Cerebrovascular Response to Carbon Dioxide in Patients with Congestive Heart Failure
http://www.100md.com
美国呼吸和危急护理医学 2005年第8期
Departments of Medicine, Orthopedics and Rehabilitation
Population Health Sciences, University of Wisconsin
the William S. Middleton Memorial Veterans Hospital, Madison, Wisconsin
ABSTRACT
Rationale: Cerebrovascular reactivity to CO2 provides an important counterregulatory mechanism that serves to minimize the change in H+ at the central chemoreceptor, thereby stabilizing the breathing pattern in the face of perturbations in PaCO2. However, there are no studies relating cerebral circulation abnormality to the presence or absence of central sleep apnea in patients with heart failure. Objectives: To determine whether patients with congestive heart failure and central sleep apnea have an attenuated cerebrovascular responsibility to CO2. Methods: Cerebral blood flow velocity in the middle cerebral artery was measured in patients with stable congestive heart failure with (n = 9) and without (n = 8) central sleep apnea using transcranial ultrasound during eucapnia (room air), hypercapnia (inspired CO2, 3 and 5%), and hypocapnia (voluntary hyperventilation). In addition, eight subjects with apnea and nine without apnea performed a 20-second breath-hold to investigate the dynamic cerebrovascular response to apnea. Measurements and Main Results: The overall cerebrovascular reactivity to CO2 (hyper- and hypocapnia) was lower in patients with apnea than in the control group (1.8 ± 0.2 vs. 2.5 ± 0.2%/mm Hg, p < 0.05), mainly due to the prominent reduction of cerebrovascular reactivity to hypocapnia (1.2 ± 0.3 vs. 2.2 ± 0.1%/mm Hg, p < 0.05). Similarly, brain blood flow demonstrated a smaller surge after a 20-second breath-hold (peak velocity, 119 ± 4 vs. 141 ± 8% of baseline, p < 0.05). Conclusion: Patients with central sleep apnea have a diminished cerebrovascular response to PETCO2, especially to hypocapnia. The compromised cerebrovascular reacticity to CO2 might affect stability of the breathing pattern by causing ventilatory overshooting during hypercapnia and undershooting during hypocapnia.
Key Words: central sleep apnea; cerebral blood flow; hypercapnia; hypocapnia
Breathing pattern and brain blood flow are closely linked by PaCO2. Normally, elevated PaCO2 causes vasodilation and increased cerebral blood flow (CBF) (1), which will wash out brain CO2, and consequently reduce H+ at the central chemoreceptor. As a result of the lower H+ at the central chemoreceptor, the ventilatory stimulation to the elevated PaCO2 will be blunted. Conversely, reduced PaCO2 causes vasoconstriction and decreased CBF, which will increase H+ at the central chemoreceptor. The ventilatory inhibition in response to the reduced CO2 will be lessened as a result of the higher H+ at the central chemoreceptors. Thus, the effect of CO2 on CBF provides an important counterregulatory mechanism that serves to minimize the change in H+ at the central chemoreceptor, thereby stabilizing the breathing pattern in the face of perturbations in PaCO2.
In patients with congestive heart failure (CHF), we suspected that reduced baseline CBF and reduced cerebrovascular responsiveness to CO2 could, at least partly, account for the high baseline level of ventilation (2, 3), the high ventilatory response to CO2 (4, 5), and the high prevalence of breathing instability and apnea during sleep. Such a reduction in baseline CBF (6, 7) and the CBF response to CO2 (8) has been demonstrated in patients with CHF as well as in a porcine model of CHF (9). Previous studies have demonstrated an increased ventilation (10eC12) and respiratory sensitivity (13) in response to brain ischemia, indicating the importance of brain blood flow on breathing stability. However, there are no studies relating cerebral circulation abnormality to the presence or absence of central sleep apnea (CSA) in patients with CHF. The purpose of this study was to determine whether patients with CHF and CSA have an attenuated cerebrovascular responsibility to CO2 compared with patients with CHF but without CSA. Such an abnormality may predispose to periodic breathing during sleep by both narrowing the difference between the eupneic PCO2 and apneic threshold (14) and by increasing the ventilatory response to CO2 (4, 5), both of which have been implicated in periodic breathing during sleep.
Some of the results of these studies have been previously reported in the form of an abstract (15).
METHODS
Subjects
We studied 22 patients with stable CHF (no unstable angina or myocardial infarction within 6 months): 11 with CSA (CSA group) and 11 without CSA (control group). The CSA group had an apneaeChypopnea index of 10 or greater per hour of sleep, with at least 75% of the events scored as apneas. Of the apneas, at least 75% were classified as central apnea. The definitions of apneas and hypopneas are provided in the online supplement. The control group had an apneaeChypopnea index of 10 or less per hour, with a central apnea index less than 5 per hour of sleep. The patients in the two groups were matched with respect to age, sex, body mass index, and left ventricular ejection fraction. Patients were excluded for a history of cerebrovascular abnormality, carotid artery occlusive disease, previous cardiopulmonary resuscitation, primary valvular heart disease, or overt neurologic lesions. The study was approved by the University of Wisconsin Health Sciences Human Subjects Committee, and all patients provided informed, written consent before the study.
Measurements
The cerebrovascular regulation was evaluated by transcranial Doppler ultrasonography technique. A 2-MHz pulsed Doppler ultrasound system (Neurovision 500 M; Multigon Industries, Yonkers, NY) was used to continuously measure peak CBF velocity (CBFV) in the proximal segment of the middle cerebral artery (MCA). The MCA was insonated through the right temporal bone window using previously described search techniques (16). After detection and optimization of the Doppler signal, the probe was mechanically secured using a headband device and probe holder to provide a fixed angle of insonation for the duration of the experiment. Subjects were studied in the semirecumbent position and were asked to keep their heads still and eyes open throughout the experiment. If the signal quality deteriorated, the probe was readjusted, and the trial would be restarted. End-tidal CO2 (PETCO2) was sampled from a leak-free mask and measured by a gas analyzer (Model CD-3A; Ametek, Pittsburgh, PA). Heart rate, arterial pressure in arm and finger, VT, breathing frequency, and SaO2 were measured as previously described (17).
All variables were recorded continuously into a computer (sampling rate, 120 Hz) for off-line analysis.
Experimental Protocols
CO2 CHALLENGE.
Subjects were requested not to eat or drink caffeine-containing beverages within 4 hours before reporting to the laboratory. Each trial of hyper- and hypocapnia lasted 5 minutes. During hypercapnia trials, gas mixtures of 3 and 5% CO2 with 21% O2 and a balance of nitrogen (76eC74%) were added to the breathing circuit. At the last minute of each exposure, the individual's VT and frequency were noted as the target for voluntary hyperventilation. During the hypocapnia trials, subjects used auditory (a computer-generated tone for the respiratory duty cycle) and visual (the trace of each individual's VT with a target line to reach with each breath) feedback to mimic breathing patterns they exhibited during inhalation of 3% (mild hypocapnia) and 5% CO2 (moderate hypocapnia). The trials were separated by at least 5 minutes of room air breathing to allow the circulatory and ventilatory variables to return to baseline levels. All experiments were started between 8:00 and 9:00 A.M. to standardize the effect of the diurnal variability of cerebral vasomotor reactivity (18).
Breath-hold.
Subjects performed 20-second breath-hold starting at functional residual capacity. They exhaled at the end of each breath-hold so that an end-tidal sample could be acquired for CO2 analysis. Three to five breath-hold trials were repeated in each individual.
Data Analysis
All data were averaged during the fifth minute of each trial. The slope of CBFV/PETCO2 was determined by the least squares linear regression analysis for each subject as well as for each group (CSA vs. control). The correlation between left ventricular ejection fraction and overall CBFV/PETCO2 slope was examined by performing a logistical regression. The parameters in Table 1 were compared between the two groups by using an unpaired t test. The parameters in Table 2 and the slopes of CBFV/PETCO2 were compared by two-way analysis of variance with repeated measurements (one factor repetition), and interaction effects were resolved using post hoc comparisons (Tukey). The speed of CBFV response to change of CO2 was evaluated during the 5% CO2 inhalation at the transition period of on- and off-CO2 administration as indicated in Figure 1. The peak CBFV after the breath-hold was calculated as a 5-second average, which contained 2.5 seconds before and after the cardiac cycle that was associated with the highest CBFV. The peak CBFV after the breath-hold was compared with the baseline and between groups by a two-way analysis of variance with repeated measurements with post hoc comparisons as necessary. The time from the end of the breath-hold to the peak CBFV (Figure 2) as well as to the subsequent nadir of SaO2 was compared between the control and CSA groups by an unpaired t test. A value of p < 0.05 was considered statistically significant. Data are expressed as means ± SE.
RESULTS
Characteristics of Subjects
Both groups consisted of elderly men with moderately increased body mass index (Table 1). In eight subjects of the control group and 10 subjects of the CSA group, CHF was secondary to ischemic cardiomyopathy, with the remainder of patients having nonischemic dilated cardiomyopathy. Patients in the two groups had a comparable low left ventricular ejection fraction (control vs. CSA, 32 ± 3 vs. 30 ± 3%; p = 0.45). More patients in the CSA group had atrial fibrillation (3 vs. 0), mitral regurgitation (3 vs. 1), and pulmonary hypertension (4 vs. 2). Patients received similar medications, with eight subjects in the control and nine subjects in the CSA group on angiotensin converting enzyme inhibitors and diuretics, five subjects in the control and six in the CSA group on digoxin, and seven subjects in the control and five in the CSA group on -blockers. The apneaeChypopnea index was pathologically higher in the CSA group than in the control group (43 ± 6 vs. 5 ± 2 events/hour, p < 0.01).
