Exercise-Induced Hypoxaemia Developed at Sea-Level Influences Responses to Exercise at Moderate Altitude

Purpose The aim of this study was to investigate the impact of exercise-induced hypoxaemia (EIH) developed at sea-level on exercise responses at moderate acute altitude. Methods Twenty three subjects divided in three groups of individuals: highly trained with EIH (n = 7); highly trained without EIH (n = 8) and untrained participants (n = 8) performed two maximal incremental tests at sea-level and at 2,150 m. Haemoglobin O2 saturation (SpO2), heart rate, oxygen uptake (VO2) and several ventilatory parameters were measured continuously during the tests. Results EIH athletes had a drop in SpO2 from 99 ± 0.8% to 91 ± 1.2% from rest to maximal exercise at sea-level, while the other groups did not exhibit a similar decrease. EIH athletes had a greater decrease in VO2max at altitude compared to non-EIH and untrained groups (-22 ± 7.9%, -16 ± 5.3% and -13 ± 9.4%, respectively). At altitude, non-EIH athletes had a similar drop in SpO2 as EIH athletes (13 ± 0.8%) but greater than untrained participants (6 ± 1.0%). EIH athletes showed greater decrease in maximal heart rate than non-EIH athletes at altitude (8 ± 3.3 bpm and 5 ± 2.9 bpm, respectively). Conclusion EIH athletes demonstrated specific cardiorespiratory response to exercise at moderate altitude compared to non-EIH athletes with a higher decrease in VO2max certainly due to the lower ventilator and HRmax responses. Thus EIH phenomenon developed at sea-level negatively impact performance and cardiorespiratory responses at acute moderate altitude despite no potentiated O2 desaturation.


Introduction
In endurance sport, most training camps and competitions take place at altitude around 2000 m. At this altitude, aerobic performance and maximal oxygen uptake (VO 2max ) are reduced due to the decrease in partial pressure of inspired O 2 (PIO 2 ) [1]. However, altitude does not appear to affect every individual in an equal manner. Indeed it is now well documented that endurancetrained athletes demonstrate a larger decline in VO 2max with increasing altitude compared to untrained subjects [2][3][4]. This particular sensitivity of endurance-trained athletes to moderate altitude has been related to an important decrease in arterial O 2 saturation (SaO 2 ) [2,3,[5][6][7]. At sea level, 50% of healthy endurance-trained athletes develop exercise-induced hypoxaemia (EIH) [8]. "EIH athletes" show a difference between rest and maximal arterial O 2 pressure (PaO 2 ) values for at least 10 mmHg and/or a delta of SaO 2 of at least 4% [9]. It is reported that EIH subsequently affects performance in response to high intensity exercise [10]. Physiopathology of EIH seems to be 1) a relative hypoventilation related to training adaptation and 2) gas exchange abnormality [11]. Each of these mechanisms appear to participate in EIH occurrence, with specific contribution depending on the training status and/or exercise mode and/or muscle mass involved [12]. However on average~20% of the variation in PaO 2 between individuals during exercise is due to the variation in hyperventilation response, with the remaining~80% of the variance in PaO 2 drop roughly divided evenly between ventilation-perfusion mismatch and diffusion limitation [13]. Because PIO 2 is reduced with altitude, EIH exhibited at sea level by some endurance-trained athletes could be amplified at acute moderate altitude. Only a few studies have examined the consequences of EIH during a maximal exercise performed at altitude while the number of athletes involved in endurance mountaineering sports, who are prone to develop EIH, was greatly increased in the last decade. In fact, literature reports only four studies that have compared O 2 desaturation and aerobic performance during maximal exercise at altitude between EIH athletes identified at sea level and non-EIH athletes [4,6,14,15]. If these studies have reported a greater decrease in VO 2max in EIH athletes than in non-EIH athlete, they also have reported controversial results about SpO 2 decrement and cardio-respiratory responses. Moreover none of these studies has examined the effect of a moderate altitude despite it is an usual altitude for training camps and competitions in endurance sports or mountaineering. The general aim of this study was to better understand the constraints applied on the cardiorespiratory system of EIH athletes at moderate altitude. More specifically, the purpose of this study was to evaluate 1) the level of O 2 desaturation and 2) cardio-ventilatory responses of EIH athletes identified at sea level during a maximal exercise performed at acute moderate altitude.

