INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: blood flow; blood pressure; Doppler; muscle sympathetic nervous system; obstructive sleep apnea; vasoconstriction

Obstructive apneas during sleep in patients with the obstructive sleep apnea syndrome (OSA) are associated with transient increases in arterial pressure, which coincide with oxyhemoglobin desaturation (1). Because these pressor responses are preceded by surges of muscle sympathetic nerve activity (MSNA), it has been postulated that transient peripheral vasoconstriction occurs (4). Alternatively, these transient pressor responses could also be due to an increase in cardiac output when breathing resumes after apnea.

Using the thermodilution technique, Guilleminault and coworkers examined the cardiac output response during apnea- hyperpnea cycles in OSA (7). These investigators demonstrated a reduction in cardiac index from baseline during apneic events and an increase above baseline when breathing resumed. This suggested to these authors that a transient increase in cardiac output contributed to the observed postapnea blood pressure elevations (7). However, because of the limited time resolution of the thermodilution technique, a precise assessment of such transient hemodynamic events is difficult. More recently, Garpestad and coworkers continuously monitored left ventricular stroke volume, using a radioisotope-based ambulatory ventricular function monitor during nonrapid eye movement sleep (NREM) in OSA (8). These data suggested that the transient hypertension at apnea termination was associated with a decrease in left ventricular stroke volume. Although the heart rate increased at apnea termination, it was insufficient to compensate for the decrease in left ventricular stroke volume, resulting in a decrease in cardiac output (8). Stoohs and Guilleminault examined the hemodynamic changes during apnea with an electrical impedance method and found left ventricular stroke volume to decrease during apnea. However, chest motion in the hyperventilatory phase of the apnea- hyperpnea cycle interfered with the measurement of stroke volume, limiting their measurements to the apnea phase (9). Thus, it is not clear what role changes in blood flow play in mediating the prominent apnea-induced pressor response. In addition to changes in cardiac output, the pressor response at apnea termination may be due to peripheral vasoconstriction. Surprisingly, little is known about the effects of obstructive apnea on peripheral limb blood flow in OSA.

To further elucidate the mechanism underlying the transient blood pressure elevation following obstructive apnea, we simultaneously measured MSNA, limb blood flow by Doppler technique, and arterial pressure in obstructive apneas occurring during sleep. Changes in peripheral vascular resistance were calculated from the relationship of mean arterial pressure divided by blood flow. Our hypothesis was that the apnea-induced arterial pressure elevations are due to sympathetically mediated vasoconstriction. Furthermore, we reasoned that if apnea-induced vasoconstriction was attenuated by the administration of supplemental oxygen (O₂), then hypoxia would be implicated as a trigger for this response.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

We studied 11 patients (8 men, 3 women) with documented OSA. The patients were recruited through our sleep disorders clinic (Hershey, PA) and their characteristics are shown in Table 1. The studies were performed in a quiet and dimly lit clinical research laboratory during the daytime hours, with the patients in the supine position and in a postabsorptive state. Chronic medications were not withheld. Written informed consent was obtained. The protocol had been approved by the Institutional Review Board at the Milton S. Hershey Medical Center.

Blood Pressure and Heart Rate Measurements

Throughout the apnea events, arterial pressure was measured on a beat-by-beat basis by finger photoplethysmography, using a Finapres device (Ohmeda, Madison, WI). This device has been shown to reliably register transient blood pressure changes (10). Care was taken to fit the finger with an appropriate cuff and to stabilize the hand at the level of the heart. The device was turned off every 30 min for several minutes to prevent finger edema, which might interfere with the accuracy of finger blood pressure readings. Baseline arterial pressure measurements obtained with the Finapres device were validated with an automated sphygmomanometer (Dinamap; Critikon, Tampa, FL). Mean arterial pressure () was calculated as diastolic pressure plus one-third of pulse pressure. The heart rate (HR) was derived from the electrocardiogram (ECG) or from the arterial pressure curve.