Cardiocerebrovascular Response to Hypercapnia and Hypocapnia
Nine patients with CSA and eight without CSA participated in the cerebrovascular response to hypercapnia and hypocapnia project. The two groups had comparable cardiorespiratory values under resting and all other experimental conditions, except that SaO2 tended to be lower in subjects in the CSA group than in control subjects, although only statistically significant during mild hypocapnia (Table 2). Patients with CSA had a lower CBFV response to CO2 over the entire range of CO2 (Figure 3), compared with those without CSA (1.8 ± 0.2 vs. 2.5 ± 0.2%/mm Hg, p < 0.05). However, when we looked at the hypercapnic and hypocapnic responses separately, we found that only the hypocapnic response was significantly lower in the CSA group (1.2 ± 0.3 vs. 2.2 ± 0.1%/mm Hg, p < 0.05), whereas the hypercapnic response showed a directional trend toward reduction that was not statistically different between the two groups (2.4 ± 0.4 vs. 3.1 ± 0.4%/mm Hg, p = 0.19; Figure 4). There was no correlation between the CBFV response to CO2 and left ventricular ejection fraction (slope, 0.02; r2 = 0.1; p = 0.20).
We determined whether the time course of the CBFV response to CO2 was different between patients with and without CSA. Both groups showed similar time delay from the beginning of the CO2 inhalation to the onset of the rise in CBFV (control vs. CSA, 25 ± 6 vs. 30 ± 9 seconds; p = 0.68) and from the termination of CO2 administration to the beginning of CBVF falling (13 ± 1 vs. 16 ± 3 seconds, p = 0.29). The time from the establishment of the high and stable level of PETCO2 to the peak response of CBFV was not different in the two groups (31 ± 4 vs. 21 ± 6 seconds, p = 0.21). Nor were differences noted in the time from termination of CO2 administration and the nadir response of the CBFV (17 ± 2 vs. 20 ± 4 seconds, p = 0.44). The CBFV response to the off-CO2 stimulation was faster than to the on-CO2 stimulation in both groups as shown in Figure 1. After the initial rising/declining, CBFV slightly fluctuated around the new level with no obvious adaptation (Figure 1).
Hemodynamic Effects of Breath-Hold
Nine control subjects and eight patients with CSA (six in each group also participated in the CBF reactivity to CO2 protocol described above) were able to hold their breath for 20 seconds. In these 17 subjects, the breath-hold caused a similar increase in PETCO2 (control, from 35 ± 1 to 44 ± 2 mm Hg; CSA, from 36 ± 2 to 44 ± 2 mm Hg) and decrease in SaO2 (2.1 ± 0.6 vs. 2.2 ± 0.4%, p = 0.85). The time from the end of breath-hold to the subsequent nadir of SaO2, known as lung-to-ear circulatory delay (19), was not different between the two groups (16.2 ± 1.7 vs. 17.2 ± 1.3 seconds, p = 0.65), although obviously longer than the normal value (20). The maximum CBFV response to the breath-hold took place after a similar time delay in both groups (time delay to the peak velocity, 11.7 ± 2.6 seconds in control and 11.3 ± 2.3 seconds in CSA). The peak CBFV was significantly smaller in the CSA group compared with the control group (119 ± 4 vs. 141 ± 8%, p < 0.05). CBFV did not increase during the breath-hold itself (Figures 2 and 5). There was no significant change in blood pressure during the breath-hold. At the time of the peak CBFV in the recovery period after the breath-hold, the mean arterial pressure showed a greater increase in the control group (85 ± 6 to 93 ± 6 mm Hg, p < 0.05) compared with the CSA group (91 ± 10 to 93 ± 11 mm Hg, p = 0.36). The peak CBFV was significantly correlated to the change of peak blood pressure (r2 = 0.60, p = 0.01).
In addition, a few subjects underwent long, cyclic breathing pauses after the hypocapnia trials, which gave us an opportunity to examine the CBFV response to spontaneous apnea. Figure 6 shows examples from two subjects: one with CSA and the other without. Even though the patient with CSA had longer apneas and more severe desaturation, the oscillations in CBF were smaller in this patient compared with the patient without CSA.
DISCUSSION
The major finding of this study is an attenuated cerebrovascular reactivity to the changes in PCO2 in patients with CHF and CSA compared with those with similar cardiac dysfunction but without CSA. The impairment of reactivity is more substantial during hypocapnia. Because CBF response to CO2 is believed to minimize respiratory overshooting and undershooting during periodic breathing, our finding reveals a potential cerebrovascular mechanism contributing to the breathing instability in patients with CHF and CSA.
Methodologic Considerations and Limitations
The transcranial Doppler technique provides a noninvasive and real-time measurement of CBFV rather than CBF. The CBFV in the MCA can be a valid measure of blood flow through the artery if the diameter of the vessel remains constant throughout all experimental conditions. This assumption is likely to be valid during changing levels of CO2, the primary intervention in our study, because CO2 is acting mainly on small arteries and arterioles in humans with minimal changes (< 4%) in the MCA diameter (21eC24). Although the brainstem is perfused by the vertebraleCbasilar system, we chose to measure CBFV in the MCA as a surrogate for changes in mean brain blood flow because the MCA is technically more accessible, and previous studies have shown a good correlation between the MCA blood flow and global and regional changes in cerebral perfusion (25). The MCA carries 80% of the hemispheric blood (26). When comparing the responsiveness to CO2 of CBFV in the MCA measured by the Doppler and of CBF measured by magnetic resonance imaging on four normal subjects in our laboratory, we found virtually identical slopes between the two techniques (magnetic resonance imaging: y = 3.9x eC 58, r2 = 0.93; Doppler: y = 3.9x eC 61, r2 = 0.96). During the MRI study, no change in MCA diameter was noted with hyper- or hypocapnia. The cerebrovascular responses to CO2 are qualitatively similar throughout the brain, but more pronounced in gray versus white matter and in cerebellar and brainstem regions versus cortical areas (27eC29). We therefore believe that our measurement of CBFV reactivity in the MCA should represent the blood flow changes in other regions of the brain, including the ventrolateral part of the brainstem, where the change of CO2 and pH is directly sensed by the central chemoreceptors and thereby affects the respiratory neurons (30, 31).
We used the linear regression analysis to describe each subject's CBFV/PETCO2 relationship. Although the published data showed uncertainty as to whether the relationship between CBFV and change of CO2 is linear across the hypo- and hypercapnic range of PETCO2 (32), we found that linear regression provided the best fit with r2 0.9 in most, though not all, of our subjects' data. We also analyzed the CBFV/PETCO2 above and below the eupneic PCO2 separately, which showed the same linear tendency as did the overall regression.
Mechanism of the Low CBFV/PETCO2 in Patients with CHF and CSA
Both the control and CSA groups showed the expected CBFV response to CO2 in the hyper- and hypocapnic ranges and demonstrated a similar differential responsiveness to hyper- and hypocapnia that has been observed in healthy subjects (33). The greater responsiveness to hypercapnia compared with hypocapnia may be related the observation that intracranial vascular tone is more influenced by vasodilator mediators compared with vasoconstrictive mediators (34). However, in the patients with CHF, the gain of the response is relatively small in both groups (2.4eC3.1% increase/mm Hg) compared with young, healthy subjects (4.7 ± 0.3% mm HgeC1) (17). This difference is likely to be related to the poor heart function and/or endothelial dysfunction. CHF is associated with impaired endothelium-dependent vasodilation (35). A low cerebrovascular CO2 reactivity with the heart dysfunction has been reported by Georgiadis and coworkers (8). Moreover, most of our subjects had an underlying ischemic cardiomyopathy, which increases the likelihood of endothelial dysfunction (36) and coexistent cerebrovascular disease. Thus, all patients with CHF are likely to have an abnormal responsiveness to CO2, but this does not explain the greater impairment in reactivity in patients with CSA.
The mechanism underlying the reduction in cerebrovascular reactivity in patients with CSA is not clear. CO2 modulates CBF via direct effect of extracellular pH on the smooth muscle of cerebral pial vessels (1). The vascular smooth muscle tone in the brain is also regulated by other neural, humoral, and physical factors. Our experiment does not allow us to determine the pathologic linkage between the impaired cerebrovascular response to CO2 and periodic breathing. The impaired cerebrovascular reactivity may also be a direct consequence of the periodic breathing. The supportive evidence is that a reduced CBF response to CO2 has also been observed in patients with obstructive sleep apnea, and this deficiency reverses after treatment with nasal continuous positive airway pressure (37). Both recurrent central and obstructive apneas are associated with asphyxia and surges in arterial pressure that could cause endothelial dysfunction and impairment of cerebrovascular reactivity to CO2. Sleep fragmentation as a result of overnight periodic breathing could also diminish hypercapnic vasomotor response (38). However, conflicting evidence has also been reported in the literature, with a normal or even higher CBF reactivity to PCO2 in patients with obstructive sleep apnea (39), and whether the endothelium is involved in the CBF response to CO2 is still controversial (40).
Patients with CSA were usually associated with excessive sympathetic activation (41), but sympathetic vasoconstrictor outflow does not appear to influence the cerebrovascular response to changes in PETCO2 (17, 42). Even if sympathetic vasoconstrictor outflow could blunt the hypercapnia-induced CBF increase as reported by Jordan and colleagues (43), it still cannot explain why patients with CSA also had a less vasoconstriction in response to hypocapnia. Although patients with CSA have a greater peripheral chemosensitivity to CO2 (5, 44), CBF responses to CO2 were not influenced by cutting peripheral chemoreceptor afferent nerves (29). Drug treatment received in the two groups was basically similar. All of our subjects were on medications for CHF that have vascular effects to improve cardiac output. However, literature shows that these medications have no overt effect on baseline CBF (45) or its dynamic changes (46).
