INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

WHEN A NORMAL SUBJECT exercises while breathing air at sea level (SL), heart rate (HR) and cardiac output (T) rise linearly in proportion to work rate (38) until maximal exercise and maximal O₂ consumption (O_2 max) are reached. When the same subject ascends rapidly to altitude, O_2 max is reduced but maximal HR and T are both similar to results at SL (46) or are slightly reduced (8, 26, 42). When the same subject remains at this altitude for 2 wk or more, acclimatization occurs with increases in ventilation, hemoglobin (Hb) concentration, and renal excretion of bicarbonate and water. However, O_2 max is not altered over 2 wk at such altitudes, whereas maximal T clearly decreases (1, 32, 43), implying an increase in arterial O₂ concentration, an increase in peripheral O₂ extraction, or both. This independence of O_2 max from T does not prove that T per se has no effect on maximal exercise capacity at altitude. In fact, because a high T is generally considered to be a key factor enabling high exercise levels (6), the reduction in maximal T on acclimatization has the potential for limiting O_2 max at altitude. Thus O_2 max might have been higher after acclimatization had peak T not fallen.

Several hypotheses have been put forward to explain the acclimatization-induced reduction in maximal T (reviewed in Ref. 45). First, water shifting out of the vascular space, together with fluid loss through greater sweating, respiration, and urine production with altitude, leads to reduced plasma volume, which, when not compensated for by an increase in erythrocyte volume, leads to a reduced blood volume. This in turn could lead to lower cardiac filling pressures (preload), compromising maximal T and O_2 max (37). Second, increased Hb concentration due to enhanced renal erythropoietin release and plasma volume loss elevates blood viscosity, which could reduce maximal T. Third, myocardial contractility could be directly reduced by the hypoxia of altitude. Fourth, maximal T might be reduced by adaptations in the autonomic nervous system (ANS), affecting HR in the absence of compensatory changes in stroke volume. A fifth more or less "passive" hypothesis states that none of the above pathophysiological mechanisms is responsible and that the reduced maximal T is simply the result of a low maximal work rate and O_2 max. Accordingly, the skeletal muscle at altitude has a reduced ability to work because of the reduced PO₂, and, because T is closely correlated with O₂ uptake (O₂), maximal T is reduced.

The present study was undertaken to investigate the fourth hypothesis, i.e., the role of the ANS. As has been shown previously, acclimatized humans and rats both show evidence of cardiac -receptor desensitization (21, 33, 34) and increased muscarinic receptor activity (22), which could provide the basis for a role of the ANS in this phenomenon. With the use of pharmacological interventions to block either arm (sympathetic and parasympathetic) of the ANS, clear effects have been shown on maximal HR (reduction with propranolol and increase with atropine), whereas peak exercise capacity and O₂ were unchanged (14, 29). As of yet, no studies with these interventions have been performed in which maximal T at SL and after acclimatization were compared. We hypothesized that, despite evidence of altered ANS function after altitude acclimatization with attendant effects on HR, these alterations are not the cause of the reduced maximal exercise T. We used sympathetic and parasympathetic blockade (separately) before and after altitude acclimatization to determine whether maximal T was sensitive to the ANS state. In this paper, we report that, after altitude acclimatization, such blockade predictably altered maximal HR but had no significant effects on maximal T, O₂, or work rate.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. The Human Subjects Committee of the University of California (San Diego, CA) approved this study. Six (5 men, 1 woman) healthy nonsmoking subjects were included in the study. They were all physically active, but not competing athletes, and had no prior history of respiratory disease, cardiac disease, or high-altitude pulmonary or cerebral edema. After subjects gave written, informed consent, a history was obtained and a physical examination was performed to exclude cardiopulmonary abnormalities. All subjects were then screened for pulmonary disease by standard spirometry. Subject characteristics are summarized in Table 1.