Materials and Methods Participants
Twenty-three healthy non-smoker males were recruited for this study. All the participants were sea-level residents. Fifteen participants were endurance-trained athletes involved in running or cycling activities (at least 10 hours per week for the past ten years). The remaining eight were sedentary or active in recreational sports, but had never been engaged in systematic endurance training. Before their inclusion in the protocol, each participant was informed about the procedures and the potential risks inherent to the experiments. They signed a written informed consent form. All procedures were approved by the ethics committee of the Consell General de l'Esport, Catalunya. performed at sea level (SL) and the second at 2150 meters in an altitude hut (ALT). After a 6-minutes warm-up at 30 Watts (W) for untrained participants and 60 W for trained participants, the power was increased every minute by 30 W until exhaustion. The test was considered to be maximal if at least three of the following criteria were presented: 1) an increase of VO 2 of < 100 ml with the last increase in work rate, 2) achievement of age-predicted maximal heart rate (HR) [210-(0.65 age) ± 10%], 3) a respiratory exchange ratio (RER) above 1.1, and 4) the incapacity to maintain the pedalling frequency imposed (70 rpm minimum) despite maximum effort and verbal encouragement. During each test, cardiorespiratory parameters and O 2 saturation were recorded. SL test allowed us to determine two groups of endurance-trained athletes: one group exhibiting EIH (n = 7) and a second group without EIH (n = 8). Mean characteristics of the three groups are presented in Table 1.

Measurements and EIH definition
Oxygen saturation level of haemoglobin was assessed by the peripheral capillary oxygen saturation (SpO 2 ). SpO 2 was measured continuously during tests using an ear-lobe pulse oximeter (Nonin, Minnesota, USA). PureSAT1 technology used in Nonin Medical pulse oximeter guarantees an measurement accuracy of ± 2.1% compared to the gold standard which is CO-oximetry analysis of arterial blood samples. The ear was pre-warmed by a vasodilating capsaicin cream (Finalgon, Fher, Spain) to avoid poor perfusion during exercise. EIH may be considered to exist when SpO 2 decreases of 4% between rest and maximal effort of the test at SL of at least the last 3 minutes [9]. At the same time, respiratory data were collected by a portable automatic breath-by-breath metabolic system (K4b 2 , Cosmed, Rome, Italy): VO 2 (ml.min -1 .kg -1 ), RER (VCO 2 /VO 2 ), minute ventilation (VE, l.min -1 ), tidal volume (V T , l), breathing frequency (B F , breaths.min -1 ). K4b 2 system was calibrated before each test according to the manufacturer's specifications: using a 3-l syringe and a gas bottle of known O 2 and CO 2 concentrations (16 and 5%, respectively). Each subject was also equipped with a chest belt (Polar Electro, Kempele, Finland) to collect HR continuously (beats.min -1 ).

Statistical treatment
The results are expressed as means ± SD. Differences with no repeated measures among the three groups were analysed using one-way analysis of variance (ANOVA). Two-way ANOVA were used for repeated measures to analyse the main effect and interaction of altitude and group on measured parameters. Correlations between the variables were tested using Pearson's product-moment correlation coefficient test. For all tests, the level of statistical significance was set at p < 0.05. Analyses were conducted using SigmaStat software (Ver 3.5).

Characteristics at rest and in response to exercise at SL
There was no difference in anthropometric data between EIH and non-EIH groups (Table 1). Untrained group had a higher body mass and body mass index compared to athletes (p < 0.05). Untrained participants presented significant differences in physiological data compared to endurance-trained groups (Table 1). At SL, there was no significant difference in SpO 2 resting values between groups. SpO 2 was normal (99 ± 0.8%) for all the participants at rest. Seven of the 15 endurance-trained athletes (~50%) exhibited EIH at SL. Delta of SpO 2 (ΔSpO 2, difference between rest and end of exercise values) at SL was significantly higher in EIH group than in non-EIH and untrained groups (Fig 1). Regarding SpO 2 kinetics at SL in EIH group reported a greater SpO 2 drop from the beginning until 60% of VO 2max and the difference was reinforced from 90% to 100% of VO 2max (Fig 2A). Performance and cardiorespiratory maximal parameters are presented in Table 2. Endurance-trained groups had a significant greater VO 2max and P max than untrained group. EIH group exhibited a lower ventilatory equivalent for O 2 (VE/VO 2max ) and CO 2 (VE/VCO 2max ) at maximal exercise while no difference occurred in HR max , B Fmax , V Tmax , VE max and duration of test.