Finally, published evidence suggests that brain hypoperfusion could attenuate cerebrovascular reactivity to CO2 (13, 47), and conversely, brain hyperperfusion could enhance the reactivity (48). A low resting CBF value has been observed in patients with CHF (6, 7), although the investigators did not clarify whether or not the patients had apneas. From the point of teleology, the low CBF response to hypocapnia could protect patients from severe ischemia in the presence of an already low cerebral perfusion. Transcranial Doppler technique does not allow us to compare the absolute CBF between individuals. Accordingly, we cannot rule out the possibility that patients with CSA may have a lower resting CBF, which results in a lower CBF response to CO2.
We investigated the possibility that a temporal abnormality in the CO2 stimulus and the vascular response could account for some of the difference between the patients with and without CSA. However, the same time lag between the two groups with and without CSA in our study suggests a similar circulatory transit delay from lung to brain and a similar reflection speed of cerebrovascular tone to increase of PaCO2.
We used breath-holds to mimic sleep apnea and to determine if patients with CSA would respond differently than patients without CSA. The CBFV response to breath-holds yielded a similar result observed during CO2 breathing (i.e., less increase in the CSA group than in the control subjects). The prolonged time delay between the onset of breath-hold and the peak in CBFV in both groups is due to the long circulation time, which was 16.2 to 17.2 seconds in our patients, compared with 6 to 10 seconds in subjects with a normal heart (20, 49). Thus, unlike normal subjects where two-thirds of the total increase in CBFV occurred before the end of the breath-hold (17), neither of our patient groups demonstrated an increase of CBFV during the breath-hold. Likewise, no increase in CBFV was seen during the spontaneously occurring apneas (Figure 6) or with previously reported examples of Cheyne-Stokes' respiration (46, 50). Thus, the absence of rise in CBFV during the apnea in both groups of our patients was likely due to the delay in arrival of the increasing alveolar PCO2 during the apnea to the cerebral vessels due to a prolonged circulation time. However, when the rising PCO2 secondary to the absence of ventilation eventually did reach the cerebral circulation after the breath-hold, the patients with CSA showed a smaller increase in CBFV because they have less of a cerebrovascular responsiveness to CO2.
We noticed that patients with CSA had a smaller increase in blood pressure in response to both exogenous hypercapnia and hypercapnia during the breath-hold. However, because of the marked autoregulation activity of the cerebral circulation, the ability of the perfusion pressure to influence CBF is minimal. Our previous study in normal subjects showed that approximately 16% of the increase in CBFV after the breath-hold was due to the associated increase in arterial pressure during and after the apnea (17). Therefore, part of the reason for the smaller rise in CBFV after the breath-hold in the patients with CSA may have been related to their absence of a blood pressure response to CO2 compared with the control subjects. However, the small rise in blood pressure noted in the control patients could not account for all of the difference in rise in CBFV after the breath-hold between the two groups.
Influence of Attenuated Cerebrovascular Reactivity to CO2 on Breathing Stability
A close relationship between changes in CBF and ventilatory output has been shown in animal studies. Simultaneous measurements of the peak diaphragmatic electromyogram and brain blood flow in anesthetized goats yielded a reciprocal relationship between changes in respiratory drive and CBF (51). Evidence for the modulation of increased ventilatory chemosensitivity by reduced CBF was provided by Chapman and his colleagues (13), who showed that when they reduced CBF by 30%, ventilatory responsiveness to CO2 was increased if expressed as E against PETCO2 or PaCO2, but was unchanged from baseline if plotted as E against cerebral venous PCO2. This observation suggests that CBF affects ventilation through modification of tissue PCO2 in the region of the medullary chemoreceptor area. With a poor brain perfusion, the central chemoreceptor will sense a relatively higher H+ for a given change of PaCO2 and will correspondingly trigger a more vigorous ventilatory response.
Cerebrovascular dysfunction may play a role in the etiology of CSA in patients with CHF. PaCO2 is a powerful regulator of the respiration as well as the cerebral circulation, and its ventilatory excitatory effect at the central chemoreceptor may be modulated by its influence on the cerebral circulation. In all mammals studied, hypercapnia causes cerebral vasodilation, whereas hypocapnia causes vasoconstriction (1). The increased CBF during hypercapnia leads to less increase in H+ at the central chemoreceptor, whereas decreased CBF during hypocapnia leads to less decrease in H+ at the central chemoreceptor. By buffering the change in H+ at the central chemoreceptor, an adequate change in CBF in response to CO2 will stabilize the breathing pattern in the face of transient perturbations in arterial CO2. Because patients with CHF and CSA have a reduced CBF reactivity to CO2, we expect them to have a greater increase in PCO2 and H+ at the central chemoreceptor during hypercapnia and a greater reduction in PCO2 and H+ during hypocapnia for a given change in PaCO2 compared with patients without CSA. Therefore, reduced CBF reactivity to CO2 should increase the ventilatory response to CO2, a major feature of CSA (4, 5, 44). When PaCO2 is reduced below eupneic levels by a ventilatory overshoot, brain PCO2 will fall more than normal because of a reduced cerebral vasoconstrictive response and thereby elevate the PaCO2 at which apnea occurs. The increased apneic threshold for PaCO2 will bring it closer to the eupneic PCO2, another critical feature of CSA (14).
Although posthyperventilation apnea is usually initiated by peripheral chemoreceptors (52), a large body of evidence supports the involvement of central mechanisms in the development of posthyperventilation apnea. First, carotid body hypocapnia by itself is not sufficient to produce apnea as shown by the presence of only reduced VT but not apnea during hypocapnic perfusion of an isolated carotid body preparation in the unanesthetized sleeping dog (53). Second, the delayed posthyperventilation apnea in the carotid body denervated dog unmasked the ability of the central chemoreceptors to cause apnea in response to hypocapnia (52). Finally, we have observed contradictory effects on breathing stability of two ventilatory stimulants whose effect is predominantly mediated by the peripheral chemoreceptors, hypoxia and almitrine, whereby hypoxia was destabilizing and almitrine was stabilizing (54). Likewise, two inhibitory agents mediated, in part, by the peripheral chemoreceptors, hyperoxia and dopamine, have different effects on ventilatory stability, with hyperoxia stabilizing and dopamine destabilizing (55). We suspect that hypoxia and hyperoxia may affect posthyperventilation breathing pattern through a combination of their peripheral and central effects (56, 57) as well as their influence on CBF (58). In other words, even though posthyperventilation apnea is mainly mediated by peripheral chemoreceptors, additional central influences are required to produce periodic breathing. The impaired CBF reactivity to CO2 might predispose patients to periodic breathing through a central mechanism.
In summary, we observed an attenuated cerebrovascular response to CO2 in patients with CHF and CSA. The compromise of the ability of CO2 to regulate CBF might disturb the stability of the breathing pattern by causing overshooting during hypercapnia and undershooting during hypocapnia. Thus, altered CBF may be an important contributor to breathing instabilities during sleep, which are associated with large fluctuations in CO2 and ventilation.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
REFERENCES
Heistad DD, Kontos HA. Cerebral circulation. In: Shepherd JT, Abboud FM, editors. Handbook of physiology section 2: the cardiovascular system. Bethesda, MD: American Physiological Society; 1983. pp. 137eC182.
Hanly P, Zuberi N, Gray R. Pathogenesis of Cheyne-Stokes respiration in patients with congestive heart failure: relationship to arterial PCO2. Chest 1993;104:1079eC1084.
Tkacova R, Hall ML, Liu PP, Fitzgerald FS, Bradley TD. Left ventricular volume in patients with heart failure and Cheyne-Stokes respiration during sleep. Am J Respir Crit Care Med 1997;156:1549eC1555.
Javaheri S. A mechanism of central sleep apnea in patients with heart failure. N Engl J Med 1999;341:949eC954.
Solin P, Roebuck T, Johns DP, Haydn Walters E, Naughton MT. Peripheral and central ventilatory responses in central sleep apnea with and without congestive heart failure. Am J Respir Crit Care Med 2000;162:2194eC2200.
Lee CW, Lee JH, Lim TH, Yang HS, Hong MK, Song JK, Park SM, Park SJ, Kim JJ. Prognostic significance of cerebral metabolic abnormalities in patients with congestive heart failure. Circulation 2001;103:2784eC2787.
Rajagopalan B, Raine AE, Cooper R, Ledingham JG. Changes in cerebral blood flow in patients with severe congestive cardiac failure before and after captopril treatment. Am J Med 1984;76:86eC90.
Georgiadis D, Sievert M, Cencetti S, Uhlmann F, Krivokuca M, Zierz S, Werdan K. Cerebrovascular reactivity is impaired in patients with cardiac failure. Eur Heart J 2000;21:407eC413.
Caparas SN, Clair MJ, Krombach RS, Hendrick JW, Houck WV, Kribbs SB, Mukherjee R, Tempel GE, Spinale FG. Brain blood flow patterns after the development of congestive heart failure: effects of treadmill exercise. Crit Care Med 2000;28:209eC214.
Chapman RW, Santiago TV, Edelman NH. Effects of graded reduction of brain blood flow on ventilation in unanesthetized goats. J Appl Physiol 1979;47:104eC111.
Lane DJ, Rout MW, Williamson DH. Mechanism of hyperventilation in acute cerebrovascular accidents. BMJ 1971;3:9eC12.
Schmidt CF. The influence of cerebral blood-flow on respiration. I. The respiratory responses to changes in cerebral blood flow. Am J Physiol 1928;84:202eC222.
Chapman RW, Santiago TV, Edelman NH. Effects of graded reduction of brain blood flow on chemical control of breathing. J Appl Physiol 1979;47:1289eC1294.
