Sympathetic neural overactivity in healthy humans after prolonged exposure to hypobaric hypoxia

Abstract

Acute exposure to hypoxia causes chemoreflex activation of the sympathetic nervous system. During acclimatization to high altitude hypoxia, arterial oxygen content recovers, but it is unknown to what degree sympathetic activation is maintained or normalized during prolonged exposure to hypoxia. We therefore measured sympathetic nerve activity directly by peroneal microneurography in eight healthy volunteers (24 ± 2 years of age) after 4 weeks at an altitude of 5260 m (Chacaltaya, Bolivian Andes) and at sea level (Copenhagen). The subjects acclimatized well to altitude, but in every subject sympathetic nerve activity was highly elevated at altitude vs. sea level (48 ± 5 vs. 16 ± 3 bursts min⁻¹, respectively, P < 0.05), coinciding with increased mean arterial blood pressure (87 ± 3 vs. 77 ± 2 mmHg, respectively, P < 0.05). To examine the underlying mechanisms, we administered oxygen (to eliminate chemoreflex activation) and saline (to reduce cardiopulmonary baroreflex deactivation). These interventions had minor effects on sympathetic activity (48 ± 5 vs. 38 ± 4 bursts min⁻¹, control vs. oxygen + saline, respectively, P < 0.05). Moreover, sympathetic activity was still markedly elevated (37 ± 5 bursts min⁻¹) when subjects were re-studied under normobaric, normoxic and hypervolaemic conditions 3 days after return to sea level. In conclusion, acclimatization to high altitude hypoxia is accompanied by a striking and long-lasting sympathetic overactivity. Surprisingly, chemoreflex activation by hypoxia and baroreflex deactivation by dehydration together could account for only a small part of this response, leaving the major underlying mechanisms unexplained.

Activation of the sympathetic nervous system produces many of the circulatory adjustments to stressors such as exercise, orthostatic stress, bleeding and hypoxia. Acute hypoxia stimulates oxygen sensitive chemoreceptors in the carotid body (Marshall, 1994; Prabhakar, 2000) and the brainstem (Reis et al. 1994; Solomon, 2000), and increases efferent sympathetic outflow in humans, as demonstrated by a number of studies using direct microneurographic recordings of sympathetic discharge to the skeletal muscle vascular bed (Saito et al. 1988; Rowell et al. 1989; Somers et al. 1989, 1991; Leuenberger et al. 1991; Morgan et al. 1995; Duplain et al. 1999; Hansen et al. 2000; Xie et al. 2001). This increase in vasoconstrictor drive is thought to counteract hypoxic vasodilator mechanisms (Rowell et al. 1986; Saito et al. 1988; Rowell et al. 1989) and maintain arterial blood pressure.

With prolonged exposure to hypobaric hypoxia at high altitude, acclimatization processes partially restore arterial oxygen content within a few weeks, thus apparently reducing the sympathoexcitatory chemoreceptor drive. It has therefore previously been thought that sympathetic nerve activity would decrease towards sea level values over a similar time-course of acclimatization. However, studies using norepinephrine (noradrenaline) (NE) concentrations or spillover as indirect measures of sympathetic nerve activity have suggested continued or only slowly declining sympathoexcitation during the first weeks of acclimatization (Cunningham et al. 1965; Bärtsch et al. 1991; Reeves et al. 1992; Mazzeo et al. 1994; Antezana et al. 1994; Kanstrup et al. 1999). Furthermore, it has been documented that sojourn at high altitude is accompanied by increased blood pressure, possibly sympathetically mediated (Reeves et al. 1992; Kanstrup et al. 1999). These observations suggest that acclimatization to high altitude paradoxically is accompanied by maintained sympathoexcitation. A major methodological concern has been that urine and plasma NE levels may not accurately reflect sympathetic neural activation during hypoxia, because of alterations in catecholamine re-uptake and metabolism (Rowell et al. 1989; Leuenberger et al. 1991). So far, direct recordings of sympathetic nerve activity in humans during prolonged exposure to hypobaric hypoxia have not been available.