Microneurography

Peroneal microneurography was used to determine MSNA as described previously (5, 11). The microneurographic recordings were made with tungsten microelectrodes that were inserted percutaneously into peripheral nerves to record multiunit action potentials from C fiber-containing nerve fascicles. These signals are thought to reflect postganglionic vasoconstrictor nerve traffic (12). A suitable recording site for MSNA was obtained when spontaneous pulse-synchronous bursts with a greater than 2:1 to 3:1 signal-to-noise ratio were observed and nerve traffic directed to skin was excluded by the absence of responses to arousal stimuli and to stroking of the skin. The nerve traffic signal was filtered, amplified, and integrated and was recorded on a TA 4000 recorder (Gould, Valley View, OH). Hard copies were analyzed by hand. The data were expressed as bursts per minute and as the total amplitude of the bursts per minute (arbitrary units).

Limb Blood Flow and Vascular Resistance

To estimate changes in brachial or femoral blood flow during obstructive apnea we used the Doppler technique as described previously (13, 14). This technique permits a beat-by-beat measurement of mean blood velocity (MBV) and thereby allows the detection of rapid and transient changes of vascular resistance. Measurements of MBV were obtained with continuous-wave ultrasound (4 MHz, Multigon 500 M; Multigon Industries, Yonkers, NY) using a flat 4-MHz ultrasound probe with an angle of insonation fixed at 45° relative to the skin. With audio feedback, the Doppler probe was adjusted manually over the brachial artery (~ 2 cm proximal to the antecubital crease) or the femoral artery (~ 2 cm distal to the inguinal ligament) to yield flow signals with a maximal Doppler frequency shift. Instantaneous MBV (cm/s) was determined from the Doppler spectra generated by a Doppler signal processor and was recorded at 100 Hz together with the arterial pressure signal and the ECG. The Doppler recordings were later analyzed with a Macintosh computer equipped with a MacLab analysis system (ADInstruments, Castle Hill, Australia).

Because in a prior study femoral artery diameter remained constant during transient blood flow and pressor responses to voluntary apnea (14), we based our estimation of blood flow on MBV alone. Assuming a flat, nonturbulent velocity profile and cylindrical shape of the artery, changes of volumetric flow can than be estimated from changes in MBV (15). The relationship

was used to estimate apnea-induced changes of limb vascular resistance.

Skin blood flow was estimated with a laser Doppler flow probe (BPM 403 TSI; Laser Flow, St. Paul, MN) (16), which was placed on the extremity where arterial blood flow was measured as described previously (17). Skin blood flow was expressed in relative units.

Ventilatory Parameters

Chest movements were recorded with abdominal and chest strain gauges. Arterial O₂ saturation (Sa_O2%) was monitored with an ear oximeter connected to an Ohmeda respiratory gas monitor (model 5250; Ohmeda) with a response time of 8 s.

Sleep Monitoring

Polysomnography was performed in a standard fashion and in accordance with recommendations by the American Thoracic Society (18). These recordings were made with a Medilog SAC 847 system (Oxford Instruments, Oxford, UK) and included a two-channel electroencephalogram, electro-oculogram, submental electromyograms (surface electrodes), as well as nasal/oral airflow probes (thermocouples). Apnea was defined as an absence of airflow for > 10 s. Airway obstruction was documented by the presence of inspiratory effort (chest/abdominal strain gauge) and the absence of airflow (18). Sleep stage was assessed according to the criteria of Rechtschaffen and Kales (19).

Experimental Protocol

Once the polysomnographic electrodes were in place and a suitable site to record MSNA was found, the Doppler probe was positioned over the brachial artery in the antecubital fossa (n = 9), or over the femoral artery ~ 2 cm below the inguinal ligament (n = 2). The skin blood flow probe was placed on the anterior aspect of the lower extremity or on the volar aspect of the forearm (n = 4). After baseline measurements over a 5-min period of quiet wakefulness, the subjects were allowed to sleep. Mean arterial pressure, HR, MBV, and MSNA were again recorded during 10-20 obstructive apneas. Supplemental O₂ was then administered via a nonrebreather face mask at 15 L/min and spontaneous apneas during O₂ supplementation were observed. Thereafter the patients were awakened and were asked to perform five separate 20-s end-expiratory voluntary apneas each on room air and during O₂ supplementation. The patients were instructed to avoid Valsalva or Müller maneuvers.