Xie AL, Skatrud JB, Puleo DS, Rahko PS, Dempsey JA. Apnea-hypopnea threshold for CO2 in patients with congestive heart failure. Am J Respir Crit Care Med 2002;165:1245eC1250.
Xie A, Khayat RN, Puleo DS, Morgan BJ, Skatrud JB. Cerebrovascular reponse to CO2 in patients with congestive heart failure . Am J Respir Crit Care Med 2003;167:A173.
Otis SM, Ringelstein EB. The transcranial Doppler examination: principles and applications of transcranial Doppler sonography. In: Tegeler CH, Babikian VL, Gomez CR, editors. Neurosonology. St. Louis, MO: Mosby; 1996. pp. 113eC128.
Przybylowski T, Bangash MF, Reichmuth K, Morgan BJ, Skatrud JB, Dempsey JA. Mechanisms of the cerebrovascular response to apnoea in humans. J Physiol 2003;548:323eC332.
Ameriso SF, Mohler JG, Suarez M, Fisher M. Morning reduction of cerebral vasomotor reactivity. Neurology 1994;44:1907eC1909.
Lorenzi-Filho G, Rankin F, Bies I, Douglas Bradley T. Effects of inhaled carbon dioxide and oxygen on Cheyne-Stokes respiration in patients with heart failure. Am J Respir Crit Care Med 1999;159:1490eC1498.
Dahan A, DeGoede J, Berkenbosch A, Olievier IC. The influence of oxygen on the ventilatory response to carbon dioxide in man. J Physiol 1990;428:485eC499.
Czosnyka M, Richards HK, Whitehouse HE, Pickard JD. Relationship between transcranial Doppler-determined pulsatility index and cerebrovascular resistance: an experimental study. Neurosurgery 1996;84:79eC84.
Giller CA, Bowman G, Dyer H, Mootz L, Krippnoer W. Cerebral arterial diameters during changes in blood pressure and carbon dioxide curing craniotomy. Neurosurgery 1993;32:737eC742.
Serrador JM, Picot PA, Rutt BK, Shoemaker JK, Bondar RL. MRI measures of middle cerebral artery diameter in conscious humans during simulated orthostasis. Stroke 2000;31:1672eC1678.
Valdueza JM, Balzer JO, Villringer A, Vogl TJ, Kutter R, Einhaupl KM. Changes in blood flow velocity and diameter of the middle cerebral artery during hyperventilation: assessment with MR and transcranial Doppler sonography. AJNR Am J Neuroradiol 1997;18:1929eC1934.
Pandit JJ, Mohan RM, Paterson ND, Poulin MJ. Cerebral blood flow sensitivity to CO2 measured with steady-state and Read's rebreathing methods. Respir Physiol Neurobiol 2003;137:1eC10.
Balfors EM, Franklin KA. Impairment of cerebral perfusion during obstructive sleep apneas. Am J Respir Crit Care Med 1994;150:1587eC1591.
Heistad DD, Marcus ML, Ehrhardt JC, Abboud FM. Effect of stimulation of carotid chemoreceptors on total and regional cerebral blood flow. Circ Res 1976;38:20eC25.
Ramsay SC, Murphy K, Shea SA, Friston KJ, Lammertsma AA, Clark JC, Adams L, Guz A, Frackowiak RSJ. Changes in global cerebral blood flow in humans: effect on regional cerebral blood flow during a neural activation task. J Physiol 1993;471:521eC534.
Sato A, Trzebski A, Zhou W. Local cerebral blood flow responses in rats to hypercapnia and hypoxia in the rostral ventrolateral medulla and in the cortex. J Auton Nerv Syst 1992;41:79eC86.
Kawai A, Ballantyne D, Muckenhoff K, Scheid P. Chemosensitive medullary neurones in the brainstemeCspinal cord preparation of the neonatal rat. J Physiol 1996;492:277eC292.
Thomas T, Spyer KM. ATP as a mediator of mammalian central CO2 chemoreception. J Physiol 2000;523:441eC447.
Poulin MJ, Liang PJ, Robbins PA. Fast and slow components of cerebral blood flow response to step decreases in end-tidal PCO2 in humans. J Appl Physiol 1998;85:388eC397.
Ide K, Eliasziw M, Poulin MJ. The relationship between middle cerebral artery blood velocity and end-tidal PCO2 in the hypocapnic-hypercapnic range in humans. J Appl Physiol 2003;95:129eC137.
Toda N, Okamura T. Cerebral vasodilators. Jpn J Pharmacol 1998;76:349eC367.
Nakamura M, Sugawara S, Arakawa N, Nagano M, Shizuka T, Shimoda Y, Sakai T, Hiramori K. Reduced vascular compliance is associated with impaired endothelium-dependent dilatation in the brachial artery of patients with congestive heart failure. J Card Fail 2004;10:36eC42.
Tentolouris C, Tousoulis D, Antoniades C, Bosinakou E, Kotsopoulou M, Trikas A, Toutouzas P, Stefanadis C. Endothelial function and proinflammatory cytokines in patients with ischemic heart disease and dilated cardiomyopathy. Int J Cardiol 2004;94:301eC305.
Diomedi M, Placidi F, Cupini LM, Bernardi G, Silvestrini M. Cerebral hemodynamic changes in sleep apnea syndrome and effect of continuous positive airway pressure treatment. Neurology 1998;51:1051eC1056.
Qureshi AI, Christopher Winter W, Bliwise DL. Sleep fragmentation and morning cerebrovasomotor reactivity to hypercapnia. Am J Respir Crit Care Med 1999;160:1244eC1247.
Droste DW, Ludemann P, Anders F, Kemeny V, Thomas M, Krauss JK, Ringelstein EB. Middle cerebral artery blood flow velocity, end-tidal pCO2 and blood pressure in patients with obstructive sleep apnea and in healthy subjects during continuous positive airway pressure breathing. Neurol Res 1999;21:737eC741.
Toda N, Ayajiki K, Enokibori M, Okamura T. Monkey cerebral arterial relaxation caused by hypercapnic acidosis and hypertonic bicarbonate. Am J Physiol 1993;265:H929eCH933.
Van de Borne P, Oren R, Abouassaly C, Anderson E, Somers VK. Effect of Cheyne-Stokes respiration on muscle sympathetic nerve activity in severe congestive heart failure secondary to ischemic or idiopathic dilated cardiomyopathy. Am J Cardiol 1998;81:432eC436.
LeMarbre G, Stauber S, Khayat RN, Puleo DS, Skatrud JB, Morgan BJ. Baroreflex-induced sympathetic activation does not alter cerebrovascular CO2 responsiveness in humans. J Physiol 2003;551:609eC616.
Jordan J, Shannon JR, Diedrich A, Black B, Costa F, Robertson D, Biaggioni I. Interaction of carbon dioxide and sympathetic nervous system activity in the regulation of cerebral perfusion in humans. Hypertension 2000;36:383eC388.
Xie A, Rutherford R, Rankin F, Wong B, Bradley TD. Hypocapnia and increased ventilatory responsiveness in patients with idiopathic central sleep apnea. Am J Respir Crit Care Med 1995;152:1950eC1955.
Waldemar G, Paulson OB. Angiotensin converting enzyme inhibition and cerebral circulation: a review. Br J Clin Pharmacol 1989;28:177SeC182S.
Franklin KA, Sandstrom E, Johansson G, Balfors EM. Hemodynamics, cerebral circulation, and oxygen saturation in Cheyne-Stokes respiration. J Appl Physiol 1997;83:1184eC1191.
Kagstrom E, Smith ML, Siesjo BK. Cerebral circulatory responses to hypercapnia and hypoxia in the recovery period following complete and incomplete cerebral ischemia in the rat. Acta Physiol Scand 1983;118:281eC291.
Hauge A, Nicolaysen G, Thoresen M. Acute effects of acetazolamide on cerebral blood flow in man. Acta Physiol Scand 1983;117:233eC239.
Khoo MC, Kronauer RE, Strohl KP, Slutsky AS. Factors inducing periodic breathing in humans: a general model. J Appl Physiol 1982;53:644eC659.
Karp HR, Sieker HO, Heyman A. Cerebral circulation and function in Cheyne-Stokes respiration. Am J Med 1961;30:861eC870.
Parisi RA, Neubauer JA, Frank MM, Santiago TV, Edelman NH. Linkage between brain blood flow and respiratory drive during rapid-eye-movement sleep. J Appl Physiol 1988;64:1457eC1465.
Nakayama H, Smith CA, Rodman JR, Skatrud JB, Dempsey JA. Carotid body denervation eliminates apnea in response to transient hypocapnia. J Appl Physiol 2003;94:155eC164.
Smith CA, Saupe KW, Henderson KS, Dempsey JA. Ventilatory effects of specific carotid body hypocapnia in dogs during wakefulness and sleep. J Appl Physiol 1995;79:689eC699.
Nakayama H, Smith CA, Rodman JR, Skatrud JB, Dempsey JA. Effect of ventilatory drive on carbon dioxide sensitivity below eupnea during sleep. Am J Respir Crit Care Med 2002;165:1251eC1260.
Chenuel BJ, Smith CA, Henderson KS, Dempsey JA. Ventilatory instability induced by selective carotid body inhibition in the sleeping dog. Adv Exp Med Biol 2004;551:197eC201.
Horn EM, Waldrop TG. Oxygen-sensing neurons in the caudal hypothalamus and their role in cardiorespiratory control. Respir Physiol 1997;110:219eC228.
Neubauer JA, Sunderram J. Oxygen-sensing neurons in the central nervous system. J Appl Physiol 2004;96:367eC374.
