.
A study reported in 1993 by Somers et al11 described the influence of rapid sympathetic nerve activity changes linked to muscle tone changes in REM sleep and drew the conclusion that this phenomenon could play a part in triggering ischemic events in patients with vascular disease.
An intermediate factor for stroke might be alterations in cerebral blood flow, but for this variable there are limited observations. Rehan et al12 showed that blood flow velocity either was unchanged or increased during central apneas in healthy term infants. He showed, however, in another observation that blood flow and blood flow velocity were influenced by respiratory patterns but not by sleep stages in these infants.13 Hajak et al14 found that average blood flow showed slight reduction in patients with sleep apnea in wakefulness and stage I NREM sleep compared with healthy control subjects. In REM sleep and NREM stages II through IV, however, they found higher cerebral blood flow in sleep apnea patients than in control subjects. In an earlier report by the same group, Klingelhöfer et al15 reported increased blood flow during apneic episodes in sleep apnea patients and blood flow decrease during episodes with many arousals. Neither study reported the consequences of an individual apnea or hypopnea on blood flow because they were focusing on average blood flow during sleep stages and apneic episodes as well as episodes with many arousals.
In a study of simulated obstructed breaths with Müller maneuvers (high negative intrathoracic pressures against an obstruction), Andreas et al16 demonstrated that during the obstructed effort a significant reduction in blood flow to the MCA occurred, an event that correlated with a fall in blood flow across the mitral and aortic valves. This study in awake humans predicts that blood flow might fall during an obstructive apnea or hypopnea during sleep. According to the observations of Podszus et al,17 systemic blood pressure decreases during obstructive apneas in elderly men.
The present study was planned to observe sleeping subjects for changes in cerebral blood flow resulting from a single apneic or hypopneic event. We postulated that central and obstructive events would differ in regard to changes in cerebral blood flow as a result of the absence or presence, respectively, of negative intrathoracic pressures. Using continuous Doppler sonography, we examined patients with sleep apnea to correlate MCA blood flow with apneas and hypopneas.
Subjects and Methods
Studies were performed in 11 men and 1 woman referred to one sleep center for evaluation of snoring, disturbed sleep, and daytime hypersomnolence. Patients were selected with the use of the following criteria: (1) no disabling cardiopulmonary illnesses, such as chronic obstructive lung disease, congestive heart failure, or asthma; (2) presence of central apneas, obstructive apneas, and obstructive hypopneas on a previous polysomnographic diagnostic study; and (3) willingness to participate. This study was approved by the ethics committee of the University of Freiburg, Germany. Consent was obtained from all patients before the study, as approved by the ethics committee of the University of Freiburg.
All patients were self-reported heavy snorers, with a mean age of 54 years. Patient characteristics are shown in Table 1⇓. As a group, the patients were obese, with a mean Broca index of 128.5%. Hypertension, defined as a blood pressure above 140/90 mm Hg, was present in 5 of 12 patients. There was a range of sleep-disordered breathing; on the average, this group of patients had a moderately severe sleep apnea syndrome, defined as an average apnea index above 30/h.
Sleep and breathing parameters were assessed with the use of 12-channel polysomnography (CNS Sleep Laboratory): C3 and C4 EEG, two-channel electro-oculogram, chin electromyogram, leg electromyogram, nasal and oral airflow assessed by thermistor, ECG, snoring sounds assessed by a microphone placed over the extrathoracic trachea, abdominal and chest wall movement assessed by a respiratory belt, and arterial oxygen saturation assessed by pulse oximeter (Nellcor Monitor).
Blood flow in the MCA was recorded with sonographic equipment (EME type TC2–64) over one of the three sonographic windows of the left or right temporal bone. The device was held in a special head harness,18 modified for use during sleep with a snug headband and straps to hold the position of the sonographic probe. The Doppler signal, representative of velocity and flow in the MCA, was recorded by the CNS Sleep Laboratory computer so that a real-time comparison and analysis of the data could be obtained (Fig 1⇓). A significant reduction in the signal was predetermined to be a 50% reduction in velocity. The criterion for unacceptable measurements was lack of a flow signal, usually the result of movement of the harness relative to the head caused by a change in body position. If this occurred, the probe was either reset to another one of the sonographic windows or repositioned to the other temple. Sleep efficiency was predictably disturbed, and few patients exhibited REM sleep.