Data Analysis and Statistics

For simplicity, all obstructive and voluntary apnea events were divided in half: from the beginning of apnea to the midpoint in the apnea cycle was termed early apnea (early). From the midpoint to the point just before the resumption of breathing was termed late apnea (late). The point of resumption of breathing to 3-5 beats after the blood pressure peak was termed postapnea. The apnea responses were analyzed as follows: , MBV, and HR and MSNA were averaged over three representative cardiac cycles at the end of the early, late, and postapnea phases, respectively. The postapnea measurements were made during the time of the blood pressure peak. Apnea events that were accompanied by technically unsatisfactory MSNA, Doppler, or arterial pressure tracings were excluded from analysis. One-way analysis of variance was performed to determine differences in these parameters between early apnea, late apnea, and postapnea. To examine differences between voluntary apneas during room air exposure versus O₂ supplementation, we used two-way analysis of variance to test for effects of apnea and level of oxygenation. Post hoc analysis was performed with the Tukey test. Linear correlation analysis was used to assess the relationship between AHI or BMI and the pressor or vasoconstrictor responses to obstructive apnea. The data are presented as means ± standard error (SE). A p value < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Neurocirculatory Effects of Obstructive Apneas during Room Air Exposure

Spontaneous obstructive apneas while exposed to room air occurred in seven subjects (see Table 1 for patient characteristics). Twenty-six obstructive apneas (duration, 19 ± 3 s; Sa_O2 nadir, 88 ± 3%) were analyzed. All apneas analyzed occurred during NREM sleep. A typical recording from one patient with OSA is shown in Figure 1 and illustrates the typical sequence of events that occurred during the apnea-hyperventilation cycle. In general, during obstructive apnea Sa_O2 decreased progressively and MSNA rose. Arterial pressure decreased in the early phase of apnea and increased toward the end of apnea whereas MBV did not change significantly. Resumption of breathing (postapnea) was associated with a sudden marked decline in MSNA, and a transient surge in arterial pressure and heart rate. However, despite the rise in pressure, MBV did not increase. From early to late apnea, MSNA amplitude increased by 232 ± 108% (p < 0.05). From late to postapnea, increased by 18 ± 2% (p < 0.001), HR increased by 19 ± 8% (p = 0.058), and vascular resistance rose by 29 ± 8% (p < 0.002). Thus, the rise in coincided with a rise in vascular resistance. No significant differences were noted in the neurocirculatory responses to obstructive apnea between the subjects who were taking vasoactive medications and those who were not. In addition, in our small patient sample, we found no significant correlation between the severity of OSA as expressed by the apnea-hypopnea index (AHI) or between the body mass index (BMI) and the pressor (r² = 0.02 and r² = 0.17, respectively) or vasoconstrictor (r² = 0.07 and r² = 0.001, respectively) responses to obstructive apnea. The effect of obstructive apneas on MSNA, , MBV, and vascular resistance are shown in Figure 2.

View larger version (48K): [in this window] [in a new window]   Figure 1.   Recording of electroencephalogram (EEG), electromyogram (EMG), integrated muscle sympathetic nerve activity (MSNA), systemic blood pressure (BP), mean brachial blood velocity (MBV), arterial oxygen saturation (SaO2), and respiration in a patient with obstructive sleep apnea during sleep. Arrows indicate the onset of arousal. The lag between onset of apnea and oxygen saturations nadir reflects normal circulatory delay.