Cohen PJ, Alexander SC, Smith TC, Reivich M, Wollman H. Effects of hypoxia and normocarbia on cerebral blood flow and metabolism in conscious man. J Appl Physiol 1967;23:183eC189.(Ailiang Xie, James B. Ska)
Population Health Sciences, University of Wisconsin
the William S. Middleton Memorial Veterans Hospital, Madison, Wisconsin
ABSTRACT
Rationale: Cerebrovascular reactivity to CO2 provides an important counterregulatory mechanism that serves to minimize the change in H+ at the central chemoreceptor, thereby stabilizing the breathing pattern in the face of perturbations in PaCO2. However, there are no studies relating cerebral circulation abnormality to the presence or absence of central sleep apnea in patients with heart failure. Objectives: To determine whether patients with congestive heart failure and central sleep apnea have an attenuated cerebrovascular responsibility to CO2. Methods: Cerebral blood flow velocity in the middle cerebral artery was measured in patients with stable congestive heart failure with (n = 9) and without (n = 8) central sleep apnea using transcranial ultrasound during eucapnia (room air), hypercapnia (inspired CO2, 3 and 5%), and hypocapnia (voluntary hyperventilation). In addition, eight subjects with apnea and nine without apnea performed a 20-second breath-hold to investigate the dynamic cerebrovascular response to apnea. Measurements and Main Results: The overall cerebrovascular reactivity to CO2 (hyper- and hypocapnia) was lower in patients with apnea than in the control group (1.8 ± 0.2 vs. 2.5 ± 0.2%/mm Hg, p < 0.05), mainly due to the prominent reduction of cerebrovascular reactivity to hypocapnia (1.2 ± 0.3 vs. 2.2 ± 0.1%/mm Hg, p < 0.05). Similarly, brain blood flow demonstrated a smaller surge after a 20-second breath-hold (peak velocity, 119 ± 4 vs. 141 ± 8% of baseline, p < 0.05). Conclusion: Patients with central sleep apnea have a diminished cerebrovascular response to PETCO2, especially to hypocapnia. The compromised cerebrovascular reacticity to CO2 might affect stability of the breathing pattern by causing ventilatory overshooting during hypercapnia and undershooting during hypocapnia.
Key Words: central sleep apnea; cerebral blood flow; hypercapnia; hypocapnia
Breathing pattern and brain blood flow are closely linked by PaCO2. Normally, elevated PaCO2 causes vasodilation and increased cerebral blood flow (CBF) (1), which will wash out brain CO2, and consequently reduce H+ at the central chemoreceptor. As a result of the lower H+ at the central chemoreceptor, the ventilatory stimulation to the elevated PaCO2 will be blunted. Conversely, reduced PaCO2 causes vasoconstriction and decreased CBF, which will increase H+ at the central chemoreceptor. The ventilatory inhibition in response to the reduced CO2 will be lessened as a result of the higher H+ at the central chemoreceptors. Thus, the effect of CO2 on CBF provides an important counterregulatory mechanism that serves to minimize the change in H+ at the central chemoreceptor, thereby stabilizing the breathing pattern in the face of perturbations in PaCO2.
In patients with congestive heart failure (CHF), we suspected that reduced baseline CBF and reduced cerebrovascular responsiveness to CO2 could, at least partly, account for the high baseline level of ventilation (2, 3), the high ventilatory response to CO2 (4, 5), and the high prevalence of breathing instability and apnea during sleep. Such a reduction in baseline CBF (6, 7) and the CBF response to CO2 (8) has been demonstrated in patients with CHF as well as in a porcine model of CHF (9). Previous studies have demonstrated an increased ventilation (10eC12) and respiratory sensitivity (13) in response to brain ischemia, indicating the importance of brain blood flow on breathing stability. However, there are no studies relating cerebral circulation abnormality to the presence or absence of central sleep apnea (CSA) in patients with CHF. The purpose of this study was to determine whether patients with CHF and CSA have an attenuated cerebrovascular responsibility to CO2 compared with patients with CHF but without CSA. Such an abnormality may predispose to periodic breathing during sleep by both narrowing the difference between the eupneic PCO2 and apneic threshold (14) and by increasing the ventilatory response to CO2 (4, 5), both of which have been implicated in periodic breathing during sleep.
Some of the results of these studies have been previously reported in the form of an abstract (15).
METHODS
Subjects
We studied 22 patients with stable CHF (no unstable angina or myocardial infarction within 6 months): 11 with CSA (CSA group) and 11 without CSA (control group). The CSA group had an apneaeChypopnea index of 10 or greater per hour of sleep, with at least 75% of the events scored as apneas. Of the apneas, at least 75% were classified as central apnea. The definitions of apneas and hypopneas are provided in the online supplement. The control group had an apneaeChypopnea index of 10 or less per hour, with a central apnea index less than 5 per hour of sleep. The patients in the two groups were matched with respect to age, sex, body mass index, and left ventricular ejection fraction. Patients were excluded for a history of cerebrovascular abnormality, carotid artery occlusive disease, previous cardiopulmonary resuscitation, primary valvular heart disease, or overt neurologic lesions. The study was approved by the University of Wisconsin Health Sciences Human Subjects Committee, and all patients provided informed, written consent before the study.
Measurements
The cerebrovascular regulation was evaluated by transcranial Doppler ultrasonography technique. A 2-MHz pulsed Doppler ultrasound system (Neurovision 500 M; Multigon Industries, Yonkers, NY) was used to continuously measure peak CBF velocity (CBFV) in the proximal segment of the middle cerebral artery (MCA). The MCA was insonated through the right temporal bone window using previously described search techniques (16). After detection and optimization of the Doppler signal, the probe was mechanically secured using a headband device and probe holder to provide a fixed angle of insonation for the duration of the experiment. Subjects were studied in the semirecumbent position and were asked to keep their heads still and eyes open throughout the experiment. If the signal quality deteriorated, the probe was readjusted, and the trial would be restarted. End-tidal CO2 (PETCO2) was sampled from a leak-free mask and measured by a gas analyzer (Model CD-3A; Ametek, Pittsburgh, PA). Heart rate, arterial pressure in arm and finger, VT, breathing frequency, and SaO2 were measured as previously described (17).
All variables were recorded continuously into a computer (sampling rate, 120 Hz) for off-line analysis.
Experimental Protocols
CO2 CHALLENGE.
Subjects were requested not to eat or drink caffeine-containing beverages within 4 hours before reporting to the laboratory. Each trial of hyper- and hypocapnia lasted 5 minutes. During hypercapnia trials, gas mixtures of 3 and 5% CO2 with 21% O2 and a balance of nitrogen (76eC74%) were added to the breathing circuit. At the last minute of each exposure, the individual's VT and frequency were noted as the target for voluntary hyperventilation. During the hypocapnia trials, subjects used auditory (a computer-generated tone for the respiratory duty cycle) and visual (the trace of each individual's VT with a target line to reach with each breath) feedback to mimic breathing patterns they exhibited during inhalation of 3% (mild hypocapnia) and 5% CO2 (moderate hypocapnia). The trials were separated by at least 5 minutes of room air breathing to allow the circulatory and ventilatory variables to return to baseline levels. All experiments were started between 8:00 and 9:00 A.M. to standardize the effect of the diurnal variability of cerebral vasomotor reactivity (18).
Breath-hold.
Subjects performed 20-second breath-hold starting at functional residual capacity. They exhaled at the end of each breath-hold so that an end-tidal sample could be acquired for CO2 analysis. Three to five breath-hold trials were repeated in each individual.
Data Analysis
All data were averaged during the fifth minute of each trial. The slope of CBFV/PETCO2 was determined by the least squares linear regression analysis for each subject as well as for each group (CSA vs. control). The correlation between left ventricular ejection fraction and overall CBFV/PETCO2 slope was examined by performing a logistical regression. The parameters in Table 1 were compared between the two groups by using an unpaired t test. The parameters in Table 2 and the slopes of CBFV/PETCO2 were compared by two-way analysis of variance with repeated measurements (one factor repetition), and interaction effects were resolved using post hoc comparisons (Tukey). The speed of CBFV response to change of CO2 was evaluated during the 5% CO2 inhalation at the transition period of on- and off-CO2 administration as indicated in Figure 1. The peak CBFV after the breath-hold was calculated as a 5-second average, which contained 2.5 seconds before and after the cardiac cycle that was associated with the highest CBFV. The peak CBFV after the breath-hold was compared with the baseline and between groups by a two-way analysis of variance with repeated measurements with post hoc comparisons as necessary. The time from the end of the breath-hold to the peak CBFV (Figure 2) as well as to the subsequent nadir of SaO2 was compared between the control and CSA groups by an unpaired t test. A value of p < 0.05 was considered statistically significant. Data are expressed as means ± SE.
RESULTS
Characteristics of Subjects
Both groups consisted of elderly men with moderately increased body mass index (Table 1). In eight subjects of the control group and 10 subjects of the CSA group, CHF was secondary to ischemic cardiomyopathy, with the remainder of patients having nonischemic dilated cardiomyopathy. Patients in the two groups had a comparable low left ventricular ejection fraction (control vs. CSA, 32 ± 3 vs. 30 ± 3%; p = 0.45). More patients in the CSA group had atrial fibrillation (3 vs. 0), mitral regurgitation (3 vs. 1), and pulmonary hypertension (4 vs. 2). Patients received similar medications, with eight subjects in the control and nine subjects in the CSA group on angiotensin converting enzyme inhibitors and diuretics, five subjects in the control and six in the CSA group on digoxin, and seven subjects in the control and five in the CSA group on -blockers. The apneaeChypopnea index was pathologically higher in the CSA group than in the control group (43 ± 6 vs. 5 ± 2 events/hour, p < 0.01).
