INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

TO INVESTIGATE THE PATHOGENETIC link between sleep-disordered breathing and hypertension, our laboratory has conducted two previous studies of the neurocirculatory consequences of acute exposure to blood-gas perturbations in healthy subjects during wakefulness. The major finding of these studies is that 20-min exposure to asphyxia (combined hypoxia and hypercapnia), applied in either sustained or intermittent fashion, caused sympathoexcitation that persisted after withdrawal of the chemical stimuli (31, 49). In contrast, exposure to hypercapnia, which was delivered with a hyperoxic background that greatly attenuated input from carotid chemoreceptors (21), did not cause sustained increases in muscle sympathetic nerve activity. These findings suggest that either hypoxia or peripheral chemoreceptor stimulation (by hypoxia or hypercapnia) is required to produce persistent sympathetic activation. To clarify the contribution of hypoxia vs. hypercapnia to the sympathetic carryover effect produced by asphyxia, we exposed healthy subjects to isocapnic hypoxia and normoxic hypercapnia on 2 separate days and examined the neurocirculatory and ventilatory responses to these two chemical stimuli individually. Our rationale was that, if the carryover effect on sympathetic outflow requires nonspecific peripheral chemoreceptor stimulation, we should see persistent sympathoexcitation after exposure to both hypoxia and hypercapnia. On the other hand, if this carryover effect is dependent on hypoxia acting either peripherally or centrally, a persistent increase in sympathetic outflow should be observed only after the hypoxic exposure.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Fourteen healthy volunteers [9 men and 5 women, aged 28 ± 9 (SD) yr] served as subjects. All were free from cardiovascular, pulmonary, and neurological diseases as evaluated by history and physical examination. All subjects were nonsmokers, and they were taking no medications. This study was approved by the University of Wisconsin Health Sciences Human Subjects Committee. All subjects provided informed consent.

General procedures. All experiments commenced at the same time of day (1300), and the room temperature was maintained at 24 ± 1°C. Subjects were studied in the supine position. Respiratory and cardiovascular variables and sympathetic nerve activity were recorded continuously on paper (model TA 4000; Gould, Cleveland, OH) and on magnetic tape (model 4000A, PCM; Vetter, Rebersburg, PA). The signals were also routed to a computer for off-line analysis of the data.

Respiratory variables. Subjects breathed through a leak-free nasal mask to which a pneumotachograph (model 5719; Hans Rudolph, Kansas City, MO) was attached for measurement of tidal volume and breathing frequency. Minute ventilation (E) was calculated by multiplying tidal volume by breathing frequency. End-tidal O₂ (PET_O2) and CO₂ tensions (PET_CO2) were sampled from the mask and measured by gas analyzers (models S-3A/I and CD-3; Ametek, Pittsburgh, PA). Arterial O₂ saturation (Sa_O2) was measured by using a pulse oximeter (Biox model 3740; Ohmeda, Madison, WI).

Cardiovascular variables. Heart rate was taken from the electrocardiogram, and arterial pressure was measured at 1-min intervals by using an automated arm-cuff sphygmomanometer (Dinamap model 1846SX/P; Critikon, Tampa, FL). Blood flow was measured in the left leg by venous occlusion plethysmography (14). During this measurement, the limb was elevated above the level of the right atrium to collapse the veins. A collecting cuff positioned above the knee was rapidly inflated to 40 mmHg for four to five cardiac cycles every 30 s. Sixty-second averages of the blood flow measurements were used in the computation of results. Vascular resistance was calculated as mean arterial pressure (one-third pulse pressure + diastolic pressure) divided by blood flow (in ml · 100 ml · tissue¹ · min¹).