Study design. After a standard progressive maximal cycle exercise test at SL to determine peak O₂ and workload, noninvasive measurements were made of T and HR at five different levels of exercise: rest, at 30, 60, and 90% of SL, and at 100% of SL maximum. All measurements were made over the course of 12 such exercise bouts spaced out over 6 experimental days, at 2 bouts per day. On the first 3 days, measurements were made at SL in San Diego. After a 2-wk acclimatization period (WM) at the White Mountain Research Station (Barcroft Station, Pace Laboratory), near Bishop, California, where the altitude is 12,470 feet (3,800 m) and barometric pressure is ~482-486 mmHg, another 3 days of measurements were carried out. On each experimental day, subjects received 1) no drug (control), 2) sympathetic blockade with propranolol, or 3) parasympathetic blockade with glycopyrrolate. On each day, two exercise tests were performed, one with subjects breathing ambient air [inspiratory fraction of O₂ (FI_O2) = 0.209] and one with subjects breathing a gas mixture with an FI_O2 of 0.125 at SL or 0.34 at WM. In this way, it was ensured that there were comparable high and low inspired PO₂ (PI_O2) (further referred to as normoxia and hypoxia, respectively) at both locations. The order of sessions was chosen at random. On a day with sympathetic blockade, 6 mg of propranolol were administered intravenously before the first exercise bout. After this bout and a 1-h resting period, another 2 mg of the drug were administered before the second bout. On a day with parasympathetic blockade, 0.8 mg of glycopyrrolate was administered intravenously before the first bout and another 0.2 mg before the second. All drug administrations and exercise tests were performed under close and continuous electrocardiogram (ECG) and blood pressure monitoring. At both SL and WM, at least 24 h were allowed between any 2 experimental days to ensure washout of propranolol and glycopyrrolate. On the control (no drug) day at SL only, a 20-gauge radial artery cannula was inserted to allow comparison of arterial O₂ saturation (Sa_O2) by cooximetry and by noninvasive pulse oximetry. These comparisons are not germane to the present study and are reported elsewhere (48).

T measurement by short-term acetylene uptake. T was measured in all subjects by a nonrebreathing acetylene (C₂H₂) technique, based on the principles of mass balance. The method, described previously in detail, has been validated in healthy subjects up to maximal exercise (3). In short, it relies on the fact that the rate of alveolar absorption of a gas soluble in blood, such as C₂H₂, is proportional to pulmonary blood flow. A subject breathes from a bag containing 1% C₂H₂ and 5% helium for 20 breaths. Bag PI_O2 is similar to that inspired during the particular exercise run. This is a nonrebreathing method that therefore does not change PO₂ or PCO₂ during its application. T is calculated with a computer algorithm by using minute ventilation, inspired C₂H₂ concentration, helium-corrected end-tidal C₂H₂ concentration, end-tidal PCO₂, mixed expiratory PCO₂, and the blood-gas partition coefficient of C₂H₂ measured at rest and during exercise on each experimental day, as reported by Barker et al. (3).

Exercise and O₂. All exercise tests were performed on an electronically braked cycle ergometer (Excaliber, Quinton Instruments, Groningen, The Netherlands), with HR monitored by a cardiac monitor (Lifepak 6, Physio-control, Redmond, WA). Subjects breathed through a nonrebreathing valve (model 2700, Hans Rudolph, Kansas City, MO), with a dead space of 100 ml. Mixed expired gas was sampled continuously from a heated mixing chamber, and concentrations of O₂ and CO₂ were measured by mass spectrometry (model 1100, Perkin-Elmer, Pomona, CA). Expired gas flow was measured with a pneumotach (no. 3 Fleisch) and a differential pressure transducer (model DP45-14, Validyne, Northridge, CA). O₂ and CO₂ production were determined from these measurements. The coefficient of variation in O₂ and CO₂ production over several days in a single subject is 5% in our laboratory.