Cardiocerebrovascular Response to Hypercapnia and Hypocapnia
Nine patients with CSA and eight without CSA participated in the cerebrovascular response to hypercapnia and hypocapnia project. The two groups had comparable cardiorespiratory values under resting and all other experimental conditions, except that SaO2 tended to be lower in subjects in the CSA group than in control subjects, although only statistically significant during mild hypocapnia (Table 2). Patients with CSA had a lower CBFV response to CO2 over the entire range of CO2 (Figure 3), compared with those without CSA (1.8 ± 0.2 vs. 2.5 ± 0.2%/mm Hg, p < 0.05). However, when we looked at the hypercapnic and hypocapnic responses separately, we found that only the hypocapnic response was significantly lower in the CSA group (1.2 ± 0.3 vs. 2.2 ± 0.1%/mm Hg, p < 0.05), whereas the hypercapnic response showed a directional trend toward reduction that was not statistically different between the two groups (2.4 ± 0.4 vs. 3.1 ± 0.4%/mm Hg, p = 0.19; Figure 4). There was no correlation between the CBFV response to CO2 and left ventricular ejection fraction (slope, 0.02; r2 = 0.1; p = 0.20).
We determined whether the time course of the CBFV response to CO2 was different between patients with and without CSA. Both groups showed similar time delay from the beginning of the CO2 inhalation to the onset of the rise in CBFV (control vs. CSA, 25 ± 6 vs. 30 ± 9 seconds; p = 0.68) and from the termination of CO2 administration to the beginning of CBVF falling (13 ± 1 vs. 16 ± 3 seconds, p = 0.29). The time from the establishment of the high and stable level of PETCO2 to the peak response of CBFV was not different in the two groups (31 ± 4 vs. 21 ± 6 seconds, p = 0.21). Nor were differences noted in the time from termination of CO2 administration and the nadir response of the CBFV (17 ± 2 vs. 20 ± 4 seconds, p = 0.44). The CBFV response to the off-CO2 stimulation was faster than to the on-CO2 stimulation in both groups as shown in Figure 1. After the initial rising/declining, CBFV slightly fluctuated around the new level with no obvious adaptation (Figure 1).
Hemodynamic Effects of Breath-Hold
Nine control subjects and eight patients with CSA (six in each group also participated in the CBF reactivity to CO2 protocol described above) were able to hold their breath for 20 seconds. In these 17 subjects, the breath-hold caused a similar increase in PETCO2 (control, from 35 ± 1 to 44 ± 2 mm Hg; CSA, from 36 ± 2 to 44 ± 2 mm Hg) and decrease in SaO2 (2.1 ± 0.6 vs. 2.2 ± 0.4%, p = 0.85). The time from the end of breath-hold to the subsequent nadir of SaO2, known as lung-to-ear circulatory delay (19), was not different between the two groups (16.2 ± 1.7 vs. 17.2 ± 1.3 seconds, p = 0.65), although obviously longer than the normal value (20). The maximum CBFV response to the breath-hold took place after a similar time delay in both groups (time delay to the peak velocity, 11.7 ± 2.6 seconds in control and 11.3 ± 2.3 seconds in CSA). The peak CBFV was significantly smaller in the CSA group compared with the control group (119 ± 4 vs. 141 ± 8%, p < 0.05). CBFV did not increase during the breath-hold itself (Figures 2 and 5). There was no significant change in blood pressure during the breath-hold. At the time of the peak CBFV in the recovery period after the breath-hold, the mean arterial pressure showed a greater increase in the control group (85 ± 6 to 93 ± 6 mm Hg, p < 0.05) compared with the CSA group (91 ± 10 to 93 ± 11 mm Hg, p = 0.36). The peak CBFV was significantly correlated to the change of peak blood pressure (r2 = 0.60, p = 0.01).
In addition, a few subjects underwent long, cyclic breathing pauses after the hypocapnia trials, which gave us an opportunity to examine the CBFV response to spontaneous apnea. Figure 6 shows examples from two subjects: one with CSA and the other without. Even though the patient with CSA had longer apneas and more severe desaturation, the oscillations in CBF were smaller in this patient compared with the patient without CSA.
DISCUSSION
The major finding of this study is an attenuated cerebrovascular reactivity to the changes in PCO2 in patients with CHF and CSA compared with those with similar cardiac dysfunction but without CSA. The impairment of reactivity is more substantial during hypocapnia. Because CBF response to CO2 is believed to minimize respiratory overshooting and undershooting during periodic breathing, our finding reveals a potential cerebrovascular mechanism contributing to the breathing instability in patients with CHF and CSA.
Methodologic Considerations and Limitations
The transcranial Doppler technique provides a noninvasive and real-time measurement of CBFV rather than CBF. The CBFV in the MCA can be a valid measure of blood flow through the artery if the diameter of the vessel remains constant throughout all experimental conditions. This assumption is likely to be valid during changing levels of CO2, the primary intervention in our study, because CO2 is acting mainly on small arteries and arterioles in humans with minimal changes (< 4%) in the MCA diameter (21eC24). Although the brainstem is perfused by the vertebraleCbasilar system, we chose to measure CBFV in the MCA as a surrogate for changes in mean brain blood flow because the MCA is technically more accessible, and previous studies have shown a good correlation between the MCA blood flow and global and regional changes in cerebral perfusion (25). The MCA carries 80% of the hemispheric blood (26). When comparing the responsiveness to CO2 of CBFV in the MCA measured by the Doppler and of CBF measured by magnetic resonance imaging on four normal subjects in our laboratory, we found virtually identical slopes between the two techniques (magnetic resonance imaging: y = 3.9x eC 58, r2 = 0.93; Doppler: y = 3.9x eC 61, r2 = 0.96). During the MRI study, no change in MCA diameter was noted with hyper- or hypocapnia. The cerebrovascular responses to CO2 are qualitatively similar throughout the brain, but more pronounced in gray versus white matter and in cerebellar and brainstem regions versus cortical areas (27eC29). We therefore believe that our measurement of CBFV reactivity in the MCA should represent the blood flow changes in other regions of the brain, including the ventrolateral part of the brainstem, where the change of CO2 and pH is directly sensed by the central chemoreceptors and thereby affects the respiratory neurons (30, 31).
We used the linear regression analysis to describe each subject's CBFV/PETCO2 relationship. Although the published data showed uncertainty as to whether the relationship between CBFV and change of CO2 is linear across the hypo- and hypercapnic range of PETCO2 (32), we found that linear regression provided the best fit with r2 0.9 in most, though not all, of our subjects' data. We also analyzed the CBFV/PETCO2 above and below the eupneic PCO2 separately, which showed the same linear tendency as did the overall regression.
Mechanism of the Low CBFV/PETCO2 in Patients with CHF and CSA
Both the control and CSA groups showed the expected CBFV response to CO2 in the hyper- and hypocapnic ranges and demonstrated a similar differential responsiveness to hyper- and hypocapnia that has been observed in healthy subjects (33). The greater responsiveness to hypercapnia compared with hypocapnia may be related the observation that intracranial vascular tone is more influenced by vasodilator mediators compared with vasoconstrictive mediators (34). However, in the patients with CHF, the gain of the response is relatively small in both groups (2.4eC3.1% increase/mm Hg) compared with young, healthy subjects (4.7 ± 0.3% mm HgeC1) (17). This difference is likely to be related to the poor heart function and/or endothelial dysfunction. CHF is associated with impaired endothelium-dependent vasodilation (35). A low cerebrovascular CO2 reactivity with the heart dysfunction has been reported by Georgiadis and coworkers (8). Moreover, most of our subjects had an underlying ischemic cardiomyopathy, which increases the likelihood of endothelial dysfunction (36) and coexistent cerebrovascular disease. Thus, all patients with CHF are likely to have an abnormal responsiveness to CO2, but this does not explain the greater impairment in reactivity in patients with CSA.
The mechanism underlying the reduction in cerebrovascular reactivity in patients with CSA is not clear. CO2 modulates CBF via direct effect of extracellular pH on the smooth muscle of cerebral pial vessels (1). The vascular smooth muscle tone in the brain is also regulated by other neural, humoral, and physical factors. Our experiment does not allow us to determine the pathologic linkage between the impaired cerebrovascular response to CO2 and periodic breathing. The impaired cerebrovascular reactivity may also be a direct consequence of the periodic breathing. The supportive evidence is that a reduced CBF response to CO2 has also been observed in patients with obstructive sleep apnea, and this deficiency reverses after treatment with nasal continuous positive airway pressure (37). Both recurrent central and obstructive apneas are associated with asphyxia and surges in arterial pressure that could cause endothelial dysfunction and impairment of cerebrovascular reactivity to CO2. Sleep fragmentation as a result of overnight periodic breathing could also diminish hypercapnic vasomotor response (38). However, conflicting evidence has also been reported in the literature, with a normal or even higher CBF reactivity to PCO2 in patients with obstructive sleep apnea (39), and whether the endothelium is involved in the CBF response to CO2 is still controversial (40).
Patients with CSA were usually associated with excessive sympathetic activation (41), but sympathetic vasoconstrictor outflow does not appear to influence the cerebrovascular response to changes in PETCO2 (17, 42). Even if sympathetic vasoconstrictor outflow could blunt the hypercapnia-induced CBF increase as reported by Jordan and colleagues (43), it still cannot explain why patients with CSA also had a less vasoconstriction in response to hypocapnia. Although patients with CSA have a greater peripheral chemosensitivity to CO2 (5, 44), CBF responses to CO2 were not influenced by cutting peripheral chemoreceptor afferent nerves (29). Drug treatment received in the two groups was basically similar. All of our subjects were on medications for CHF that have vascular effects to improve cardiac output. However, literature shows that these medications have no overt effect on baseline CBF (45) or its dynamic changes (46).