View larger version (20K): [in this window] [in a new window]   Figure 2.   Responses of muscle sympathetic nerve activity (MSNA; A), mean arterial pressure (Psa [MAP]; B), mean blood velocity (MBV, C ), and limb vascular resistance (D) to obstructive apnea during sleep. Early, late, and post refer to the phase of the apnea cycle.

Skin blood flow (SBF) measurements during obstructive apneas were obtained in three patients. In all three subjects SBF increased progressively during apnea and peaked during the early postapnea period in parallel with the rise in blood pressure and in contrast to total limb blood flow, which did not change during the postapnea period. Compared with late apnea, in these three subjects, SBF increased by 33, 81, and 100%, respectively in the postapnea period.

Neurocirculatory Effects of Obstructive Apneas during Administration of Supplemental O₂

Only two of seven patients who received supplemental O₂ experienced obstructive apneas during O₂ exposure (see Table 1 for patient characteristics). Five of seven were unable to sleep. Five spontaneous obstructive apneas in one subject and four apneas in the other subject were analyzed (duration, 32 ± 3 s; Sa_O2 nadir, 99 ± 0.3%). Compared with obstructive apneas while exposed to room air, in these subjects supplemental O₂ resulted in the following responses in , MBV, and vascular resistance from late to postapnea (room air versus O₂: , +16 versus +16%; MBV, 0.0 versus 35%; resistance, +18 versus +8% for Subject 1 and , +19 versus 9%; MBV 15 versus +6%; resistance +53 versus +17% for Subject 2). In addition, compared with room air exposure, during O₂ supplementation, the increase in MSNA amplitude from early to late apnea was attenuated in these subjects (+73 versus +50% and +150 versus 0.0% for Subjects 1 and 2, respectively).

Neurocirculatory Effects of Voluntary Apnea

Ten subjects performed voluntary apneas. Fifty-nine voluntary apneas during room air exposure and 47 apneas during supplemental O₂ administration were analyzed. The responses of MSNA, , MBV and vascular resistance to voluntary apneas are shown in Table 2. The patterns of these responses were similar to those of obstructive apneas. From early to late apnea, MSNA on room air increased by 129 ± 80% (p < 0.02). From late to postapnea, rose by 13 ± 5% (p < 0.002) whereas MBV was unchanged. Thus, from late to postapnea vascular resistance increased by 25 ± 7% (p < 0.001). Supplemental O₂ resulted in a significant attenuation in the vasoconstrictor responses (p < 0.02).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study we examined the mechanism of the transient pressor responses associated with obstructive apnea during sleep. To this end we simultaneously measured blood pressure, sympathetic nerve activity (MSNA), and limb blood velocity as an index of limb blood flow. From the changes in blood pressure and blood velocity we calculated transient apnea-induced changes in limb vascular resistance. Our principal finding was that despite a transient elevation of arterial pressure and heart rate in the immediate postapnea period, there was no commensurate increase in limb blood velocity. This suggests that obstructive apnea is associated with transient vasoconstriction. Because the pressor response was preceded by discharges of vasoconstrictor sympathetic nerves, it is likely that the apnea-induced transient vasoconstriction is at least in part mediated by the sympathetic nervous system. The pressor and vasoconstrictor responses were similar in patients with OSA independent of BMI and AHI and were also similar to those observed in healthy subjects who performed the maximal voluntary apnea maneuver while exposed to hypoxia (14). In this section we discuss the study rationale, potential explanations for our findings, and potential limitations.

Repetitive pressor responses and hypoxia during sleep in OSA may have adverse effects on the cardiovascular system. Therefore, understanding the mechanism(s) underlying these responses is of considerable interest. In a series of prior studies, investigators have attempted to define the hemodynamic events associated with obstructive apnea. On the basis of these reports, the factors responsible for the acute neurocirculatory responses to obstructive apnea are thought to include hypoxia, hypercapnia, and arousal (20, 21). Some authors also believe that the swings in intrathoracic pressure that occur during obstructive apneas may play a major role in the blood pressure increase postapnea through an increase in cardiac output (9). From a pathophysiologic point of view, the rise in blood pressure at apnea termination could result from a rise in blood flow (cardiac output) or from increased vascular resistance. On the one hand, a transient increase in cardiac output postapnea would be plausible because the heart rate rises at apnea termination. On the other hand, surges of sympathetic vasoconstrictor nerve traffic at the end of apnea would be expected to lead to transiently increased vascular resistance. However, release of vasodilator metabolites during apnea-induced hypoxia could counteract vasoconstrictor nerve signals.