At completion of the polysomnography, EEG data together with the Doppler signals were analyzed by inspection of the record.19 Determination of obstructive and central apneas as well as obstructive hypopneas in all 12 patients was accomplished according to the following criteria. Apnea was defined as an absence of inspiratory flow for at least 10 seconds.20 Central apneas were defined as total absence of nasal and oral inspiratory flow and chest/abdominal movement during the period of absence of inspiratory flow. Obstructive apneas were defined by a decrease of greater than 80% in airflow in the presence of paradoxical movements of the ribcage and abdomen. Obstructive hypopneas were defined as a 50% to 80% reduction in airflow accompanied by lack of synchrony of ribcage and abdominal motion.
A reduction in the blood flow was defined as a 50% or more reduction in amplitude of the signal. Results of this decision were matched separately with the type and duration of apnea and with the decrease in arterial oxygen saturation from baseline value with each abnormal breathing event. A representative sample of the record is shown in Fig 1⇑.
Results are expressed as mean and SD. Differences between groups and correlation with variables such as duration of apnea, desaturation from baseline, and incidence of decreased blood flow of the MCA were analyzed with the use of the Wilcoxon signed rank test and, when indicated, Friedman two-way ANOVA (logistic regression analysis). Differences with values of P≤.05 were considered statistically significant.
The tables present the number of apneas examined for each patient in regard to the duration of events (Table 2⇓) and the change in oxygen saturation with an event (Table 3⇓) in which there was a technically acceptable blood flow signal recorded along with all other polysomnographic variables. The tables demonstrate that a range of event duration and desaturation with events was observed in each patient but that the number of observations per patient was limited. There was a difference in regard to the technical ability to record MCA flow among patients that depended on the Apnea/Hypopnea Index. With more affected individuals, movements and arousals occurred more frequently than in those with less sleep disturbances. Given the selection criteria (see “Subjects and Methods”), observations were made of all types of events in all patients; however, given the circumstances of the study, the number of observations per patient was limited and too small to perform correlations for each subject or in regard to sleep stage. Very little REM sleep was observed, and as a result analyses were confined to events in NREM sleep. There were no differences between patients with or without hypertension. Thus, results were pooled across all patients according to event type (obstructive apnea, obstructive hypopnea, or central apnea).
In regard to event type, all data combined showed a significant (>50%) reduction in MCA blood flow (P<.0001); however, the frequency of this varied by event type. A reduction of MCA blood flow was observed in 98 of 123 obstructive apneas (80%) and 169 of 223 obstructive hypopneas (76%) (Fig 2⇓, Table 4⇓), but only 13 of 96 central apneas (14%) showed a reduction in MCA blood flow (Fig 2⇓, Table 4⇓). This difference in the frequency of a reduced blood flow of the MCA between central and obstructive apneas and hypopneas was significant. The level of reduced blood flow during obstructive apneas was not significantly different from that seen during obstructive hypopneas.
In regard to the duration of events, reduced MCA blood flow correlated with all apneas (P<.01). Examined separately, a correlation was significant with the duration of the hypopneas (P<.01) but not with the duration of the obstructive apneas (P=.07) or with the duration of central apneas (P=.17) (Fig 3⇓). This was not the result of event length, since the mean duration of events, either associated or not associated with a reduction in MCA flow, was similar for the hypopneic events (18.1±6.5 seconds), the central apnea group (17.2±5.9 seconds), and the obstructive apnea group (14.8±5.0 seconds) (P=.3).
In regard to the change in oxygen saturation with events, there was a statistically significant correlation (P<.05) between the fall in arterial oxygen saturation and the occurrence of reduced MCA blood flow in obstructive hypopnea (Fig 4⇓). No such correlation existed for all events (P=.1) or for obstructive apneas (P=.92) or central apneas (P=.22) (Fig 4⇓). The mean decline in oxygen saturation was similar for central apneas (10.3±11.7%), obstructive hypopneas (8.9±9.7%), and obstructive apneas (7.3±6.8%) (Table 4⇑). Similar conclusions were reached when we considered only the events leading to a decreased perfusion of the MCA.
This observational study demonstrates a correlation between obstructive apneas and hypopneas and the probability of an episodic reduction in MCA blood flow in NREM sleep. The occurrence of blood flow reduction increases with the duration of obstructive hypopnea and the fall in oxygen saturation with obstructive hypopnea. In central apneas, a reduction in MCA blood flow was less frequent. A reduction in blood flow occurred during the event and returned to baseline at the termination of the apnea.