The two aims of the present investigation were: first, to test the hypothesis that sympathetic nerve activity is elevated in humans well acclimatized to hypobaric hypoxia. Second, to probe the underlying mechanisms driving such sympathoexcitation. To do so, we performed direct microneurographic measurements of muscle sympathetic nerve activity in healthy lowlanders on three separate occasions: after a 4 week sojourn at an altitude of 5260 m; at sea level 3 days following descent and at sea level 4–6 months post expedition (sea level control). To test the contributions of hypoxic chemoreflex activation and cardiopulmonary baroreflex unloading to altitude induced sympathoexcitation, measurements were also performed during interventions designed to eliminate the contribution of these putative mechanisms alone or in combination (oxygen breathing and volume loading with intravenous saline).

Methods

The study was approved by the Ethics Committees for La Paz, Bolivia and for Copenhagen and Frederiksberg, Denmark. Informed written consent was obtained from the subjects prior to participation. We studied eight healthy Danish lowlanders (four females and four males) aged 24 ± 2 years on three occasions: (1) after a 4 week sojourn at an altitude of 5260 m (Mt. Chacaltaya in the Bolivian Andes), (2) at sea level (Copenhagen, Denmark) on day 3 after return from altitude (descent), and (3) at sea level (Copenhagen, Denmark) 4–6 months post-expedition (sea level control). The studies at high altitude were performed after 32 ± 1 days of acclimatization in all subjects; and descent studies were carried out 3 days after return in six subjects and 5 days after in one subject. The descent study on the 8th subject was abandoned since microneurographic data could not be obtained. For the descent study, all subjects travelled by air from La Paz, Bolivia to Copenhagen, Denmark. This journey takes 24 hours, and was included as 1 day of return to sea level.

Subjects were always studied in the supine position. An antecubital venous catheter was inserted for infusion of saline for the purpose of volume loading. We counted respiratory excursions over several minutes, and continuously measured heart rate (HR, ECG), arterial blood pressure (automated oscillometry, Spengler, Cachan, France) and muscle sympathetic nerve activity (muscle SNA, peroneal microneurography, UIowa Bioengineering, Iowa City, Iowa, USA). Calf blood flow (venous occlusion plethysmography, D. E. Hokanson Inc., Bellevue, Washington, USA) was measured during baseline but not during interventions. Oxygen (purity 99 % or higher) was administered via a Douglas bag reservoir and mouthpiece. Oxygen saturation was measured continuously during both baseline and interventions by pulse oximetry (Satlite, Datex, Helsinki, Finland). Analyses of blood gases (ABL 510, Radiometer, Copenhagen, Denmark), haemoglobin and haematocrits are available for ambient air breathing but not during the interventions.

Peroneal microneurography

Multiunit recordings of muscle SNA were obtained with unipolar Tungsten microelectrodes inserted into muscle nerve fascicles of the peroneal nerve (Vallbo et al. 1979). The neural signals were amplified (95.5 × 10³), filtered (bandwidth 700–2000 Hz), rectified, and integrated (time constant 0.1 s) to obtain a mean voltage neurogram. A recording of muscle SNA was acceptable when the neurogram revealed spontaneous pulse-synchronous bursts, with a minimum signal-to-noise ratio of 3:1, that increased during phases II and III of the Valsalva manoeuvre, but not during loud noise arousal. The neural recording was allowed to stabilize for at least 10 min before recording data for analysis. Sympathetic bursts were detected by inspection of the mean voltage neurograms and muscle SNA expressed as number of bursts per minute and per hundred heart beats (100 beats)⁻¹.

In eight subjects (all different from the subjects included in the present study) the day-to-day variation when microneurographic recordings of muscle SNA were performed twice at sea level was 17.9 % (the means ± s.e.m. were 15 ± 2 and 20 ± 4 bursts min⁻¹, n = 8). This variability is similar to previously published reports (Sundlöf & Wallin, 1977).