Adaptations during exercise at ALT
At 2150 m, PIO 2 is decreased and there was a significant effect of hypoxia on SpO 2 resting values (99 ± 0.8% vs 95 ± 1.3% over all subjects) without difference among groups. SpO 2 values at the end of exercise were significantly reduced in the three groups compared to SL condition, respectively: 82 ± 4.2% vs 91 ± 1.2% in EIH group; 83 ± 2.8% vs 97 ± 1.3% in non-EIH group; 90 ± 3.5% vs 97 ± 1.8% in untrained group (Fig 2). ΔSpO 2 was significantly greater in EIH and non EIH groups compared with untrained group (Fig 1). Non-EIH group defined at SL have developed the same O 2 desaturation than EIH group defined at SL. In all groups, there was an effect of ALT condition on ΔSpO 2, reflecting O 2 desaturation during exercise. Regarding SpO 2 kinetic, most of SpO 2 decrease occurred between 60 to 80% of VO 2max in EIH group whereas it decrease between 80 to 100% of VO 2max in non-EIH trained group (Fig 2B). VO 2max , power achieved at VO 2max (P-VO 2max ) and duration of exercise at ALT were significantly lower than at SL in the three groups (Table 2). VO 2max SL-ALT (difference between SL and ALT values) was significantly greater in EIH group compared to non-EIH and untrained groups in relative values (-22 ± 7.9%, -16 ± 5.3% and -13 ± 9.4%, respectively) and in absolute values. P-VO 2max fall was around 10% in the three groups. Only EIH group exhibited a significant decrease of HR max at ALT compared to SL condition ( Table 2). When the data from all groups were pooled, there were significant relationship between VO 2max SL-ALT and SpO 2max at ALT (r = 0.53, p < 0.01) and also with HR max SL-ALT (r = 0.45, p < 0.05; Fig 3). No significant effect of ALT has been observed for VE max , B Fmax and V Tmax . B Fmax and VE/VCO 2max were lower in EIH group compared to non-EIH and untrained groups. VE max tended to be lower in EIH group compared to non-EIH group (p = 0.09).

Discussion
The main finding of this investigation is that non-EIH group had the same O 2 desaturation level at ALT with EIH athletes identified at SL. However EIH athletes have a higher decline in VO 2max , HR max and an exercise ventilatory response possibly attenuated.

Methodological considerations
For ethical reasons this study used pulse oximetry, a non-invasive method to estimate SaO 2 rather than blood arterial gases. Pulse oximetry method has been reported to deliver high precision, reproducibility and validity for O 2 saturation above 75% when compared to O 2 saturation measured from arterial blood gases at rest and exercise, at SL and at ALT [16,17]. Mollard et al. [17] have shown that the correlation between SaO 2 and SpO 2 was strong (r 2 = 0.97, p < 0.001).
In our study we are in agreement with the literature because SpO 2 values were > 75%. Further we were very rigorous in the use of this method. Being aware of the limitations caused by the use of pulse oximetry, we used an EIH definition suitable at SpO 2 measurement. Prefaut et al. [9] proposed that "EIH may be considered to exist when SaO 2 decreases of 4% from baseline with 1 to 2% accounting for the accuracy of pulse oximeters and 1 to 2% accounting for the shift to the right of the oxyhemoglobin dissociation curve during exercise due to acidosis. EIH should persist for at least the last 3 steps of an incremental exercise test or the last 3 minutes of a steady state exercise test".