Our simultaneous measurements of arterial pressure and limb blood flow velocity allowed us to estimate changes in limb vascular resistance associated with obstructive apnea. In contrast to Garpestad and coworkers, who demonstrated a decrease in cardiac stroke volume at apnea termination despite an increase in heart rate (8), we found that although blood pressure rose as expected, limb blood flow velocity, an index of volumetric limb blood flow, did not change postapnea. In their report, cardiac output decreased by ~ 26% at apnea termination compared with preapnea. This discrepancy may be due to the greater length of apneas, and severity of O₂ desaturations in the study by Garpestad and coworkers (8). Alternatively, the responses of limb flow and cardiac output to obstructive apnea may differ. However, in clear agreement with Garpestad and coworkers (8), our data strongly suggest that obstructive apnea is associated with transient vasoconstriction. Our principal conclusion of apnea-induced vasoconstriction is also consistent with an observation by Schnall and coworkers (22), who found that transient decreases of pulsatile blood volume correlated with obstructive apneas.

In a small subgroup of our patients with OSA, we found skin blood flow in the limb to increase in parallel with the postapnea rise in blood pressure, thereby attenuating the apnea-induced rise in limb vascular resistance. Because skin blood flow in the limb may represent up to 10% of total limb blood flow (23), we believe that our Doppler data actually underestimate the magnitude of the apnea-induced vasoconstriction present in skeletal muscle.

What triggers the transient vasoconstriction induced by obstructive apnea? A series of prior investigations demonstrated that obstructive apnea leads to sympathetic neural excitation (4) and suggest that this is in large part due to hypoxia (5, 11, 24, 25). Studies by Schroeder and coworkers (26) and Shepard (27) demonstrated that the transient elevations of arterial pressure after apneas were directly related to the severity of the apnea-induced O₂ desaturation and were blunted by supplemental O₂ (26). In another study, when normal awake subjects performed repetitive breathholds, the sympathetic neural and pressor responses that accompanied these maneuvers were augmented during administration of a hypoxic gas mixture and attenuated during hyperoxia (11). Last, the sympathetic neural and blood pressure responses to obstructive apnea during sleep have been shown to be attenuated during administration of 100% O₂ (5). In support of these published reports, in two of our patients, supplemental O₂ attenuated the pressor and sympathetic responses to apnea. In addition, the vasoconstrictor responses to obstructive apnea were also attenuated. Taken together, these published reports and the findings from the present study are consistent with the hypothesis that the neural and vasoconstrictor responses to obstructive apnea may be linked to hypoxia.

Because swings in intrathoracic pressure that occur during obstructive apneas may contribute to their neurocirculatory consequences (9), we also examined the effects of voluntary apneas during wakefulness. In agreement with the findings of other investigators (21), we found that the responses to both spontaneous obstructive apneas and to voluntary apneas were qualitatively similar. Because we did not monitor intrathoracic pressure, our data do not allow us to comment definitively on potential contributing effects of the wide fluctuations of intrathoracic pressure commonly observed during obstructive apnea. However, in healthy humans the Müller maneuver (forced inspiration against a closed glottis) did not affect the blood pressure and sympathetic responses to voluntary apnea (21). Furthermore, in a porcine model of central versus obstructive apnea, the profound intrathoracic pressure fluctuations induced by obstructive apnea had no additional effect on the vasoconstrictor effects of apnea and the associated hypoxia, hypercapnia, and acidosis (28). On the basis of these considerations it appears unlikely that intrathoracic pressure changes act as a crucial determinant of these responses.