Specific protocols

Resting data were averaged over 10 min. The contribution of chemoreflex activation and cardiopulmonary baroreflex deactivation to sympathetic nerve activity was assessed by: (1) 100 % oxygen breathing for 15 min to eliminate hypoxic chemoreceptor drive (oxygen alone); (2) infusion of 1000 ml of saline intravenously over 15 min to minimize cardiopulmonary baroreceptor unloading during continued oxygen breathing (oxygen + volume expansion) and (3) volume expansion alone 4 min after discontinued oxygen breathing. Data for analysis were averaged over the last 2 min of each intervention.

During oxygen breathing, decreased ventilatory drive could cause confounding hypercapnia-induced muscle SNA activation (Somers et al. 1989; Seals et al. 1993). To minimize hypercapnia, subjects performed paced breathing by following a metronome set to the individual resting respiratory frequency (counted during ambient air breathing).

The rationale for the ‘descent’ study on day 3 after return from altitude to sea level was to further assess the contribution of chemoreflex activation and cardiopulmonary baroreflex deactivation. At this time point, the subjects (n = 7) were studied after prolonged re-exposure to normoxia and normobaria. Furthermore, subjects were expected to be in a state of hypervolaemia, as plasma and blood volume rapidly increase upon descent from altitude (Svedenhag et al. 1997; Robach et al. 2000).

Data analysis

Mean arterial pressure (MAP) was calculated as diastolic pressure + ⅓ pulse pressure. Calf vascular resistance was calculated as MAP divided by calf blood flow. All data are expressed as means ± s.e.m. Statistical analysis was performed using Student's t test for paired comparisons of baseline parameters between the three different study days and parameters before and after intervention within the same study day. The significance level was set at P = 0.05 and adjusted using Bonferroni's method as appropriate.

Results

Acclimatization to high altitude is accompanied by marked increases in muscle sympathetic nerve activity

Two out of the eight subjects reported mild to moderate degrees of high altitude sickness during the first week of high altitude exposure. However, after 4 weeks at altitude all eight subjects were well acclimatized. Although pulse oximetry and arterial blood gasses showed significantly decreased oxygen saturation compared with sea level, blood oxygen content more than recovered due to increased haemoglobin concentration (Table 1). Despite this expected acclimatization, muscle SNA burst frequency was tripled after 4 weeks at high altitude vs. 4–6 months post-expedition sea level control (48 ± 5 vs. 16 ± 3 bursts min⁻¹, respectively, P < 0.05; Fig. 1 and Fig. 2). HR, MAP and calf vascular resistance increased moderately but significantly, whereas calf blood flow was decreased at altitude (Fig. 2, Table 2).

Table 1.

Arterial blood gasses and haematocrit after acclimatization to high altitude, 3 days after return (descent), and 4–6 months after return to sea level

	Sea level(n = 8)	Altitude 5260m (n = 8)	Descent (n = 7)
pH (units)	7.40 ± 0.01	7.47 ± 0.01*	7.40 ± 0.01
Pa,O2, (kPa)	13.3 ± 0.3	6.7 ± 0.1*	14.5 ± 0.4
Pa,CO2(kPa)	5.3 ± 0.2	3.1 ± 0.1*	4.3 ± 0.1
O2 sat. (%)	97.5 ± 0.2	84.8 ± 0.1*	97.5 ± 0.2
Hb(gl−1)	134 ± 5	182 ± 5*	138 ± 6
Hct (fraction)	0.42 ± 0.01	0.51 ± 0.01*	0.40 ± 0.01
CaO2 (ml O2 l−1)	175 ± 6	207 ± 7*	181 ± 8

P_a,O2 arterial oxygen pressure; P_a,CO2, arterial carbon dioxide pressure. O₂ sat., oxygen saturation; Hb, haemoglobin; Hct, haematocrit; C_a,O2, oxygen content in arterial blood

^*P < 0.05vs. sea level (n = 8), and P < 0.05vs. descent (n = 7).

Figure 1. Segments of mean voltage neurograms of muscle SNA.