Data analysis. Calculations were made of stroke volume (stroke volume = T/HR). Because no direct measurements were made of Hb and arterial PO₂, O₂ delivery was estimated as T × 1.39 × (Hct/3) × Sa_O2. Dissolved O₂ was neglected. Distributions of all variables were evaluated by Levene's test of equality of variance; parametric tests were used only in case of normal distributions. Differences between the 12 conditions, two levels for acclimatization (SL or WM) times three levels for drug (control, propranolol, or glycopyrrolate) times two levels for PI_O2 (high or low), were assessed by repeated-measure ANOVA (parametric) or Friedman's test (nonparametric). Post hoc testing was done only in case of a significant condition effect in the ANOVA with paired Student's t-tests (parametric) or Wilcoxon's signed-rank test (nonparametric). Bonferroni protection of the significance level was used for multiple comparisons. Pearson's (parametric) or Spearman's (nonparametric) correlation coefficients were calculated for the assessment of relationships between variables. Where the relationships examined used repeated measures data, the effects of between-subject differences were controlled for with the use of dummy encoding (9). All calculations were made with the SPSS-advanced statistics package, with P  0.05 taken as reflecting significance. Data are reported as means ±SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All six subjects completed the SL studies. For the WM studies, one subject developed a brief episode of chest pain at the end of the first exercise test, which was performed under propranolol treatment. Although there were no other cardiopulmonary symptoms or signs (normal ECG, normal auscultation), it was considered unsafe to proceed testing in this subject. He was subsequently sent down to the closest hospital (Bishop, CA), where further testing provided no evidence of myocardial ischemia or other pathological phenomenon. He remains normal and has resumed recreational climbing activities as before, with no recurrence of symptoms. Because this subject did not complete the entire test sequence, the data that are presented here are from the other five subjects only. With acclimatization, resting Hct rose from 43 ± 1 to 50 ± 1%. As expected, peak workload, peak O₂, and maximal exercise Sa_O2 were lower in ambient air than in high PI_O2 (Table 2). Acclimatization resulted in an increase in maximal exercise Sa_O2 but did not affect peak workload or peak O₂ (Table 2). Moreover, there were no significant drug effects on peak Sa_O2, O₂, or workload. At WM, resting and exercise venous norepinephrine concentrations were significantly higher (peak exercise SL vs. WM, both ambient air: 2,813 ± 359 vs. 3,506 ± 668 pg/ml, P < 0.001).

Peak HR. At WM, peak HR was reduced by 9% from 186 ± 3 at SL to 170 ± 5 beats/min in ambient air (chronic hypoxia, P < 0.0001) (Fig. 1). At WM, peak HR in normoxia was similar to SL (183 ± 4 vs. 186 ± 3 beats/min, respectively, P = not significant). Propranolol caused a further reduction in WM peak HR both in hypoxia and in normoxia (145 ± 1 and 139 ± 2 beats/min, respectively, P < 0.0001 for both). With glycopyrrolate, WM peak HR was brought back to SL levels (184 ± 3 beats/min, P < 0.001 compared with WM, ambient air, control).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Mean ± SE peak heart rate (HR, top), peak cardiac output (T, middle), and stroke volume (SV, bottom) in 5 subjects exercising at sea level (SL) and after 2 wk of acclimatization at 3,800 m (WM). Exercise was performed in normoxia [solid bars; ambient air at SL and inspiratory fraction of O2 (FIO2) = 0.34 after acclimatization] and hypoxia (open bars; FIO2 = 0.125 at SL and ambient air after acclimatization), either without drug (control, C), under sympathetic blockade with propranolol (PRO), or under parasympathetic blockade with glycopyrrolate (GLY). aWM-C hypoxia significantly different from SL-C normoxia (P < 0.0001 for HR and T; P < 0.01 for SV) and WM-C hypoxia significantly different from SL-C hypoxia (P < 0.0005 for HR; P < 0.01 for T; not significant for SV). bWM-PRO normoxia significantly different from WM-C normoxia (P < 0.0001 for HR; P < 0.005 for SV). cWM-PRO hypoxia significantly different from WM-C hypoxia (P < 0.0001 for HR; P < 0.002 for SV). dWM-GLY hypoxia significantly different from WM-C hypoxia (P < 0.001 for HR).