Sympathetic nerve activity. Postganglionic muscle sympathetic nerve activity in the right fibular (peroneal) nerve was recorded by using intraneural microelectrodes (42). Placement of the recording electrode in a muscle nerve fascicle was confirmed by the 1) presence of muscle twitches, not paresthesias, in response to electrical stimulation; 2) characteristic pattern of discharge in pulse-synchronous bursts; 3) afferent activity in response to tapping or stretching of muscle, but not gentle stroking of skin, in the appropriate receptive fields; and 4) absence of neural activation in response to arousal stimuli. The neural signals were amplified (total gain = 100,000), filtered (band pass = 700-2,000 Hz), and integrated (time constant = 100 ms); and signal-to-noise ratio > 3:1 was required for an acceptable neural recording. Sympathetic bursts were identified by inspection of the mean voltage neurogram (48). For purposes of quantification, muscle sympathetic nerve activity was expressed as burst frequency (bursts/min) and burst amplitude (arbitrary units).

Experimental protocols. The 20-min hypoxic and hypercapnic exposures occurred on 2 separate days at least 1 mo apart. The order of the exposures was randomized, with 7 out of 14 subjects receiving the hypoxic exposure first. In the isocapnic hypoxia trial, 95% nitrogen was added to the breathing circuit for two to three breaths, followed by 10-12% O₂, so that Sa_O2 fell to 80%. PET_CO2 was maintained at the eupneic level by increasing the inspired CO₂ fraction as necessary. In the normoxic hypercapnia trial, a gas mixture of 5-8% CO₂ was added to the breathing circuit so that PET_CO2 rose by +7 Torr. To prevent the rise in arterial PO₂ that normally accompanies exposure to hypercapnia, we added N₂, as necessary, to the breathing circuit. Thus Sa_O2 was maintained at the room-air baseline level during the CO₂ breathing. In the second trial, we titrated the chemical stimuli to match, as closely as possible, the amount of sympathetic activation caused by the initial trial. Each exposure was preceded by at least 10 min of stable baseline measurements of neurocirculatory and ventilatory variables, and it was followed by a 20-min recovery period of room-air breathing.

Data analysis. In five subjects, we were not able to maintain stable nerve recordings during and after both exposures. Therefore, the nerve traffic and ventilatory data analyses are based on the responses of nine subjects [3 women and 6 men, aged 29 ± 9 (SD) yr]. In three subjects, we were not able to obtain adequate leg blood flow measurements during both exposures. The cardiovascular data analyses are based on 11 subjects [4 women and 7 men, aged 26 ± 7 (SD) yr]. To characterize the chemical stimuli associated with each exposure, average values for end-tidal gases and Sa_O2 were computed for the baseline, intervention, and recovery periods. To quantify the neurocirculatory and ventilatory responses to hypoxia and hypercapnia, 5-min averages of neurocirculatory and ventilatory variables obtained during the baseline, exposure, and recovery periods were compared by two-way, repeated-measures ANOVA. When the overall F tests for time-by-condition (i.e., hypoxia or hypercapnia) interactions were statistically significant (P < 0.05), Student-Newman-Keuls post hoc tests were used to identify significant differences between the hypoxic and hypercapnic exposures at each time point and to compare measurements made during and after the exposures with the respective baseline values. We examined the association between the ventilatory response and amount of sympathetic activation during the chemical challenges by applying Pearson's product-moment correlation analysis to 5-min averages of E and sympathetic burst frequency measured before, during, and after the interventions. Except where otherwise noted, data are reported as means ± SE in the text, Table 1, and Figs. 1-4. P values < 0.05 were considered statistically significant.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Original record showing a representative subject's ventilatory and neurocirculatory responses before, during, and after exposure to isocapnic hypoxia (A) and normoxic hypercapnia (B). Ventilation increased during both exposures and returned to baseline after return to room-air breathing. Sympathetic burst frequency was increased approximately twofold with both exposures. Burst frequency remained elevated at the 20th and 60th min after withdrawal of hypoxic stimulus, whereas it returned to baseline during the recovery period after hypercapnia. Note: for subject comfort, the respiratory circuit was disconnected after 20 min of recovery. SaO2, arterial O2 saturation; PETCO2, end-tidal CO2 tension.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Group mean ± SE values (n = 9) for sympathetic burst frequency (A) and minute ventilation (B) before, during, and after 20-min exposure to isocapnic hypoxia () and normoxic hypercapnia (). The dashed vertical lines indicate the duration of exposure. * Significant difference vs. baseline, P < 0.05. + Significant difference hypoxia vs. hypercapnia, P < 0.05.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Group mean ± SE values (n = 11) for heart rate (A) and arterial blood pressure (B) before, during, and after 20-min exposure to isocapnic hypoxia ( or ) and normoxic hypercapnia ( or ). The dashed vertical lines indicate the duration of exposure. SBP, systolic blood pressure; DBP, diastolic blood pressure. * Significant difference vs. baseline, P < 0.05. There was no difference between hypoxia vs. hypercapnia at any time point.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Group mean ± SE values (n = 11) showing the changes in leg blood flow (A) and leg vascular resistance (B) before, during, and after 20-min exposure to isocapnic hypoxia () and normoxic hypercapnia (). Neither exposure produced statistically significant changes in leg blood flow or vascular resistance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ventilatory response to hypoxia and hypercapnia. During the initial visit, the chemical challenges were designed to produce either a decrease in Sa_O2 to 80% (if hypoxia was administered first) or an increase in PET_CO2 of +7 Torr (if hypercapnia was administered first). During the second visit, we titrated the chemical stimuli to match the amount of sympathetic activation observed in the first visit. During isocapnic hypoxia, the average Sa_O2 was 80.2 ± 1% (range, 77-87%) with PET_CO2 maintained at the eupneic level (43.2 ± 1.0 vs. 41.8 ± 2.1 Torr, P > 0.05). During normoxic hypercapnia, the average increase in PET_CO2 was +6.5 Torr (range, 5.3-8.6 Torr) above the eupneic level, and neither Sa_O2 (97.6 ± 0.3 vs. 97.3 ± 0.3%) nor PET_O2 (110 ± 4 vs. 102 ± 4 Torr) was changed relative to the eupneic levels (both P > 0.05).