Recordings were obtained after 4 weeks at an altitude of 5260 m (middle column), 3 days after descent (right column), and 4–6 months after return to sea level (left column). The neurograms are representative for the three subjects with the lowest (A), median (B) and highest (C) muscle SNA burst frequency at altitude.

Figure 2. Haemodynamic and sympathetic neural responses to high altitude acclimatization.

Summary data (means ± s.e.m.) for mean arterial pressure, heart rate and muscle SNA during supine rest after 4 weeks at an altitude of 5260 m, 3 days after return (descent) and 4–6 months after return from altitude (sea level). *P < 0.05vs. sea level; n = 8.

Table 2.

Effects of high altitude acclimatization on haemodynamics and sympathetic nerve activity

	Sea level (n = 8)	Altitude 5260m (n = 8)	Descent(n = 7)
Mean arterial pressure (mmHg)	77 ± 2	87 ± 3*	83 ± 3*
Heart rate (beats min−1)	56 ± 2	72 ± 4*	61 ± 3*
Muscle SNA (bursts min−1)	16 ± 2	48 ± 5*	37 ± 5*
Muscle SNA (bursts (100 beats)−1)	27 ± 5	64 ± 4*	56 ± 5*
Calf blood flow (ml min−1 (100g)−1)	2.34 ± 0.2	1.53 ± 0.2*†	2.65 ± 0.3
Calf vascular resistance (units)	35 ± 4	63 ± 8*†	35 ± 5

^*P < 0.05vs. sea level (n = 8 for comparison to altitude, n = 7 for comparison to descent)

^†P < 0.05vs. descent (n = 7).

Oxygen breathing and volume loading do not reverse the major part of high altitude sympathoexcitation

Paced breathing by itself had no effects on MAP, HR or muscle SNA (data not shown). At sea level, administration of 100 % oxygen or intravenous saline infusion alone, had no effects on any variable, but in combination produced a minor increase in MAP accompanied by a minor decrease in muscle SNA (Table 3, Fig. 3). At altitude, oxygen saturation as expected normalized during oxygen breathing, but quickly returned when oxygen administration was stopped (Table 3). Oxygen restored MAP to sea level values and returned HR towards sea level values. The addition of saline increased MAP, but had little effects on HR (Table 3). At altitude, the effects of oxygen breathing and saline infusion were reductions in muscle SNA of 6 ± 1 and 7 ± 2 bursts min⁻¹, respectively (both P < 0.05). These effects were partly additive, however during the combined intervention the resulting muscle SNA burst frequency was still greatly elevated (38 ± 4 bursts min⁻¹; Fig. 3, Table 3).

Table 3.

Responses to oxygen breathing and saline infusion alone and in combination at altitude and sea level

		Responses at altitude (n = 8)
	Control	O2	O2+ saline	Saline
O2 sat. (%)	82 ± 2	98 ± 0.3*	98 ± 0.2*	86 ± 1
Mean arterial pressure (mmHg)	87 ± 3	78 ± 2*	89 ± 2	84 ± 2
Heart rate (beats mm−1)	72 ± 4	63 ± 3*	61 ± 4*;	70 ± 4
Muscle SNA (bursts min−1)	48 ± 5	41 ± 4*	38 ± 4*	41 ± 6*
Muscle SNA (bursts (100 beats)−1)	64 ± 4	62 ± 4	56 ± 4	54 ± 4*
		Responses at sea level (n = 7)
	Control	O2	O2+ saline	Saline
O2 sat. (%)	98 ± 0.1	99 ± 0.3	100 ± 0.3*	98 ± 0.4
Mean arterial pressure (mmHg)	77 ± 2	79 ± 3	85 ± 3*	83 ± 3
Heart rate (beats min−1)	55 ± 3	54 ± 1	54 ± 4	56 ± 4
Muscle SNA (bursts min−1)	15 ± 2	13 ± 3	12 ± 3*	13 ± 3
Muscle SNA (bursts (100 beats)−1)	26 ± 5	23 ± 5	20 ± 6	21 ± 6

O₂ sat., oxygen saturation

^*P < 0.05vs. control.

