Journal Logo


Impairment of 3000-m Run Time at Altitude Is Influenced by Arterial Oxyhemoglobin Saturation


Author Information
Medicine & Science in Sports & Exercise: September 2011 - Volume 43 - Issue 9 - p 1649-1656
doi: 10.1249/MSS.0b013e318211bf45
  • Free


Athletic competitions involving endurance exercise at moderate altitude are characterized by impaired performances compared with sea level. Several studies in the late 1960s and early 1970s, coincident with the awarding of the 1968 Olympic Games to Mexico City (altitude = 2290 m), examined the relationship between altitude and distance running performance impairment and attempted to predict which particular physiological characteristics may be related to competitive performance decline. In 1971, Dill and Adams (11) noted that highly trained runners might be handicapped "to an unusual extent" at altitude compared with lesser trained individuals. In the 40 yr that have followed, several investigations have confirmed a strong positive relationship between maximal oxygen uptake (V˙O2max) at sea level and the magnitude of the decline in V˙O2max at altitude (14,15,24,36,40).

One of the most important mechanisms behind the relationship between V˙O2max and the decline in V˙O2max during acute exposure to altitude seems to be the degree of pulmonary gas exchange limitations during exercise. For example, significant correlations have been demonstrated between the decline in V˙O2max at altitude and a) arterial oxyhemoglobin saturation (SaO2) during maximal exercise (7,14,15,24) and b) the ratio of lung diffusing capacity to V˙O2max (5). Therefore, athletes with smaller degrees of pulmonary gas exchange limitations during heavy exercise seem to be better able to maintain V˙O2max at altitude. However, whereas the ability to consume oxygen at a high rate has been deemed essential to success in endurance activities (29,32), the subsequent link connecting the maintenance of SaO2 and V˙O2max to actual running performance with acute altitude exposure has not been established. In fact, among elite athletes, it has been argued that V˙O2max provides little discriminating power concerning actual race performance (32). Whether or not this relationship is more robust at altitude is unknown.

The principal aim of this study was to determine the relationship between SaO2 during heavy exercise and the impairment in race performance with acute altitude exposure in elite distance runners. We hypothesized that athletes who demonstrated prominent oxyhemoglobin desaturation during race pace running at sea level would demonstrate the largest reductions in 3000-m time trial performance at altitude. If true, prescreening SaO2 during heavy exercise at sea level may help predict which athletes would experience the greatest performance limitations with acute altitude exposure. In addition, because of the minimal amount of data in the literature on elite female athletes, physiological and performance-based sex differences were also examined to determine whether elite men and women runners are affected to a similar extent at moderate altitude.



Twenty-seven US national-class distance runners (18 men and 9 women) agreed to participate in the project. All were current or former US collegiate NCAA All-Americans in track and field, and the group included two US Olympians. Twenty-five of the 27 athletes were ranked in the top 50 in the United States in their primary event in the year of the study. Exclusion criteria included altitude residence (>1000 m) >7 d in the previous 6 months, permanent altitude residence of >3 months during their lifetimes, or injury or illness that impaired normal training and racing before the study. All subjects gave written informed consent to a protocol approved by the Institutional Review Board of The University of Texas Southwestern Medical Center at Dallas and Presbyterian Hospital of Dallas. Subjects' characteristics are presented in Table 1. The specialty competitive events of these athletes included the 1500 m (n = 3), 3000-m steeplechase (n = 9), 5000 m (n = 8), 10,000 m (n = 6), and the marathon (n = 1). This study was part of a larger research project examining chronic altitude training effects, the primary outcomes of which have been reported elsewhere (34).

Descriptive data.

Experimental sequence.

Each subject participated in a treadmill testing session at sea level (Bloomington, IN; elevation = 230 m; barometric pressure = 736 ± 1 mm Hg, mean ± SD), 2-7 d before exposure to altitude. On the evening before departure for altitude, each subject participated in a 3000-m time trial on a standard 400-m all-weather outdoor track. The morning after the sea-level time trial, the subjects were flown by commercial jet to Salt Lake City, UT, and were immediately transported by van to Deer Valley, UT (elevation = 2500 m), arriving at noon local time. Forty-eight hours after arrival, each subject participated in a second 3000-m time trial on a 400-m all-weather outdoor track at an elevation of 2100 m. The 3000-m time trial distance was chosen because it best matched the specialty events of the athlete cohort, it results in energetic supply (i.e., ATP utilization) primarily from oxidative sources (13), and the recovery between successive trials separated by 66 h would be more complete than time trials utilizing a longer distance (such as 5000 or 10,000 m). The sea-level 3000-m time trial was performed at dusk (20:00), and the altitude time trial was performed at noon, in an attempt to match ambient temperatures as closely as possible between the two trials. The nature of the experimental design made it impossible to control other ambient conditions between the trials, such as wind velocity and relative humidity (sea level: temperature = 26.8°C, relative humidity = 75%; altitude: temperature = 25.1°C, relative humidity = 48%). Time was recorded for each subject to the nearest 0.1 s.