Peak T and oxygen delivery. At WM, peak T for subjects breathing ambient air was reduced by 18% compared with those breathing air at SL (16.2 ± 1.2 vs. 20.0 ± 1.8 l/min, P < 0.0001) and was also lower when comparing acute vs. chronic hypoxia (P < 0.01) (Fig. 1). Without drug administration, peak T WM was not significantly different from SL values with normoxia (18.4 ± 2.1 l/min, difference with SL not significant) (Fig. 1). Exactly the same pattern was seen with either propranolol or glycopyrrolate administration. In terms of oxygen delivery, the reduction in peak T with acclimatization was compensated for by an increase in Hct and Sa_O2 (Fig. 2). Although hypoxia greatly reduced oxygen delivery under all conditions (P < 0.0001), there were no significant differences between SL and WM oxygen delivery index levels (comparing equal FI_O2 conditions).

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Mean ± SE peak oxygen delivery (in arbitrary units) in 5 subjects exercising at SL and after 2 wk of acclimatization at 3,800 m (WM). Exercise was performed in normoxia (solid bars; ambient air at SL and FIO2 = 0.34 after acclimatization) and hypoxia (open bars; FIO2 = 0.125 at SL and ambient air after acclimatization), either without drug (C), under sympathetic blockade with propranolol, or under parasympathetic blockade with glycopyrrolate. aP < 0.0001 vs. normoxic condition at similar location and with similar drug treatment.

Peak stroke volume. Similar to HR, WM peak stroke volume was reduced by 10% (ambient air at both locations: 95 ± 4.9 vs. 108 ± 9.4 ml, P < 0.01; acute vs. chronic hypoxia not significant) and was similar to SL values with low PI_O2 (Fig. 1). The reduction in WM peak HR induced by propranolol was compensated for by an increase in WM peak stroke volume (propranolol vs. control: P < 0.002 in hypoxia and <0.005 in normoxia). Despite a numerically lower stroke volume after glycopyrrolate (higher HR without change in T), the reduction in stroke volume did not reach statistical significance (P = 0.036, not significant with Bonferroni correction) because of the combined random errors in the independently measured HR and T.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main findings of the present study are that, at WM, 1) T at peak exercise for subjects breathing air at this altitude was 18% lower than at SL before ascent; 2) maximal T was immediately restored to levels not significantly different from SL values when exercise was performed in normoxia; 3) propranolol reduced and glycopyrrolate increased maximal HR, without significant effects on maximal T. Also, although O_2 max and power outputs were both reduced by hypoxia, neither propranolol nor glycopyrrolate reduced these variables at SL or WM.

Thus we confirmed the basic observations made previously by others in points 1 and 2 above (1, 32, 43, 46). In addition, the effects of sympathetic and parasympathetic blockade on HR and exercise capacity were similar as in previous studies (14, 29). Overall, these results confirm the adequacy of our dosing of propranolol and glycopyrrolate and therefore establish the setting required to test the hypothesis of this paper: that the ANS reduction of HR is not responsible for the reduced maximal T at altitude. It is, however, appropriate to point out that, based on just five subjects, the conclusions of this study need to be made with some caution despite their statistical significance.

Attainment of peak O₂ under all conditions. Before dealing with any explanation for the reduced maximal T at altitude, the issue of whether our subjects reached peak O₂ must be discussed. There were no significant differences in peak O₂ or workload in any of the different exercise bouts (at SL or WM), including those of the preliminary test, under similar PI_O2. Evidence for maximum effort in the preliminary test comes from the fact that peak HR was higher than 95% of the age predicted maximum in all subjects. Additionally, the peak exercise respiratory exchange ratios were 1.09 ± 0.01 and 1.11 ± 0.01 in normoxia and hypoxia, respectively. Lysed whole blood lactate concentrations of 9.7 ± 0.8 and 8.2 ± 0.7 mmol/l and norepinephrine concentrations of 2,813 ± 359 and 2,727 ± 384 pg/ml were observed, respectively. Together, these findings support the conclusion that subjects attained peak O₂ under all experimental conditions.