The qualitative similarities between the responses to obstructive and voluntary nonobstructive apneas also suggest indirectly that arousal, which did not occur during voluntary apneas, may not be necessary to evoke postapnea pressor and vasoconstrictor responses. However, arousal has been shown to result in mild transient increases of sympathetic activity and blood pressure (20), and in our study we did not examine the neurocirculatory effects of arousal alone. Therefore, we cannot rule out that arousal alone or in combination with chemical stimuli may contribute to the pressor effects of obstructive apnea.

In addition to sympathoexcitation, it is possible that engagement of the myogenic reflex could contribute to the observed peripheral vasoconstriction. This reflex can evoke vasoconstriction when perfusion pressure rises (29). It is also known that the myogenic reflex can be enhanced by heightened sympathetic tone (30). Therefore, the rise in blood pressure coupled with the neural release of norepinephrine may augment the magnitude of vasoconstriction, which in turn further elevates blood pressure. However, because sympathetic activation and increased intravascular pressure may act in concert to raise vasoconstrictor tone, the relative contribution of neurogenic and myogenic vasoconstrictor influences is unclear.

Compared with the large apnea-related surges of sympathetic activity, the responses of blood pressure and vascular resistance were relatively small. In this context, it is important to note that during apnea events, both vasoconstrictor mechanisms such as sympathoexcitation and/or the myogenic reflex, and nonneural vasodilator mechanisms (e.g., release of metabolic vasodilators during hypoxia) may be activated simultaneously (31). However, because hypoxia-induced vasodilation may be impaired in OSA (32), we doubt that augmented activity of vasodilator systems contributed to the relatively modest increases in vascular resistance and blood pressure.

Several important limitations of our study need to be considered. We did not observe rapid eye movement (REM) sleep in any of our subjects. Therefore, our findings cannot be extrapolated to obstructive apneas that occur during REM sleep. Despite our efforts to study prolonged periods of sleep, the physical restraint caused by the extensive instrumentation prevented the extended observation necessary to observe REM sleep. Because REM sleep is accompanied by accentuated hemodynamic fluctuations and larger sympathetic surges than those seen during lower stages of sleep (33, 34), the apnea- related vasoconstrictor responses might be even greater during REM sleep compared with those observed during lighter stages of sleep.

We did not monitor arterial diameter throughout obstructive apnea events. An increase in arterial diameters during apnea could attenuate an increase in MBV if limb blood flow increased and could thus lead to "apparent" vasoconstriction. Continuous and accurate diameter measurements in these often obese subjects would be difficult. However, in our experience, in healthy (nonobese) subjects femoral artery diameter does not change during moderate to severe hypoxia compared with baseline conditions and is unaffected by sympathetic and blood pressure responses to voluntary apnea (14). For our purposes we therefore assumed that arterial diameter remained constant. Although we may not be able to determine the precise magnitude of the changes of flow and resistance, we contend that our data are clear regarding the direction of change, namely that vasoconstriction (rather than vasodilation) does occur in response to obstructive apnea.

Although our data suggest that the apnea-induced vasoconstriction is sympathetically mediated, they do not prove a causal link between sympathetic activation and vasoconstriction. Proof of this concept would require that sympathetic blockade abolished the apnea-induced vasoconstriction. In this regard, Katragadda and coworkers examined normal awake individuals before and after sympathetic blockade (35). In their study, ganglionic blockade with trimethaphan completely abolished the MSNA and pressor responses to simulated obstructive apneas (voluntary apneas with Müller maneuver) and voluntary apneas (35).

In conclusion, our findings suggest that the transient pressor response observed after obstructive apnea during NREM sleep is due to transient vasoconstriction, which is in part mediated by hypoxia-induced surges of vasoconstrictor nerve traffic. We speculate that transient vasoconstriction, hypertension, and hypoxia in association with obstructive apnea may contribute to myocardial ischemia, left ventricular hypertrophy, and cardiac dysfunction commonly seen in OSA.