Both Partinen and Palomaki8 and later Koskenvuo et al9 showed a strong epidemiological correlation between loud snoring and the risk of stroke development. For instance, in one study the unadjusted risk for stroke was 40 times higher in heavy snorers than in nonsnorers.8 Causality could not be clearly elucidated from these studies; however, it seemed reasonable to the authors to assume a relation to effects of sleep-disordered breathing. Subsequent studies examining nocturnal blood pressure in sleep apnea patients,2 3 4 as well as the study by Fischer et al21 measuring intracranial hemodynamics in sleep apnea, suggest a direct effect of each sleep-related respiratory event on blood pressure and cerebral perfusion, respectively. Other pathophysiological pathways potentially operative include hypoxia-induced vasodilatation22 or reflex-mediated cerebral vasoconstriction in the MCA from acute systemic hypertension,23 mechanisms identified as causes of acute stroke.
Fischer et al21 documented a mean reduction of blood flow to the MCA during REM sleep, NREM sleep, and on awakening in patients suffering from sleep apnea syndrome; however, no temporal relationship between the presence or type of apnea was reported. The results of the present study establish a direct relationship between individual apneas and reduction in MCA blood flow (Fig 2⇑). This real-time correlation of blood flow with apneic events is not currently possible to achieve with alternative methods of determining blood flow based on metabolic tracers or spin resonance imaging. Based on the data obtained, obstructive apneas and hypopneas compared with central apneas lead much more frequently to a reduction in cerebral blood flow, as assessed by the qualitative method of MCA Doppler ultrasonography, during NREM sleep.
Duration of the event plays a role in the development of reduced perfusion of the MCA, ie, the longer the event, the greater the likelihood for a reduction in blood flow. This was particularly observed for obstructive hypopneas, the physiological marker of self-reported heavy snoring. A weak association exists between the degree of desaturation and the frequency of decreased blood flow of the MCA, and again this is true only for obstructive hypopneas. These observations are unlikely to be explained by differences in patient physiology, since each patient demonstrated all event types. However, it must be noted that within all analyzed events a duration longer than 22 seconds as well as an oxygen desaturation greater than 15% was present in obstructive hypopneas more than in obstructive apneas. This might be explained by recruiting a patient group that presents all types of apneas, so that obstructive hypopneas may be present in this patient group more than in a more typical patient group with obstructive sleep apnea and mainly obstructive apneas only. Because the majority of obstructive hypopneas are in the group of events with a duration longer than 22 seconds and this possibly interferes with the fall in arterial oxygen saturation in this group of events, it gives the impression that oxygen desaturation is less likely to be present in obstructive apneas.
An explanation for the relationship between partial airway obstruction and decreased perfusion of the MCA might be provided by the observation of Andreas et al,16 who were able to demonstrate reduced blood flow during Müller maneuvers. Our study did not measure pleural pressures or cardiac output and used the patterns of change in air flow and chest wall motion as indicators of central or obstructive events.
Our observation that a reduction in blood flow of the MCA occurred more often with a longer duration of hypopnea suggests that there is an interaction between respiratory efforts and the duration of the event in regard to their impact on cerebral blood flow and on systemic blood pressure.17 24 25 26 Increased time of partial obstructions might lead to a greater likelihood for development of a high cardiac preload, a lower cardiac afterload, activation of carotid body baroreceptors, and vasodilatation by both increasing arterial carbon dioxide and decreasing oxygenation, all of which can contribute to a reduction in cerebral blood flow. Termination of the hypopnea with an unobstructed breath, however, rapidly results in a return to pre-event levels of blood flow, suggesting that not all of these mechanisms are operative. Termination may protect from decreasing blood flow resulting from reflex or the direct effects of asphyxia. Thus, increased blood flow velocity could be related to arousals or to arousal threshold; however, in contrast to the previous conclusions of Klingelhöfer et al,15 arousals are not the reason for blood flow reductions.
We focused in our study on single events rather than on episodes or sleep stages, and this may also explain the difference between our findings and those of Klingelhöfer et al,15 ie, our results do not suggest an increase in cerebral blood flow during apneic events.
If negative intrathoracic pressures are indeed a major cause of the decrease in MCA blood flow, one would not expect a decrease in MCA perfusion during central apneas. This conclusion, namely that central apneas are not associated with changes in cerebral blood flow, was reached by Rehan et al12 in a study of infants. We observed that reductions in MCA flow occurred much less frequently in central than in obstructive events: 14% compared with 70% to 80%. By strict definition, no inspiratory efforts and thus no negative inspiratory pressure developments occur during central apneas. However, misclassification of apnea type can occur in the absence of direct measures of intrathoracic pressure.27 While such a misclassification is less likely when both ribcage and abdominal motions are measured, we acknowledge that upper airway obstructions may have occurred and could be responsible for reduced MCA blood flow found in the minority of central events. An alternative explanation is that the decreased blood flow in a central apnea possibly reflects hypoxic/hypercapnic cerebral vasodilatation; however, we found only a weak significant correlation with MCA blood flow and the fall in arterial oxygen saturation in obstructive hypopneas. This observation may be influenced by the greater number of hypopneas analyzed with longer duration and greater fall in arterial oxygen saturation; in addition, a greater number of analyzed obstructive apneas with oxygen desaturation would have led to a stronger correlation between the fall in oxygen saturation and decreased MCA blood flow.