Effect of hypoxia and hypercapnia on muscle sympathetic nerve activity. By design, sympathetic burst frequency increased to the same extent during the hypoxia and hypercapnia trials (Figs. 1 and 2). During the final 5 min of the exposures, sympathetic burst frequency had increased to 167 ± 29% of baseline in hypoxia and 171 ± 23% of baseline in hypercapnia (P > 0.05). After the hypoxic exposure, burst frequency remained elevated for the duration of the recovery period. In contrast, after hypercapnic exposure, burst frequency returned to the baseline level within 5 min after resumption of room-air breathing (114 ± 8% of baseline, P > 0.05). The difference in burst frequency between hypoxia and hypercapnia trials was statistically significant during the final 10 min of the recovery period. Neither hypoxia nor hypercapnia affected sympathetic burst amplitude. There was a very strong association between sympathetic burst frequency and E in the hypercapnia trial (r = 0.98, P < 0.05), but only weak association between the two variables in the hypoxia trial (r = 0.53, P > 0.05). Figure 1 shows a representative subject, in whom we were able to observe sympathetic nerve activity for 60 min after removing the chemical stimuli in both the hypoxia and hypercapnia trials. In this subject, the hypoxia-induced sympathetic activation persisted for the entire 60-min recovery period, whereas sympathetic activation was not maintained in the recovery period after hypercapnic exposure.

Cardiovascular response to hypoxia and hypercapnia. Both hypoxia and hypercapnia caused significant increases in heart rate and systolic blood pressure that returned to baseline levels after the withdrawal of chemical stimuli (Fig. 3). Neither exposure affected diastolic pressure. Hypoxic exposure caused small, statistically insignificant increases in leg blood flow and decreases in leg vascular resistance, whereas, during the hypercapnic exposure, these variables remained stable (Fig. 4).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Previously, we found that brief exposure to asphyxia (combined hypoxia and hypercapnia) produced a memory-like effect whereby sympathetic outflow to skeletal muscle remained elevated after withdrawal of the chemical stimuli. In the present study, we sought to determine the relative importance of hypoxia and hypercapnia in causing this persistent sympathetic activation. The major findings are that exposure to hypoxia evoked an increase in sympathetic nerve activity that outlasted the chemical stimulus. In contrast, exposure to hypercapnia, which caused a comparable acute increase in sympathetic nerve activity, failed to produce long-lasting sympathetic activation.