Figure 3. Haemodynamic and sympathetic neural responses to oxygen and volume loading.

Summary data (means ± s.e.m.) for mean arterial pressure, heart rate and muscle SNA during supine rest after 4 weeks at an altitude of 5260 m (n = 8) and 4–6 months after return to sea level (n = 7) under 4 conditions (left to right): (1) breathing ambient air, (2) breathing 100 % oxygen, (3) volume loading by 1000 ml of saline i.v. over 15 min while breathing 100 % oxygen and (4) volume loading while breathing ambient air. *P < 0.05vs. ambient air.

High altitude sympathoexcitation persists after return to sea level

Three days after descent from altitude to sea level, blood gases and haematocrits were normalized (Table 1). In spite of this, the sympathoexcitation induced by prolonged hypobaric hypoxia was well maintained at this time point (37 ± 5 bursts min⁻¹, P < 0.05vs. 4-6 months post-expedition sea level control). Furthermore, HR and MAP also remained elevated, whereas calf blood flow and resistance were not significantly different from sea level control (Table 2, Fig. 2).

Discussion

This study provides direct neurophysiological data on human sympathetic neural behaviour during and after prolonged exposure to high altitude. The major new finding is that acclimatization to high altitude is accompanied by marked elevations in resting levels of sympathetic neural discharge to the skeletal muscle vascular bed. The study provides the first experimental evidence that chemoreflex activation and baroreflex deactivation surprisingly account for only a small part of this sympathoexcitation. Furthermore, the high altitude-induced sympathetic activation persisted for days after return to sea level, representing a unique example of long-lasting sympathetic neural overactivity in healthy humans.

Magnitude of sympathetic activation in chronic vs. acute hypoxia: progressive activation during acclimatization?

Acute exposure to hypoxia has previously been well documented to cause increased muscle SNA as measured by microneurography (Saito et al. 1988; Rowell et al. 1989; Somers et al. 1989, 1991; Leuenberger et al. 1991; Morgan et al. 1995; Duplain et al. 1999; Hansen et al. 2000; Xie et al. 2001). The most frequently used approach has been hypoxic breathing mimicking the ambient air at altitude (typically an F_I,O2 of 0.10 corresponding to ≈5500 m). The sympathetic response has been somewhat variable with increases typically in the order of 6–12 bursts min⁻¹ (Leuenberger et al. 1991; Morgan et al. 1995; Duplain et al. 1999; Hansen et al. 2000). In a hypobaric chamber study, Saito et al. (1988) found an increase of 13 and 14 bursts min⁻¹ with acute exposure to simulated altitudes of 5000 and 6000 m, respectively. In an acute high altitude microneurographic study comparing muscle SNA recordings performed at sea level and on Monte Rosa (4559 m) 18–24 h after ascent, muscle SNA was increased by 15 bursts min⁻¹ in healthy subjects at altitude (Duplain et al. 1999). All these acute responses to hypoxaemia contrast with the present finding of a massive 32 bursts min⁻¹ increase in muscle SNA after 4 weeks at 5260 m. The main obvious difference between this study and the previous studies is the prolonged time of exposure to hypoxic air, suggesting an apparently paradoxical further sympathetic neural activation as acclimatization to hypobaric hypoxia is progressing. All our subjects were in good health and performed exercise and daily activities that far exceeded their capacity during initial exposure to altitude. Furthermore, arterial blood gases and haemoglobin concentrations at altitude documented: (1) return of blood oxygen saturation towards sea level values and (2) higher blood oxygen content compared with sea level values. Despite this evidence for the expected functional and physiological acclimatization, muscle SNA was three times higher after 4 weeks at high altitude compared with the 4–6 months post-expedition sea level control. Although the subjects had sustained hypocapnia, which may increase the chemoreflex sensitivity (Somers et al. 1989), the paradox of the markedly elevated muscle SNA is further emphasized by the relatively high oxygen saturation of 85 %. This level of saturation is well above the 80 % suggested as the threshold for acute hypoxia-induced sympathoexcitation (Smith & Muenter, 2000).