C₂H₂ uptake method for measuring T. Because of the large number of individual measurements of peak T performed in each subject (about 60 determinations in each subject, spread over 12 exercise sessions on 6 different days), it was critical that a noninvasive method was used. Because of the motion associated with exercise, the only class of noninvasive techniques that could be used had to be based on gas exchange. A further critical constraint was that, at all exercise levels in both normoxia and hypoxia, alveolar PO₂ and PCO₂ during the periods of T measurement had to be maintained at the levels occurring naturally at the designated PI_O2 during each exercise bout before and after such measurements. This was important because acute changes in PO₂ can change T and HR within seconds, thus invalidating the entire study. The short-term C₂H₂ uptake method (3) was developed for just this circumstance. It requires only that the subject be switched from the particular inspired O₂-N₂ gas mix dictated by the protocol to a gas mix of the same PI_O2 containing trace levels of C₂H₂ and He for ~20 breaths. Moreover, during these 20 breaths, the subject continues to breathe exactly as during the remainder of the exercise segment. Rebreathing methods on the other hand (CO₂ or C₂H₂) have the potential of causing substantial changes in alveolar PO₂ and/or PCO₂, unless special precautions are taken to buffer inhaled levels of these gases, a difficult task. Thus the C₂H₂ uptake method was selected as the most appropriate technique for our particular protocol design. This required measurements of C₂H₂ partition coefficient in each subject on every study day, both at rest and during exercise, to be sure that apparent differences in T were not the result of changes in C₂H₂ solubility under any condition. Our data therefore reflect the use of each subject's own solubility data for each exercise session. In the paper by Barker et al. (3), the C₂H₂ method was validated against the direct Fick method and found to agree well. Johnson et al. (20) compared C₂H₂ uptake data (by using two computational methods different from ours) with the Fick method and found strong correlations (r² = 0.89 to 0.90). However, T was underestimated significantly with the use of one of these two calculation techniques, likely a result of either increased ventilation-perfusion (A/) inequality or failure to account for C₂H₂ recirculation to the lungs at high levels of exercise. Therefore, an important constraint on the use of such methods is their vulnerability to underestimation of T in the presence of A/ inequality. Our laboratory noted previously that exercise, especially in hypoxia, causes an increase in A/ inequality (12). Thus the hypoxic T values during peak exercise could be somewhat lower than the actual values. This is unlikely to affect interpretation of the present study for two reasons. First, the important comparison was for maximal T for subjects breathing ambient air at 3,800 m between control conditions, after propranolol, and after glycopyrrolate. The peak work rates and O₂ were unaffected by the two drugs (see RESULTS), and the study design used each subject as their own control. Thus, even if A/ inequality led to an underestimate of T, the relative values under the three conditions should be comparable. The second reason comes from a study of the quantitative effects of A/ inequality on C₂H₂-based T measurements, for which we constructed a two-compartment tidally ventilated model of A/ mismatch and computed the effects of increasing degrees of A/ inequality on the estimated T. The increase in A/ inequality from rest to exercise, even in hypoxia, is relatively small: log SD, the second moment of the perfusion distribution, increasing from ~0.4 to ~0.6 (12). This change was calculated to spuriously decrease estimated T by only 3%, a perturbation that would be very difficult to detect experimentally.