Critique of methods. Because the hypoxia and hypercapnia exposures were conducted on 2 separate days, day-to-day variability in baseline levels of sympathetic outflow or in sympathetic responsiveness may have contributed to our findings. To minimize this possibility, we conducted the hypoxia trials and hypercapnia trials in random order. Also, our primary analysis of the carryover effect on muscle sympathetic nerve activity was based on burst frequency measurements. In our experience and that of other investigators, indexes of burst frequency are stable in repeated measures made on separate days (<15%) (42). Finally, we normalized the sympathetic nerve activity data by expressing responses to hypoxia and hypercapnia as the percentage of each individual's baseline level. Therefore, we consider it unlikely that our finding of persistent sympathetic activation after exposure to hypoxia, but not hypercapnia, can be explained on the basis of day-to-day variability in sympathetic outflow.

Neurocirculatory and ventilatory responses to hypoxia and hypercapnia. In the present study, exposures to both hypoxia and hypercapnia caused substantial increases in sympathetic outflow and E. We titrated the level of the chemical stimulus in the second exposure to match the amount of sympathetic activation produced by the initial exposure; therefore, by design, the sympathetic responses to hypoxia and hypercapnia were similar. In contrast, the ventilatory responses to the two stimuli were quite different. After comparable initial increases, E increased progressively during continued exposure to hypercapnia, whereas it declined during continued exposure to hypoxia. This "roll-off" phenomenon for E, which has been described by previous investigators (3, 22, 44), was not accompanied by a parallel decrease in sympathetic activity. Consequently, the levels of E and sympathetic activation were not correlated during exposure to hypoxia, whereas they were well-correlated during hypercapnia. This disparity in the neurocirculatory and ventilatory responses to the two chemical stimuli was especially evident during the recovery periods after the exposures. Ventilation and sympathetic nerve activity decreased in parallel to baseline levels after the hypercapnic exposure. In contrast, even though E normalized promptly after removal of the hypoxic stimulus, hypoxia-induced sympathetic activation persisted into the recovery period.

Neural pathways involved in hypoxia-induced sympathetic activation. Exposure to hypoxia evokes increases in sympathetic outflow via two partially distinct mechanisms: 1) stimulation of peripheral chemoreceptors and 2) activation of oxygen-sensitive cells within the brain stem. Hypoxic stimulation of the carotid chemoreceptors produces reflex increases in sympathetic outflow to the heart and to the renal, splanchnic, and skeletal muscle vascular beds (30). Afferent nerve signals traveling in the carotid sinus and hypoglossal nerves are relayed via the nucleus of the solitary tract to those neurons in the rostral ventrolateral medulla (RVLM) that are thought to be responsible for generating sympathetic "tone" (see Ref. 15 for detailed description). The respiratory periodicity that is evident in the discharge of these presympathetic neurons (and also in some pre- and postganglionic sympathetic neurons) suggests that they are linked to the central respiratory rhythm generator. Nevertheless, in anesthetized animals, elimination of all central respiratory activity by muscimol injection abolished respiratory oscillations in postganglionic nerve traffic but did not attenuate the sympathetic response to stimulation of peripheral chemoreceptors (25). In addition, increases in cervical preganglionic sympathetic nerve activity could be elicited by peripheral chemoreceptor stimulation even when phrenic discharge was abolished by hyperventilation hypocapnia (23). During hyperventilation-induced apnea in humans, muscle sympathetic nerve activity increased before resumption of respiratory activity in response to gradually accumulating chemoreflex stimuli (38). Taken together, these findings suggest that some chemoreceptor information can reach the presympathetic RVLM neurons without an intervening relay within the respiratory network.