One possible explanation for this apparent paradox is that blood gas values measured during wakefulness underestimate the hypoxic stimuli experienced at high altitude, since they do not reflect intermittent desaturation associated with episodic breathing during sleep, which may be involved in the sustained sympathetic activation (see below).

Sympathetic activation in chronic hypoxia: integration of neurophysiological and biochemical studies

The microneurographic technique directly measures sympathetic nerve activity with minor day-to-day variations (Sundlöf & Wallin, 1977), which was also demonstrated in a group studied twice at sea level in Copenhagen (see Methods). Furthermore, microneurography requires no assumptions about the relation between neural discharge, NE release and clearance. This is a key advantage in the present studies, because previous studies had indicated that hypoxaemia-induced increases in muscle SNA were not uniformly accompanied by increases in plasma NE (Rowell et al. 1989). The reason for this may be related to increased NE clearance from plasma (Leuenberger et al. 1991; Rostrup, 1998), such that plasma and urine NE concentration underestimates sympathetic activity during acute hypoxaemia.

Whether urine and plasma NE levels continue to underestimate sympathetic activity during sustained exposure to hypoxia is unknown. However, the data during the first days at altitude have been inconsistent, with NE levels being either decreased (Rostrup, 1998) unchanged (Cunningham et al. 1965; Reeves et al. 1992; Mazzeo et al. 1994; Kanstrup et al. 1999) or increased (Bärtsch et al. 1991). During prolonged exposure for a week or more at altitudes exceeding 4000 m, invariably, NE levels increase (Cunningham et al. 1965; Bärtsch et al. 1991; Reeves et al. 1992; Mazzeo et al. 1994; Antezana et al. 1994; Ponchia et al. 1994) and stay elevated for weeks (Mazzeo et al. 1994; Antezana et al. 1994), but may return towards sea level values after 3–4 weeks (Antezana et al. 1994; Ponchia et al. 1994). The increasing magnitude of NE levels at altitudes of 4300 m (Mazzeo et al. 1994), 5000 m (Ponchia et al. 1994), and 6500 m (Antezana et al. 1994) indicate an expected relation between altitude and degree of sympathetic activation. Two studies using unpaired comparisons of different subject groups at sea level and altitude report significantly higher NE plasma concentrations after a 10 weeks sojourn at 6000 m (Anand et al. 1993), and in high altitude residents at 3600 m (Antezana et al. 1995). While the latter two studies raise the possibility that normalization of sympathetic activity does not occur at high altitude, neurophysiological studies are needed to directly address this issue.

The vascular response to elevated sympathetic discharge in acute and chronic hypoxia

Despite sympathetic activation, acute hypoxaemia is well known to cause vasodilatation of regional vascular beds in the systemic circulation (Rowell et al. 1989), including those of skeletal muscle (Heistad & Wheeler, 1970; Gonzáles-Alonso et al. 2001) and to inhibit vascular responsiveness to sympathetic vasoconstrictor stimuli (Heistad & Wheeler, 1970; Rowell et al. 1989; Hansen et al. 2000). In one hypobaric chamber and two laboratory studies (Rowell et al. 1986, 1989; Saito et al. 1988) more severe hypoxia (F_I,O2 0.07–0.09) produced relatively large decreases in mean arterial blood pressures despite high sympathetic traffic (Saito et al. 1988; Rowell et al. 1989). In the present study sympathoexcitation at high altitude was accompanied by significantly increased calf vascular resistance and moderately elevated arterial pressure. Furthermore oxygen breathing did not increase, but rather lowered mean arterial pressure, suggesting that the level of sympathetic activity over-compensates for local vasodilatation. This makes the term sympathetic overactivity appropriate, and argues strongly against involvement of the arterial baroreflex in high altitude sympathoexcitation. In contrast, in the descent study, calf vascular resistance was similar to sea level control values despite much higher sympathetic traffic. Since hypoxia cannot be directly responsible for this uncoupling after descent, we cannot fully explain this observation based on the present data. However, it is consistent with previous studies demonstrating normal calf vascular resistance despite continued elevations in sympathetic traffic during recovery following short-term exposure to hypoxia (Morgan et al. 1995; Xie et al. 2001). Additionally, we speculate that the chronically elevated muscle sympathetic discharge in our study may be accompanied by compensatory downregulation of α-adrenoceptors or other constituents of the signal- transduction pathway linking NE release and vasoconstriction during, and after, high altitude acclimatization.