Finally, published evidence suggests that brain hypoperfusion could attenuate cerebrovascular reactivity to CO2 (13, 47), and conversely, brain hyperperfusion could enhance the reactivity (48). A low resting CBF value has been observed in patients with CHF (6, 7), although the investigators did not clarify whether or not the patients had apneas. From the point of teleology, the low CBF response to hypocapnia could protect patients from severe ischemia in the presence of an already low cerebral perfusion. Transcranial Doppler technique does not allow us to compare the absolute CBF between individuals. Accordingly, we cannot rule out the possibility that patients with CSA may have a lower resting CBF, which results in a lower CBF response to CO2.
We investigated the possibility that a temporal abnormality in the CO2 stimulus and the vascular response could account for some of the difference between the patients with and without CSA. However, the same time lag between the two groups with and without CSA in our study suggests a similar circulatory transit delay from lung to brain and a similar reflection speed of cerebrovascular tone to increase of PaCO2.
We used breath-holds to mimic sleep apnea and to determine if patients with CSA would respond differently than patients without CSA. The CBFV response to breath-holds yielded a similar result observed during CO2 breathing (i.e., less increase in the CSA group than in the control subjects). The prolonged time delay between the onset of breath-hold and the peak in CBFV in both groups is due to the long circulation time, which was 16.2 to 17.2 seconds in our patients, compared with 6 to 10 seconds in subjects with a normal heart (20, 49). Thus, unlike normal subjects where two-thirds of the total increase in CBFV occurred before the end of the breath-hold (17), neither of our patient groups demonstrated an increase of CBFV during the breath-hold. Likewise, no increase in CBFV was seen during the spontaneously occurring apneas (Figure 6) or with previously reported examples of Cheyne-Stokes' respiration (46, 50). Thus, the absence of rise in CBFV during the apnea in both groups of our patients was likely due to the delay in arrival of the increasing alveolar PCO2 during the apnea to the cerebral vessels due to a prolonged circulation time. However, when the rising PCO2 secondary to the absence of ventilation eventually did reach the cerebral circulation after the breath-hold, the patients with CSA showed a smaller increase in CBFV because they have less of a cerebrovascular responsiveness to CO2.
We noticed that patients with CSA had a smaller increase in blood pressure in response to both exogenous hypercapnia and hypercapnia during the breath-hold. However, because of the marked autoregulation activity of the cerebral circulation, the ability of the perfusion pressure to influence CBF is minimal. Our previous study in normal subjects showed that approximately 16% of the increase in CBFV after the breath-hold was due to the associated increase in arterial pressure during and after the apnea (17). Therefore, part of the reason for the smaller rise in CBFV after the breath-hold in the patients with CSA may have been related to their absence of a blood pressure response to CO2 compared with the control subjects. However, the small rise in blood pressure noted in the control patients could not account for all of the difference in rise in CBFV after the breath-hold between the two groups.
Influence of Attenuated Cerebrovascular Reactivity to CO2 on Breathing Stability
A close relationship between changes in CBF and ventilatory output has been shown in animal studies. Simultaneous measurements of the peak diaphragmatic electromyogram and brain blood flow in anesthetized goats yielded a reciprocal relationship between changes in respiratory drive and CBF (51). Evidence for the modulation of increased ventilatory chemosensitivity by reduced CBF was provided by Chapman and his colleagues (13), who showed that when they reduced CBF by 30%, ventilatory responsiveness to CO2 was increased if expressed as E against PETCO2 or PaCO2, but was unchanged from baseline if plotted as E against cerebral venous PCO2. This observation suggests that CBF affects ventilation through modification of tissue PCO2 in the region of the medullary chemoreceptor area. With a poor brain perfusion, the central chemoreceptor will sense a relatively higher H+ for a given change of PaCO2 and will correspondingly trigger a more vigorous ventilatory response.
Cerebrovascular dysfunction may play a role in the etiology of CSA in patients with CHF. PaCO2 is a powerful regulator of the respiration as well as the cerebral circulation, and its ventilatory excitatory effect at the central chemoreceptor may be modulated by its influence on the cerebral circulation. In all mammals studied, hypercapnia causes cerebral vasodilation, whereas hypocapnia causes vasoconstriction (1). The increased CBF during hypercapnia leads to less increase in H+ at the central chemoreceptor, whereas decreased CBF during hypocapnia leads to less decrease in H+ at the central chemoreceptor. By buffering the change in H+ at the central chemoreceptor, an adequate change in CBF in response to CO2 will stabilize the breathing pattern in the face of transient perturbations in arterial CO2. Because patients with CHF and CSA have a reduced CBF reactivity to CO2, we expect them to have a greater increase in PCO2 and H+ at the central chemoreceptor during hypercapnia and a greater reduction in PCO2 and H+ during hypocapnia for a given change in PaCO2 compared with patients without CSA. Therefore, reduced CBF reactivity to CO2 should increase the ventilatory response to CO2, a major feature of CSA (4, 5, 44). When PaCO2 is reduced below eupneic levels by a ventilatory overshoot, brain PCO2 will fall more than normal because of a reduced cerebral vasoconstrictive response and thereby elevate the PaCO2 at which apnea occurs. The increased apneic threshold for PaCO2 will bring it closer to the eupneic PCO2, another critical feature of CSA (14).
Although posthyperventilation apnea is usually initiated by peripheral chemoreceptors (52), a large body of evidence supports the involvement of central mechanisms in the development of posthyperventilation apnea. First, carotid body hypocapnia by itself is not sufficient to produce apnea as shown by the presence of only reduced VT but not apnea during hypocapnic perfusion of an isolated carotid body preparation in the unanesthetized sleeping dog (53). Second, the delayed posthyperventilation apnea in the carotid body denervated dog unmasked the ability of the central chemoreceptors to cause apnea in response to hypocapnia (52). Finally, we have observed contradictory effects on breathing stability of two ventilatory stimulants whose effect is predominantly mediated by the peripheral chemoreceptors, hypoxia and almitrine, whereby hypoxia was destabilizing and almitrine was stabilizing (54). Likewise, two inhibitory agents mediated, in part, by the peripheral chemoreceptors, hyperoxia and dopamine, have different effects on ventilatory stability, with hyperoxia stabilizing and dopamine destabilizing (55). We suspect that hypoxia and hyperoxia may affect posthyperventilation breathing pattern through a combination of their peripheral and central effects (56, 57) as well as their influence on CBF (58). In other words, even though posthyperventilation apnea is mainly mediated by peripheral chemoreceptors, additional central influences are required to produce periodic breathing. The impaired CBF reactivity to CO2 might predispose patients to periodic breathing through a central mechanism.
In summary, we observed an attenuated cerebrovascular response to CO2 in patients with CHF and CSA. The compromise of the ability of CO2 to regulate CBF might disturb the stability of the breathing pattern by causing overshooting during hypercapnia and undershooting during hypocapnia. Thus, altered CBF may be an important contributor to breathing instabilities during sleep, which are associated with large fluctuations in CO2 and ventilation.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
REFERENCES
Heistad DD, Kontos HA. Cerebral circulation. In: Shepherd JT, Abboud FM, editors. Handbook of physiology section 2: the cardiovascular system. Bethesda, MD: American Physiological Society; 1983. pp. 137eC182.
Hanly P, Zuberi N, Gray R. Pathogenesis of Cheyne-Stokes respiration in patients with congestive heart failure: relationship to arterial PCO2. Chest 1993;104:1079eC1084.
Tkacova R, Hall ML, Liu PP, Fitzgerald FS, Bradley TD. Left ventricular volume in patients with heart failure and Cheyne-Stokes respiration during sleep. Am J Respir Crit Care Med 1997;156:1549eC1555.
Javaheri S. A mechanism of central sleep apnea in patients with heart failure. N Engl J Med 1999;341:949eC954.
Solin P, Roebuck T, Johns DP, Haydn Walters E, Naughton MT. Peripheral and central ventilatory responses in central sleep apnea with and without congestive heart failure. Am J Respir Crit Care Med 2000;162:2194eC2200.
Lee CW, Lee JH, Lim TH, Yang HS, Hong MK, Song JK, Park SM, Park SJ, Kim JJ. Prognostic significance of cerebral metabolic abnormalities in patients with congestive heart failure. Circulation 2001;103:2784eC2787.
Rajagopalan B, Raine AE, Cooper R, Ledingham JG. Changes in cerebral blood flow in patients with severe congestive cardiac failure before and after captopril treatment. Am J Med 1984;76:86eC90.
Georgiadis D, Sievert M, Cencetti S, Uhlmann F, Krivokuca M, Zierz S, Werdan K. Cerebrovascular reactivity is impaired in patients with cardiac failure. Eur Heart J 2000;21:407eC413.
Caparas SN, Clair MJ, Krombach RS, Hendrick JW, Houck WV, Kribbs SB, Mukherjee R, Tempel GE, Spinale FG. Brain blood flow patterns after the development of congestive heart failure: effects of treadmill exercise. Crit Care Med 2000;28:209eC214.
Chapman RW, Santiago TV, Edelman NH. Effects of graded reduction of brain blood flow on ventilation in unanesthetized goats. J Appl Physiol 1979;47:104eC111.
Lane DJ, Rout MW, Williamson DH. Mechanism of hyperventilation in acute cerebrovascular accidents. BMJ 1971;3:9eC12.
Schmidt CF. The influence of cerebral blood-flow on respiration. I. The respiratory responses to changes in cerebral blood flow. Am J Physiol 1928;84:202eC222.
Chapman RW, Santiago TV, Edelman NH. Effects of graded reduction of brain blood flow on chemical control of breathing. J Appl Physiol 1979;47:1289eC1294.
Xie AL, Skatrud JB, Puleo DS, Rahko PS, Dempsey JA. Apnea-hypopnea threshold for CO2 in patients with congestive heart failure. Am J Respir Crit Care Med 2002;165:1245eC1250.