Exercise testing.

The treadmill test consisted of three phases: a warm-up, a constant-load bout at a simulated "race pace" breathing normoxic and hypoxic inspirates, and an incremental test to exhaustion breathing a normoxic inspirate. After a 10- to 15-min warm-up outside of the laboratory, each subject began the exercise test by walking on a motor-driven treadmill at 4.8 km·h−1 with 0% grade. After 2 min of walking, a 4-min warm-up period was completed at 15.2 km·h−1 for men or 13.6 km·h−1 for women to become acclimated to treadmill running. The speed of the treadmill was then increased to either 20.6 or 21.5 km·h−1 for men or 18.0 km·h−1 for women with 0% grade, and each subject was asked to hold this constant simulated race pace for 4 min. At the end of 2 min at this simulated race pace, the inspirate was switched from room air to 16.3% O2 (simulating an altitude of ∼2100 m) for the final 2 min of exercise. Data for analysis were averaged during the final minute of exercise in each of the normoxic and hypoxic conditions. This 4-min constant-load race pace protocol with 2 min each in normoxia and hypoxia was used because separate incremental tests to exhaustion in normoxia and hypoxia on separate days were not logistically possible.

After a rest/recovery period of no less than 5 min, an incremental treadmill test to exhaustion was performed at a constant speed of 16 km·h−1 for men and 13.6 km·h−1 for women. After 2 min of running at 0% grade, the slope of the treadmill was increased by 2% every 2 min until volitional fatigue. The highest V˙O2 recorded was chosen as the subject's V˙O2max because all subjects met at least two of the following three criteria: 1) an HR in excess of 90% of the age-predicted maximum (220 − age), 2) an RER of ≥1.10, and 3) a plateau (<150-mL increase) in V˙O2 with an increase in workload.

Ventilatory and metabolic variables were continuously measured and monitored during exercise using a computer-interfaced open-flow indirect calorimetry system. Minute ventilation (E) was determined using a dual thermistor flow probe (Torrent 1200; Hector Engineering, Ellettsville, IN) on the inspired side. Subjects breathed through a low-resistance two-way valve (model 2700; Hans Rudolph, Kansas City, MO), and a 5-L mixing chamber was used for the collection of expired gases. Fractional concentrations of O2 and CO2 were determined from a continuous sample of expired gas using a mass spectrometer (RAMS M-100; Marquette, Milwaukee, WI). SaO2, V˙O2, and V˙CO2 were averaged during each minute of exercise, and V˙O2 and V˙CO2 were corrected to STPD. The above variables and fractions of end-tidal O2 and CO2 were continuously measured and monitored with a data acquisition control system (WorkBench for Windows 2.0; Strawberry Tree, Sunnyvale, CA) sampling at 40 Hz.

Arterial oxygen saturation measurement.

SaO2 was estimated using ear oximetry (model 47201A; Hewlett-Packard (HP), Waltham, MA). The oximeter was calibrated using an internal protocol before each test according to the manufacturer's instructions. This instrument has been reported to be accurate across a wide range of SaO2, with arterial SaO2 values >75% underestimated by <2% (31). The "older" HP 47201A oximeter is unique because it uses eight wavelengths covering the range from 650 to 1050 nm and 18 molar extinction coefficients to determine oxy- and deoxyhemoglobin concentrations, whereas most modern oximeters use only two or three wavelengths. Measuring transmittances at this large number of suitably chosen wavelengths allows for calculation of SaO2 values that are satisfactorily independent of factors such as movement of the earpiece on the ear, skin pigmentation, or the waveform of the pulse. In addition, a standard spectrophotometer cuvette was specially adapted in our laboratory to fit in the sample-viewing slit of the oximeter. Freshly drawn whole blood was tonometered to produce a range of O2 saturations and was sampled in the cuvette by the ear oximeter. One-minute computer averages of SaO2 via the ear oximeter were referenced against the average of four samples analyzed by an OSM3 hemoximeter (Radiometer, Copenhagen, Denmark). In the range of 60%-95% SaO2, the two independent measures never differed by more than 1.9% and were tightly correlated (r = 0.99) (7). After data collection in this study, the ear oximeter was again checked versus blood samples analyzed with an ABL3000 blood gas analyzer (Radiometer) with similar results. For SaO2 ranging from 65% to 98%, the ear oximeter and blood gas analyzer were significantly correlated (r = 0.99), the average deviation across the range was <1.0%, and the maximum deviation at any point was 2.1% (10). The oximeter earpiece was attached to a headgear worn by the subject, and an elastic bandage was used to secure the earpiece and prevent movement from the pinna of the ear.