Possible mechanisms of reduced maximal T after altitude acclimatization. Several cardiocirculatory adaptations have been proposed to explain the reduced maximal T. It is known that hypoxia acutely increases Hb concentration, implying water shifting out of the plasma space (31). Subsequent altitude acclimatization results in further plasma volume reduction through greater sweating, respiration, and urine production. This plasma volume loss is enhanced by a reduction in plasma oncotic pressure due to net protein loss. In our study, Hct was elevated by 16% from 43 to 50%. Because it takes more than 2 wk before hypoxia-induced increases in erythropoiesis can significantly increase erythrocyte volume (11), plasma volume in our subjects must have been lower at WM than at SL (but we did not measure plasma volume). Accordingly, reduced cardiac filling pressures might have limited maximal T in our subjects. Although acclimatization-related changes in cardiac filling pressures were clearly shown by Reeves et al. (33) and Boussuges et al. (5), there is evidence against the hypothesis that they lead to a reduced maximal T. In those studies where blood volume was expanded, there was little or no augmentation in maximal T (10, 14, 47). Moreover, acute O₂ breathing at altitude causes maximal exercise capacity and T to rise (32) without changing blood volume. Such was the case in the present study, where breathing 34% O₂ at altitude elevated T and O₂ to values that were not significantly different from SL normoxic values. However, the possibility of a type II error cannot be ruled out. Nevertheless, Robach et al. (37) recently reported a 9% increase in O_2 max with plasma volume expansion in subjects exercising at a simulated altitude of 6,000 m.

ANS changes with altitude. It has been shown that with time at altitude, adaptations occur in the ANS and that they result in an altered HR response to exercise. These adaptations consist of increased sympathetic activation as evidenced by higher circulating norepinephrine and muscle sympathetic nerve activity (25, 27, 39). However, there is also evidence of cardiac -receptor desensitization (2, 21, 34, 35) and increased cholinergic activation (22). Prolonged exposure to altitude and, hence, prolonged exposure to high levels of adrenergic agonists can be expected to result in a decline in mRNA encoding the ₂-adrenergic receptor. This process of long-term desensitization should be considered separately from short-term desensitization and leads to lower quantities of receptors at the membrane level (13). The effects of acclimatization and autonomic blockade on the HR response to exercise that we found (lower maximal HR under control conditions, further reduction with sympathetic blockade, increase to SL maximum with parasympathetic blockade) are similar to those found in previous studies (15, 40). However, up until now, it has not been determined whether the reduced maximal HR is responsible for reduced maximal T or, rather, is compensated for by an increase in stroke volume. In previous studies, the effects of autonomic blockade on exercise T have been investigated at SL (ambient air) only, showing a slight reduction in maximal T with propranolol and no effect of parasympathetic blockade with atropine or glycopyrrolate (7). The present study confirms these results and, moreover, shows that these drug effects are similar after altitude acclimatization.

The passive hypothesis. With little or no evidence to support the aforementioned explanations for reduced maximal T (reduced blood volume, increased viscosity, myocardial hypoxia, or autonomic changes), we are left with the passive hypothesis as the best explanation for the findings. This hypothesis states that at any level of exercise, T is dictated by O₂. Certainly, there is a tight linear relationship between T and O₂ from rest to O_2 max. The slope and intercept of this relationship are very similar across many published studies in normoxia (16, 19, 23, 28, 41). Whereas the intercept is higher in acute hypoxia, the slope is not (46), and, after acclimatization, the relationship returns to that of normoxia (4, 32, 33). Our understanding of the underlying mechanisms that may tie T to O₂ is vague, but it is possible that predominantly local muscle metabolic and neural changes with exercise could dictate local vascular conductance. This in turn could signal an increase in T to maintain systemic pressure via further neural and humoral pathways (38). In this scenario, the lower maximal T would be the regulated result of the lower maximal power output and O₂ at altitude.