O 2 desaturation in EIH athletes at ALT
At ALT, the decline in PIO 2 leads to a decline in PO 2 down the cascade from the atmosphere, to the alveoli, to the arterial blood and finally into the capillary. However, impairment in PaO 2 during exercise at ALT may not be universal for all subjects. In fact previous studies have shown a greater arterial O 2 desaturation in athletes at maximal exercise in hypoxia [5,18]. Chapman [12] suggested that when PaO 2 corresponds to the shoulder of the oxyhemoglobin dissociation curve (as for EIH) a small decline in PO 2 leads to a relatively large decline in O 2 saturation. As SpO 2 during maximal exercise at SL in EIH group was on the shoulder of the oxyhemoglobin dissociation curve, we hypothesized that in response to a same exercise done at moderate altitude can result a significantly greater SpO 2 fall in EIH group. Yet in our study, the same ΔSpO 2 has been observed at ALT between EIH and non-EIH athletes. Previous studies at higher altitude (3000-4000m) have reported controversial results but seem indicate a higher decline in SpO 2 for EIH athletes compared to non-EIH athletes [4,14,15]. Our results suggest that at ALT (~2000 m), reduce in PIO 2 is not sufficient to induce a greater O 2 desaturation in EIH athletes compared to non-EIH athletes. Our study is the first at moderate altitude, thus more studies are necessary to make a general trend.

Specific cardio-ventilatory responses of EIH athletes at ALT
At ALT, our results suggest an attenuate ventilatory response in EIH athletes but the small number of subjects in each group not allows us to clearly affirm this idea. In contrast, Gavin et al. [4] have clearly reported lower VE max , VE/VO 2max et VE/VCO 2max at 4000 m in EIH athletes. It has been generally accepted that the ventilatory response to hypoxia is mainly mediated by the sensitivity of peripheral chemoreceptors [19]. Thus, the attenuated pulmonary ventilation at ALT in EIH group could be related to an alteration of the ventilatory response to hypoxia (HVR), as an index of peripheral chemosensitivity to hypoxia. Indeed, previous studies [20,21] have indicated that EIH athletes demonstrate a lower HVR suggesting that inadequate hyperventilatory response at ALT must be due to low chemoresponsiveness. However these studies have not measured ventilatory parameters during exercise in hypoxia. A potential low Correlations between decrease in VO 2max (VO 2max SL-ALT, difference between sea level and altitude values) with SpO 2 values at the end of exercise at altitude (SpO 2max ALT) and with decrease in HR max (HR max SL-ALT, difference between sea level and altitude values). All participants were considered, n = 23.
doi:10.1371/journal.pone.0161819.g003 chemoresponsiveness in EIH group during exercise at ALT can be a surprising result because intermittent hypoxia exposure (as during training in EIH group) has been repeatedly shown to increase HVR [22][23][24][25]. This discrepancy can be explained by at least two reasons. Firstly, the increase in HVR seems to enhance exercise ventilation and improve SaO 2 depending on the altitude where exercise is performed, since it was observed at high altitude (4500 m) [22] but not at moderate altitude (2500 m or below) [26,27]. An explanation can be the level of arterial oxygenation. In fact mean SpO 2 values of our subjects during exercise at moderate ALT were around 80% whereas at 4500 m the values were around 60% [22]. As a consequence the stimulus to the peripheral oxygen sensing chemoreceptors was markedly lower in our subjects and could have contributed only slightly to the regulation of exercise ventilation. Secondly, in general, athletes are known to express low exercise ventilation and are reported to illustrate low chemoresponsiveness at SL [28]. The blunted chemosensitivity to HVR in the endurance athletes has been considered to be due to endurance training over long periods [29]. Thus exercise training and exposure to intermittent hypoxia might have opposite effects on HVR. At 4500 m, Katayama et al. [30] have observed that HVR increased significantly in control group but not in training group. These results would suggest that endurance training during intermittent exposure to hypoxia depresses the increment of chemosensitivity to hypoxia. Thus, the larger relative hypoventilation in EIH group during exercise in moderate ALT could be due to a larger effect of aerobic training on chemoreceptor desensitization. Concerning cardiac response, Mollard et al. [7] have noted that HR max decreased from 1000 m in trained subjects and from 2500 m in untrained subjects. Our results confirm the decrease in HR max at acute moderate ALT in EIH athletes but not in non-EIH athletes. This greater decrease in HR max at ALT in EIH athletes compared to non-EIH athletes has been observed in previous studies but not at moderate ALT [14,15]. Autonomic changes could be involved in the reduction in HR max at ALT in EIH group. Indeed, prior studies have provided a gradual down regulation of β-adrenergic receptors [31] and an up-regulation of muscarinic receptors [32] over the time of exposure to ALT, as a possible mechanism of myocardial adaptation to hypoxia. Further it has been shown that chronic exercise, as chronic exposure to hypoxia, reduced the number of cardiac β receptors [33]. Perhaps the frequent repetition of long-duration exercises with training in subjects who exhibit an arterial O 2 desaturation could induce a similar mechanism of myocardial adaptation and could explain the decrease in HR max at ALT in EIH group.