Statistical analysis.

Group responses (men only, women only, or all subjects) in normoxia and hypoxia were compared using paired Student's t-tests. Subjects were divided into groups on the basis of SaO2 during race pace exercise in normoxia, utilizing classifications published elsewhere (7,20). Subjects with SaO2 > 93% were placed in a group designated as HiSat (n = 7, four females and three males), whereas subjects with SaO2 < 91% were placed in a LoSat group (n = 7, seven males). These classifications with a 2% gap in SaO2 measures were used because of the inherent variability in oximetry estimates of SaO2. With previously published maximum deviations of SaO2 of 1.9% (7) and 2.1% (10) between the oximeter used in this study and direct blood gas measures, this classification strategy was chosen to help ensure that group assignments are based on a true difference in SaO2. Pearson product-moment correlations were used to determine relationships between dependent variables for all subjects. A modified Bonferroni procedure was used to maintain a familywise α = 0.05, and P ≤ 0.05 was required for significance in all comparisons.


Race performance.

Data from the 3000-m time trials are presented in Table 2. One subject (male) was unable to complete the altitude time trial because of illness, leaving 26 athletes for comparison. Group 3000-m performance was significantly slower at altitude compared with sea level (Δ3000-m time = 48.5 ± 12.7 s). Absolute declines in mean performance were nearly identical in the group of 17 men (48.4 ± 14.6 s) and the group of 9 women (48.6 ± 8.9 s). However, this decline trended as a larger percentage of total running time in men (9.7% ± 2.7%, range = 4.5%-13.7%) compared with women (8.5% ± 1.6%, range = 6.3%-11.5%, P = 0.08).

The 3000-m time trial data.

Metabolic data.

Data from the race pace exercise bout in normoxia and hypoxia are shown in Table 3. Of the 27 athletes, five were unable to complete both minutes of race pace exercise in hypoxia because of volitional fatigue. These five subjects did not display common characteristics that would otherwise explain their inability to complete the exercise bout in hypoxia. Therefore, in Table 3, data are presented only from the 22 athletes who completed the entire constant-load exercise bout.

Ventilatory and metabolic measures during race pace exercise.

By design, the LoSat group had a significantly lower SaO2 during the race pace exercise in normoxia compared with the HiSat group (88.6% ± 2.5% vs 94.2% ± 0.7%). Although the HiSat group had a mix of gender (four females and three males) and the LoSat group was composed entirely of males, there was no difference in relative V˙O2max between the two groups (LoSat = 73.2 ± 4.9 mL·kg−1·min−1, HiSat = 69.8 ± 7.6 mL·kg−1·min−1). The decline in race pace V˙O2 from normoxia to hypoxia in the LoSat group was significantly larger (−9.2 ± 2.1 mL·kg−1·min−1) than the decline in the HiSat group (−3.5 ± 2.0 mL·kg−1·min−1; Fig. 1). Similarly, the LoSat group had a significantly greater slowing of 3000-m race performance from sea level to altitude, compared with the HiSat group (54.0 ± 13.7 vs 38.9 ± 9.7 s).

Differences between LoSat and HiSat groups in (left) the change in race pace V˙O2 between normoxia and hypoxia (16.3% O2) and (right) the change in 3000-m race time between sea level and an altitude of 2100 m. Values are means ± SE. *Significantly different from HiSat, P ≤ 0.05.