Xie A, Khayat RN, Puleo DS, Morgan BJ, Skatrud JB. Cerebrovascular reponse to CO2 in patients with congestive heart failure . Am J Respir Crit Care Med 2003;167:A173.
Otis SM, Ringelstein EB. The transcranial Doppler examination: principles and applications of transcranial Doppler sonography. In: Tegeler CH, Babikian VL, Gomez CR, editors. Neurosonology. St. Louis, MO: Mosby; 1996. pp. 113eC128.
Przybylowski T, Bangash MF, Reichmuth K, Morgan BJ, Skatrud JB, Dempsey JA. Mechanisms of the cerebrovascular response to apnoea in humans. J Physiol 2003;548:323eC332.
Ameriso SF, Mohler JG, Suarez M, Fisher M. Morning reduction of cerebral vasomotor reactivity. Neurology 1994;44:1907eC1909.
Lorenzi-Filho G, Rankin F, Bies I, Douglas Bradley T. Effects of inhaled carbon dioxide and oxygen on Cheyne-Stokes respiration in patients with heart failure. Am J Respir Crit Care Med 1999;159:1490eC1498.
Dahan A, DeGoede J, Berkenbosch A, Olievier IC. The influence of oxygen on the ventilatory response to carbon dioxide in man. J Physiol 1990;428:485eC499.
Czosnyka M, Richards HK, Whitehouse HE, Pickard JD. Relationship between transcranial Doppler-determined pulsatility index and cerebrovascular resistance: an experimental study. Neurosurgery 1996;84:79eC84.
Giller CA, Bowman G, Dyer H, Mootz L, Krippnoer W. Cerebral arterial diameters during changes in blood pressure and carbon dioxide curing craniotomy. Neurosurgery 1993;32:737eC742.
Serrador JM, Picot PA, Rutt BK, Shoemaker JK, Bondar RL. MRI measures of middle cerebral artery diameter in conscious humans during simulated orthostasis. Stroke 2000;31:1672eC1678.
Valdueza JM, Balzer JO, Villringer A, Vogl TJ, Kutter R, Einhaupl KM. Changes in blood flow velocity and diameter of the middle cerebral artery during hyperventilation: assessment with MR and transcranial Doppler sonography. AJNR Am J Neuroradiol 1997;18:1929eC1934.
Pandit JJ, Mohan RM, Paterson ND, Poulin MJ. Cerebral blood flow sensitivity to CO2 measured with steady-state and Read's rebreathing methods. Respir Physiol Neurobiol 2003;137:1eC10.
Balfors EM, Franklin KA. Impairment of cerebral perfusion during obstructive sleep apneas. Am J Respir Crit Care Med 1994;150:1587eC1591.
Heistad DD, Marcus ML, Ehrhardt JC, Abboud FM. Effect of stimulation of carotid chemoreceptors on total and regional cerebral blood flow. Circ Res 1976;38:20eC25.
Ramsay SC, Murphy K, Shea SA, Friston KJ, Lammertsma AA, Clark JC, Adams L, Guz A, Frackowiak RSJ. Changes in global cerebral blood flow in humans: effect on regional cerebral blood flow during a neural activation task. J Physiol 1993;471:521eC534.
Sato A, Trzebski A, Zhou W. Local cerebral blood flow responses in rats to hypercapnia and hypoxia in the rostral ventrolateral medulla and in the cortex. J Auton Nerv Syst 1992;41:79eC86.
Kawai A, Ballantyne D, Muckenhoff K, Scheid P. Chemosensitive medullary neurones in the brainstemeCspinal cord preparation of the neonatal rat. J Physiol 1996;492:277eC292.
Thomas T, Spyer KM. ATP as a mediator of mammalian central CO2 chemoreception. J Physiol 2000;523:441eC447.
Poulin MJ, Liang PJ, Robbins PA. Fast and slow components of cerebral blood flow response to step decreases in end-tidal PCO2 in humans. J Appl Physiol 1998;85:388eC397.
Ide K, Eliasziw M, Poulin MJ. The relationship between middle cerebral artery blood velocity and end-tidal PCO2 in the hypocapnic-hypercapnic range in humans. J Appl Physiol 2003;95:129eC137.
Toda N, Okamura T. Cerebral vasodilators. Jpn J Pharmacol 1998;76:349eC367.
Nakamura M, Sugawara S, Arakawa N, Nagano M, Shizuka T, Shimoda Y, Sakai T, Hiramori K. Reduced vascular compliance is associated with impaired endothelium-dependent dilatation in the brachial artery of patients with congestive heart failure. J Card Fail 2004;10:36eC42.
Tentolouris C, Tousoulis D, Antoniades C, Bosinakou E, Kotsopoulou M, Trikas A, Toutouzas P, Stefanadis C. Endothelial function and proinflammatory cytokines in patients with ischemic heart disease and dilated cardiomyopathy. Int J Cardiol 2004;94:301eC305.
Diomedi M, Placidi F, Cupini LM, Bernardi G, Silvestrini M. Cerebral hemodynamic changes in sleep apnea syndrome and effect of continuous positive airway pressure treatment. Neurology 1998;51:1051eC1056.
Qureshi AI, Christopher Winter W, Bliwise DL. Sleep fragmentation and morning cerebrovasomotor reactivity to hypercapnia. Am J Respir Crit Care Med 1999;160:1244eC1247.
Droste DW, Ludemann P, Anders F, Kemeny V, Thomas M, Krauss JK, Ringelstein EB. Middle cerebral artery blood flow velocity, end-tidal pCO2 and blood pressure in patients with obstructive sleep apnea and in healthy subjects during continuous positive airway pressure breathing. Neurol Res 1999;21:737eC741.
Toda N, Ayajiki K, Enokibori M, Okamura T. Monkey cerebral arterial relaxation caused by hypercapnic acidosis and hypertonic bicarbonate. Am J Physiol 1993;265:H929eCH933.
Van de Borne P, Oren R, Abouassaly C, Anderson E, Somers VK. Effect of Cheyne-Stokes respiration on muscle sympathetic nerve activity in severe congestive heart failure secondary to ischemic or idiopathic dilated cardiomyopathy. Am J Cardiol 1998;81:432eC436.
LeMarbre G, Stauber S, Khayat RN, Puleo DS, Skatrud JB, Morgan BJ. Baroreflex-induced sympathetic activation does not alter cerebrovascular CO2 responsiveness in humans. J Physiol 2003;551:609eC616.
Jordan J, Shannon JR, Diedrich A, Black B, Costa F, Robertson D, Biaggioni I. Interaction of carbon dioxide and sympathetic nervous system activity in the regulation of cerebral perfusion in humans. Hypertension 2000;36:383eC388.
Xie A, Rutherford R, Rankin F, Wong B, Bradley TD. Hypocapnia and increased ventilatory responsiveness in patients with idiopathic central sleep apnea. Am J Respir Crit Care Med 1995;152:1950eC1955.
Waldemar G, Paulson OB. Angiotensin converting enzyme inhibition and cerebral circulation: a review. Br J Clin Pharmacol 1989;28:177SeC182S.
Franklin KA, Sandstrom E, Johansson G, Balfors EM. Hemodynamics, cerebral circulation, and oxygen saturation in Cheyne-Stokes respiration. J Appl Physiol 1997;83:1184eC1191.
Kagstrom E, Smith ML, Siesjo BK. Cerebral circulatory responses to hypercapnia and hypoxia in the recovery period following complete and incomplete cerebral ischemia in the rat. Acta Physiol Scand 1983;118:281eC291.
Hauge A, Nicolaysen G, Thoresen M. Acute effects of acetazolamide on cerebral blood flow in man. Acta Physiol Scand 1983;117:233eC239.
Khoo MC, Kronauer RE, Strohl KP, Slutsky AS. Factors inducing periodic breathing in humans: a general model. J Appl Physiol 1982;53:644eC659.
Karp HR, Sieker HO, Heyman A. Cerebral circulation and function in Cheyne-Stokes respiration. Am J Med 1961;30:861eC870.
Parisi RA, Neubauer JA, Frank MM, Santiago TV, Edelman NH. Linkage between brain blood flow and respiratory drive during rapid-eye-movement sleep. J Appl Physiol 1988;64:1457eC1465.
Nakayama H, Smith CA, Rodman JR, Skatrud JB, Dempsey JA. Carotid body denervation eliminates apnea in response to transient hypocapnia. J Appl Physiol 2003;94:155eC164.
Smith CA, Saupe KW, Henderson KS, Dempsey JA. Ventilatory effects of specific carotid body hypocapnia in dogs during wakefulness and sleep. J Appl Physiol 1995;79:689eC699.
Nakayama H, Smith CA, Rodman JR, Skatrud JB, Dempsey JA. Effect of ventilatory drive on carbon dioxide sensitivity below eupnea during sleep. Am J Respir Crit Care Med 2002;165:1251eC1260.
Chenuel BJ, Smith CA, Henderson KS, Dempsey JA. Ventilatory instability induced by selective carotid body inhibition in the sleeping dog. Adv Exp Med Biol 2004;551:197eC201.
Horn EM, Waldrop TG. Oxygen-sensing neurons in the caudal hypothalamus and their role in cardiorespiratory control. Respir Physiol 1997;110:219eC228.
Neubauer JA, Sunderram J. Oxygen-sensing neurons in the central nervous system. J Appl Physiol 2004;96:367eC374.
Cohen PJ, Alexander SC, Smith TC, Reivich M, Wollman H. Effects of hypoxia and normocarbia on cerebral blood flow and metabolism in conscious man. J Appl Physiol 1967;23:183eC189.(Ailiang Xie, James B. Ska)