Consequences of EIH on VO 2max at moderate altitude
At SL, EIH is already known to negatively affect VO 2max [8] and subsequently exercise performance [10]. In this study at 2150 m, EIH athletes demonstrated a higher decline in VO 2max from SL than non-EIH athletes. From an investigation compiling the results of 11 studies examining the magnitude of VO 2max decline with acute ALT exposure in endurance-trained athletes, VO 2max declines by 7.7% for every 1000 m ascended above SL [1]. The mean decrease in VO 2max in non-EIH athletes in this present study was 7.4% per 1000 m, which is in accordance with the literature. In EIH athletes, VO 2max fall is higher (10% per 1000 m) exceeding the standards reported in the literature. Recently, Chapman et al. [12] suggested that the degree of arterial desaturation during maximal exercise at SL, and not baseline VO 2max levels per se, is a primary limiting factor determining VO 2max decline with exposure to acute moderate ALT. But as previously described, SpO 2 fall at moderate ALT was equal in both EIH and non-EIH groups, bringing into question the contribution of other limiting factor(s) determining VO 2max decline in EIH athletes. According to Fick equation, a fall in HR max at ALT can have an effect on VO 2max via a reduction of maximal cardiac output (Q max ). HR max decrease under ALT was higher in EIH than in non-EIH athletes and even more compared to untrained participants. Thus greater VO 2max decrement in EIH athletes could be explained by a larger reduction in O 2 transport than non-EIH athletes. However in the present study, Q max had not been measured; the effects of modification of HR max on Q max should be interpreted with caution. Further although EIH group shows a significant decrease in HR max , this value is not significantly different between the two athletes groups at altitude.
This study has also shown a decrease of P-VO 2max at altitude around 10% despite a VO 2max decrease around 22% in EIH group. Thus it seems that the decrease of P-VO 2max at altitude is not proportional with the VO 2max decrease. As in our study, Peltonen and colleagues noted that hypoxia of 2 500 m reduced VO 2max nearly twice as much as maximal power during a maximal exercise on cycle ergometer [34]. Several studies have shown that chronic [35] and acute [36,37] lack of oxygen supply to muscle, reduces the leftwards shift of the electromyogram power spectrum density during sustained static contractions of 60-80% of maximum voluntary contraction. This reduction in the power spectrum density corresponds to a preferential recruitment of fast motor units [38], which may constitute an adaptive process to limit the recruitment of slow motor units which are highly oxygen dependent [36]. The preferential recruitment of fast motor units could constitute an adaptive muscle response to a reduced oxygen supply and can explain the lesser decrease of P-VO 2max compared to VO 2max decrease in hypoxia. Further investigations are needed to confirm this hypothesis.

Conclusion
EIH athletes have shown a greater aerobic impairment than non-EIH athletes at altitude in spite of a same O 2 desaturation level. Furthermore they also demonstrated a greater decrease in HR max and potentially ventilatory adaptations at altitude. Thus, EIH athletes develop specific cardiorespiratory adaptations during exercise at acute moderate altitude maybe due to their frequent exposures to hypoxia. Based on these results, this study highlights the importance to discriminate EIH athletes in studies involving exercise adaptations in endurance athletes at altitude. Definitely further investigations into the mechanisms of EIH at moderate altitude are warranted.