Contribution of chemoreflexes and cardiopulmonary baroreflexes to high altitude induced sympathoexcitation

The mechanisms inducing increased muscle SNA during acute exposure to hypoxia are related primarily to excitatory influences from oxygen sensitive chemoreceptors in the carotid body (Marshall, 1994; Prabhakar, 2000) and brainstem neuronal pools (Reis et al. 1994; Solomon, 2000). The mechanisms underlying the sympathetic overactivity during prolonged exposure to hypobaric hypoxia could be related to continued chemoreflex activation. In addition, unloading of cardiopulmonary baroreceptors could be involved due to the decrease in blood volume occurring over time at altitude (Robach et al. 2000). To begin to assess the relative importance of these mechanisms, we performed additional experiments designed to manipulate hypoxic chemoreceptor drive and cardiopulmonary baroreceptor drive alone, and in combination. Oxygen breathing and cardiopulmonary baroreceptor loading with saline had significant, albeit relatively minor effects on muscle SNA, indicating that the major part of the sympathoexcitatory response to prolonged high altitude exposure is not produced by these mechanisms. However, some interpretational issues should be mentioned. First, it is a limitation that ventilation was not quantified in the present study. Thus, despite using paced oxygen breathing in an effort to control ventilation, we cannot exclude that diminished hypoxic ventilatory drive caused a fall in ventilation and relative hypercapnia, leading to reduced inhibitory drive from lung stretch receptors and increased central chemoreceptor stimulation, respectively. Second, sensitization of oxygen sensitive chemoreflexes controlling sympathetic outflow may take place during chronic sustained hypoxic exposure. For example, such sensitization of carotid body sensitivity to hypoxia has recently been reported during chronic intermittent hypoxic exposure in rodent models (Peng et al. 2001). Thus, although breathing 100 % oxygen in our study raised saturation to ≈100 %, it may not have completely eliminated chemoreceptor drive. Third, arterial blood pressure decreased during oxygen breathing at altitude, which may have evoked baroreflex mediated increases in muscle SNA. Finally, we have no direct measure of cardiopulmonary baroreceptor loading at altitude or during saline infusion. While central venous pressure is often used during short interventions at sea level as an index of changes in the loading status of cardiopulmonary baroreceptors, it is not a reliable estimate at high altitude. One previous hypobaric chamber study has provided evidence that right atrial pressure decreases by 2.3 mmHg at 6100 m compared with sea level, whereas right ventricular and pulmonary pressures increase significantly at high altitude (Reeves et al. 1987). The relative importance of these diverging changes in cardiac and pulmonary pressures is unknown, and precludes straightforward assessment of cardiopulmonary baroreceptor activation at high altitude. The volume loading protocol in the present study was based on previous sea level microneurographic studies using similar volumes of intravenous saline to increase central venous pressure by 3–4 mmHg compared with control (Vissing et al. 1989; Jacobsen et al. 1993) and normalize the decreased central venous pressure during heat stress (Crandall et al. 1999). In addition, peripheral vein compliance is decreased by about 20 % at high altitude compared with sea level (Weil et al. 1971; Cruz et al. 1976), and the absolute decrease in blood volume induced by high altitude is between 250 and 750 ml (Sawka et al. 2000). If similar decreases in compliance and intravascular volume are assumed at high altitude in the present study, 1000 ml of normal saline would be expected to provide significant acute volume expansion. However, residual deactivation of the cardiopulmonary baroreflex even during saline infusion at high altitude cannot be excluded.