We were unable to capture enough events to analyze differences between sleep states. Perhaps with more advanced technology there will be an opportunity to observe the effect of event type on cerebral blood flow in regard to REM as well as NREM sleep. There may well be differences between the two states, given the specific and different influences of REM sleep on cardiovascular and neurophysiological function.
Based on our data, one could speculate that patients with central sleep apneas should have the least risk for the development of stroke, and those with a predominance of partial upper airway obstruction during sleep (heavy snoring) would have the greatest risk. Epidemiological studies like those of Partinen and Palomaki8 and Koskenvuo et al9 documented a significantly increased risk for stroke in those with self-reported heavy snoring, but further distinctions in regard to respiratory events were not available. Others have jumped to the conclusion that apneas rather than snoring per se are responsible for the observed associations, but our data suggest that obstructive hypopneas, the physiological correlate for heavy snoring, have at least as great or an even greater significance than either central or obstructive apneas. Individuals with predominantly central events would be predicted to exhibit less risk for stroke; however, such individuals are rarely found, except when central apneas occur in the setting of heart failure. A prospective study would be one way to determine the relative risk for the development of stroke in regard to the type of sleep-disordered breathing.
Selected Abbreviations and Acronyms
|MCA||=||middle cerebral artery|
|NREM||=||non–rapid eye movement|
|REM||=||rapid eye movement|
We thank Prof Dr Schulte-Mönting, Institute for Biomathematics, University of Freiburg, for statistical analysis of our data, and Eugene C. Fletcher, MD, Professor of Medicine, Louisville, Ky, for his advice and reading of the manuscript.
Reprint requests to Nikolaus Netzer, MD, Case Western Reserve University, Department of Medicine, Division of Pulmonary and Critical Care Medicine, VA Medical Center 111 J (W), 10701 East Blvd, Cleveland, OH 44106.
- Received July 15, 1997.
- Revision received October 10, 1997.
- Accepted October 10, 1997.
- Copyright © 1998 by American Heart Association
Fletcher EC, DE Behnke RD, Lovoi MS, Gorin AB, Undiagnosed sleep apnea in patients with essential hypertension. Ann Intern Med. 1985;103:190–195.
Norton PG, Dunn EV. Snoring as a risk factor for disease: an epidemiological survey. BMJ. 1985;291:630–632.
Koskenvuo M, Kapnio J, Talakivi T, Partinen M Heikkila K, Sarna S. Snoring as a risk factor for stroke in men. Br Med J. 1987;294:16–19.
Dyken ME, Somers VK, Yamada T, Ren ZY, Zimmerman MB. Investigating the relationship between stroke and obstructive sleep apnea. Stroke.. 1996;27:401–407.
Klingelhöfer J, Hajak G, Sander D, Schulz-Varszegi M, Rüther E, Conrad B. Assessment of intracranial hemodynamics in sleep apnea syndrome. Stroke. 1992;23:1427–1433.
Podszus T, Koehler U, Mayer J, Penzel T, Peter JH, Wichert P. Systemic arterial blood pressure decreases during obstructive sleep apnea. Sleep Res. 1985;15:155.
Aslid R, ed. Examination Techniques in Transcranial Doppler Sonography. New York NY: Springer; 1986;39–56.
Rechtschaffen A, Kales A. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. Los Angeles, Calif: BIS/BRIUCLA; 1968.
Neubauer JA,Bisgard GE. Peripheral and central effects of hypoxia. In: Dempsey JA, Pack AI, eds. Regulation of Breathing. New York, NY: Marcel Dekker; 1994:617–668.
Guyton AC. Cerebral blood flow, the cerebrospinal fluid and brain metabolism. In: Guyton AC, ed. Basic Neuroscience: Anatomy and Physiology. 2nd ed. Philadelphia, Pa: WB Saunders; 1992:285–295.
Mayer J, Greb H, Herres B, Kloss TM, Penzel JH, Podszus T, von Wichert P. Nocturnal hemodynamics in patients with sleep apnea. In: Peter JH, Podszus T, von Wichert P, eds. Sleep Related Disorders and Internal Diseases. Berlin, Germany: Springer; 1987:315–320.
Krieger BP. Expanded options for monitoring the mechanical ventilated patient. Semin Respir Med. 1993;14:284–292.