Mechanisms of hypercapnia-induced sympathetic activation. Increases in PCO₂ stimulate sympathetic outflow by both peripheral (carotid chemoreceptor) and central neural mechanisms (12, 16, 27). The cells responsible for mediating the central component of the neurocirculatory response to hypercapnia are thought to be located in RVLM regions that have some overlap with central CO₂ chemoreceptors (27). Are these the same medullary cells that respond to hypoxia? The observation that hypercapnia did not excite presympathetic neurons in the RVLM that were activated by hypoxia (39), taken together with the finding that hypercapnia and hypoxia caused different patterns of Fos expression in medullary neurons (40), suggests that central processing of hypoxic and hypercapnic stimuli occurs at distinct locations. As is the case with hypoxia, existing evidence suggests that CO₂-induced increases in sympathetic outflow can occur independently of increases in activity of central respiratory network (24, 38, 41).

Mechanism for the persistent sympathetic activation after hypoxic exposure. The present findings demonstrate that generic chemoreceptor stimulation is not sufficient to produce an aftereffect on sympathetic discharge; instead, hypoxia is required. We consider it unlikely that the marked increase in sympathetic outflow evoked during hypoxic exposure could, per se, have had a stimulatory effect on postganglionic activity, because hypercapnia-induced sympathoexcitation of comparable magnitude failed to do so. Moreover, muscle sympathetic activity is increased substantially by other experimental interventions (e.g., lower body suction, static exercise), yet it returns promptly to baseline levels after removal of the stimuli (29, 43). We also consider it unlikely that the sustained sympathetic activation that we observed after hypoxic exposure was caused by continued chemoreceptor stimulation because end-tidal gases (and presumably arterial blood gases) normalized soon after withdrawal of the stimuli. The finding that ventilation also returned to baseline promptly after the exposure in our study is consistent with this notion. Instead, our findings point to the possibility of long-lasting facilitation of sympathetic outflow that is dependent on prior hypoxic stimulation. We speculate that this memory-like effect occurs in RVLM neurons responsible for generating sympathetic "tone" and that it is caused by a modulatory effect of brain hypoxia on central processing of afferent input from peripheral chemoreceptors.

Absence of a hypoxia-induced aftereffect on ventilation. The slowly disappearing posthypoxic effect on sympathetic outflow that we observed may be analogous to the ventilatory acclimatization that maintains breathing above baseline levels for hours and even days after exposure to the hypoxia of high altitude (2). If so, why didn't our subjects demonstrate a persistent increase in ventilation after hypoxic exposure? One possibility is that we maintained eucapnia during the hypoxic exposures. Previous studies that employed isolated carotid body perfusion to increase ventilation suggest that brain hypoxia and hypocapnia are both required to produce long-term facilitation of breathing (4).

Hemodynamic responses to hypoxia vs. hypercapnia. Both hypoxia and hypercapnia caused increases in heart rate and systolic blood pressure that returned to baseline during the recovery period. The lack of change in diastolic blood pressure is consistent with the unchanged or even decreased peripheral vascular resistance that we observed in our study. We speculate that increases in sympathetic vasoconstrictor outflow did not raise leg vascular resistance because of the overriding vasodilatory influence of hypoxia and hypercapnia on vascular smooth muscle (12, 17). This idea is consistent with recent information from a canine model of obstructive sleep apnea (36). In that study, systemic blood pressure fell during apneas that occurred under conditions of autonomic blockade, an effect that was prevented by administration of supplemental oxygen. The authors interpreted this finding to mean that the neurally mediated increase in blood pressure caused by hypoxia occurs against a background of concomitant vasodilation. Although the sympathoexcitation observed in our experiments did not cause vasoconstriction in the leg, this chemoreflex-induced increase in muscle sympathetic nerve activity may have supported blood pressure by preventing marked vasodilation.