If operative in the present experiments, the factors mentioned above might lead to an underestimation of the role of the hypoxic chemoreflex and the cardiopulmonary baroreflex. Importantly, however, these experimental data acquired at altitude should be interpreted together with the data acquired 3 days after return to sea level. Under conditions of normobaria, normoxia, slightly elevated arterial blood pressure, and hypervolaemia (as indicated by completely normalized haematocrit values), we found muscle SNA values very similar to those observed during oxygen breathing and volume loading at altitude. This strongly indicates that potential confounding factors did not account for the major part of the large sympathoexcitatory response to prolonged hypobaric hypoxia.

Additional potential mechanisms underlying sympathetic neural activation in chronic hypobaric hypoxia

The mechanisms by which acclimatization to hypobaric hypoxia produces persistent sympathetic neural activation have not been fully elucidated. In a previous study measuring NE levels during 21 days at 4300 m, Asano et al. (1997) found a correlation between NE excretion in urine and tidal volume, leading the authors to suggest a coupling between ventilatory and sympathetic acclimatization. It is unlikely that an increased ventilatory response per se could be causing an increase in muscle SNA of the magnitude observed in our study. Increased tidal volume mainly influences the distribution of sympathetic activation during the respiratory cycle, and if anything, causes relative inhibition of sympathetic outflow due to activation of inhibitory vagal fibres subserving pulmonary stretch receptors (Somers et al. 1989; Seals et al. 1993). Another interpretation is that sympathetic activation and ventilatory acclimatization share common regulatory mechanisms such as chemoreceptor sensitization (Prabhakar et al. 2001) or a coupling between central respiratory motor drive and central sympathetic outflow (Zhong et al. 1997). The continued ventilatory drive in previous studies (Dempsey & Forster, 1982; Pedersen et al. 2000) and elevation of muscle SNA along with blood gas evidence of continued hyperventilation in the present study for days after return from altitude supports this possibility. However, even brief (20 min) exposure to hypoxia has recently been reported to produce activation of muscle sympathetic discharge that is sustained after discontinuation of the stimulus, whereas similar increases in sympathetic discharge produced by exposure to hypercapnia promptly returns to baseline after withdrawal of the stimulus (Morgan et al. 1995; Xie et al. 2001). In those studies, ventilation returned to baseline promptly after withdrawal of both the hypoxic and hypercapnic stimulus. The authors speculate that hypoxic exposure may engage long-term potentiation of synaptic transmission in brain centres involved in the control of central sympathetic outflow (Xie et al. 2001). This notion is supported by several lines of experimental studies in rodents showing sympathetically mediated increases in blood pressure (Fletcher, 2001), augmentation of carotid body sensitivity to hypoxia (Prabhakar et al. 2001) and increased expression of c-fos, a marker of transcriptional activity, in the rostral ventrolateral medulla following exposure to chronic intermittent hypoxia (Hirooka et al. 1997; Greenberg et al. 1999). Clearly, further studies are needed to determine the contribution of potential long-term regulators of central sympathetic outflow, such as sensitization/resetting of reflex control mechanisms, neurohormonal changes, altered levels of circulating neuromodulatory substances and altered gene expression, to the sympathetic activation accompanying chronic sustained or intermittent exposure to hypoxia in humans.

Pathophysiological implications

The physiological significance of sustained sympathetic overactivity at high altitude is not clear. It could be an integral part of the acclimatization process, or it could be a detrimental side effect increasing the risk for high altitude related health problems such as increased blood pressure (Reeves et al. 1992; Kanstrup et al. 1999), and pulmonary oedema (Bärtsch et al. 1991; Duplain et al., 1999). The present study introduces a novel model for long-term sympathetic overactivity in healthy humans, and thereby provides the opportunity to further study mechanisms involved in the long-term control of sympathetic activity. These mechanisms may be relevant for our understanding of a variety of disease states characterized by chronic sympathoexcitation and chronic intermittent or sustained hypoxaemia, such as sleep apnoea related hypertension (Carlson et al. 1993; Narkiewicz et al. 1999b), chronic obstructive pulmonary disease (Heindl et al. 2001) and congestive heart failure (Leimbach et al. 1986; Narkiewicz et al. 1999a).