Consequences of changes in maximal T for O₂ transport in hypoxia. Theoretical and experimental work suggest that, at altitude, gains in convective O₂ transport in the circulation that might be afforded by an increase in T would be partially offset by corresponding reductions in the diffusive transport of O₂ from alveolar gas to pulmonary capillary blood and from microvascular blood to muscle mitochondria (44, 45). Somewhat by chance, the present study provides experimental data to partially test this hypothesis. Figure 3A shows the relationship between cardiac index (T divided by body surface area) and Sa_O2 at maximal exercise across the six conditions involving hypoxia (three at SL and three at WM, each under control conditions, postpropranolol, and postglycopyrrolate). Each of the five subjects is shown separately (each represented by a different symbol). There is a clear inverse relationship between the two variables compatible with the idea that, due to diffusion limitation in the lungs during hypoxic exercise, increases in T further impair O₂ loading in the lungs because of decreased pulmonary capillary transit times. The inverse relationship between cardiac index and Sa_O2 not only holds for all data grouped together but, more importantly, within the data sets of each single subject. The mean slope of this relationship is 3.7 and is significantly different from 0 (P < 0.0001; P = 0.0004 for regression lines in individual subjects). This slope translates to a mean fall of 1.5% in Sa_O2 per liter per minute increase in T. Whether this is cause and effect, Figure 3B shows that, within each individual subject, systemic oxygen delivery (indexed to body surface area) does not change with increasing cardiac index across the same six hypoxic conditions (the individual slopes are not different from 0). Thus the gain in delivery that would result from increased T is offset by a lower Sa_O2. The correlation between cardiac index and oxygen delivery for the grouped data is misleading. This positive correlation is due to between-subject differences in exercise capacity (the likely result of differences in fitness, genetics, and so forth) with subsequent differences in both cardiac index and oxygen delivery, and the relationship accounting for repeated measures on individuals is not significant (P = 0.52). Accordingly, O₂ extraction (Fig. 3C, calculated by dividing O₂ by oxygen delivery) does not change over the cardiac index range: the slope of the overall regression line is not significantly different from 0 (P = 0.12). This plot shows a lot of scatter, probably due to the fact that the variable is not directly measured. Instead, it was derived indirectly from T, Hct, Sa_O2, and O₂, thereby amplifying the separate measurement errors. Finally, Fig. 3D shows the relationship between peak cardiac index and peak O₂. Again, the individual regression lines give a picture different from the overall regression line. Overall, there is a highly significant positive correlation with a slope of 0.164, which is to be expected for exercise at moderate altitude. Knowing the average Sa_O2 under these circumstances (~80%), it can be calculated that, with every liter per minute per meter squared increase in cardiac index, oxygen delivery per meter squared is 160 ml higher, and, therefore, O₂ can theoretically increase by ~160 ml · min¹ · m². The individual lines, however, have much lower slopes. The average slope is 0.05, which is significantly different from both 0 (P = 0.03) and 0.164 (P = 0.002). This means that, within individuals, the sensitivity of peak O₂ to an increase in T is lower, at about one-third of the value expected if T were the dominant limiting factor. This is concordant with the concept that, at or near maximal exercise at moderate altitude, the benefits of an increase in convective O₂ transport will largely be offset by a decrease in diffusive O₂ transport (44).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Correlations between peak exercise cardiac index (T indexed for body surface area, expressed in ml · min1 · m2) and peak arterial O2 saturation (SaO2; in %; A), peak oxygen delivery index (oxygen delivery indexed for body surface area, expressed in arbitrary units; B), peak oxygen extraction (in %; C), and peak O2 consumption (O2; in ml · min1 · m2; D) in 5 subjects exercising in hypoxia: FIO2 = 0.125 at SL and ambient air after 2 wk of acclimatization at 3,800 m. altitude (WM). Exercise was performed either without drug, under sympathetic blockade with propranolol, or under parasympathetic blockade with glycopyrrolate. When between-subject variability is taken into account for this repeated- measures regression, only A (P = 0.0004) and D (P = 0.03) are significant. See text for details.

Conclusions. Of the several theories advanced to explain why maximal exercise T is reduced after acclimatization at altitude, that implicating the ANS as the major factor is not supported by the results of the present study. This is despite the documented changes in the ANS, including cardiac -receptor desensitization shown by the previous work of other authors (21, 33, 34). Thus separate sympathetic and parasympathetic blockade, although producing the expected changes in HR, had no significant effects on maximal exercise capacity or T at altitude. Combined with the immediate normalization of maximal T at altitude when high O₂ concentrations are breathed seen in the present and many prior studies, the likeliest cause of diminished peak T is decreased demand caused by hypoxic limitation of muscle metabolic rate.