For all subjects, SaO2 during normoxic race pace exercise was significantly correlated with the decline in V˙O2 from normoxia to hypoxia during the race pace bout (r = −0.68, n = 22; Fig. 2). Normoxic race pace SaO2 was also significantly correlated with the slowing of 3000-m race time from sea level to altitude (r = −0.38, n = 26; Fig. 3). Therefore, athletes with the lowest SaO2 during normoxic race pace exercise had the largest decline in V˙O2 from normoxia to hypoxia and also had the largest slowing of 3000-m race time from sea level to altitude.

Relationship between SaO2 during race pace running in normoxia and the reduction in V˙O2 from normoxia to hypoxia (16.3% O2) during race pace running. Closed circles indicate men; open circles, women.
Relationship between SaO2 during race pace running in normoxia and the change in 3000-m race time from sea level to an altitude of 2100 m. Closed circles indicate men; open circles, women.

Sex comparisons.

Sex differences were evident in both the V˙O2 (L·min−1 and mL·kg−1·min−1) and E achieved during race pace exercise, with the men having significantly higher values than the women. However, normoxic race pace exercise elicited similar percentages of V˙O2max across sexes (men = 88.4% ± 3.9% of V˙O2max, women = 89.7% ± 3.1% of V˙O2max). SaO2 was significantly higher during race pace exercise in normoxia in the group of women (93.2% ± 1.3%) versus the men (91.1% ± 2.7%). The group of men demonstrated a significantly larger increase in E from normoxia to hypoxia (±27.4 ± 11.0 L·min−1) compared with the women (±16.6 ± 4.5 L·min−1). However, no difference existed between sexes in the increase in the ventilatory equivalent for O2 (E/V˙O2) from normoxia to hypoxia (ΔE/V˙O2: men = 9.7 ± 2.5, women = 9.0 ± 2.2). No other sex differences were observed in the change of metabolic or ventilatory variables from normoxia to hypoxia during race pace running.


The principal new observation from this investigation is that the decline in 3000-m race performance at acute altitude in highly trained distance runners is related to the ability to maintain SaO2 and oxygen uptake during heavy exercise in hypoxia. The data confirm the already well-established relationship between the degree of pulmonary gas exchange limitations and the decline in V˙O2max from normoxia to hypoxia and add new understanding by extending this relationship to include a measure of running performance.

Maximal oxygen transport to skeletal muscle has a strong controlling influence on V˙O2max (6,12,30), although it is not the sole limiting factor (38). In acute hypoxia, maximal cardiac output is unchanged compared with normoxia (33); therefore, unless maximal flow to the working muscles is augmented, O2 transport and V˙O2max will ultimately be reduced with a falling arterial saturation during hypoxic exercise. On this theoretical basis, several studies have subsequently demonstrated that individuals who are unable to maintain SaO2 during heavy exercise at sea level are less able to maintain V˙O2max at altitude. In a mixed group of trained and untrained male subjects, Lawler et al. (24) reported a strong correlation (r = −0.86) between SaO2 during maximal exercise and the decline in V˙O2max from normoxia to acute hypoxia (14% O2). Another investigation using only highly trained male endurance athletes found a significant correlation (r = −0.54) between SaO2 at V˙O2max in normoxia and the decline in V˙O2max from normoxia to mild (18.7% O2) hypoxia (7). The significant correlation found in the present study between SaO2 during race pace exercise in normoxia and the decline in race pace V˙O2 from normoxia to hypoxia is similar to these findings from other investigations and confirms these observations.

One key difference in methodology was our use of a constant-load race pace exercise bout in normoxia and hypoxia versus separate normoxic and hypoxic incremental tests to exhaustion. Whereas the latter strategy is commonly used, we felt our race pace strategy was appropriate for several reasons. First, by selecting a high-intensity yet submaximal workload in normoxia, this work rate functionally becomes supramaximal when the inspirate is switched to hypoxia. Therefore, the athletes with the largest reduction in V˙O2 in hypoxia at a constant work rate ultimately are the same athletes who demonstrate the greatest decline in V˙O2max at altitude. Second, the logistics of the timing of this study and our access to this unique athlete population did not allow for separate incremental tests to exhaustion to be completed on separate days with the appropriate recovery. In addition, simulating the race conditions on the treadmill, in terms of a fast running pace and zero incline, best matched the specific physiological load the athlete would face while performing the time trials on the track. Therefore, the responses of metabolic variables observed using a race pace protocol were likely more appropriate, specifically in terms of characterizing and predicting race performance, than those obtained through a progressive work rate and a changing treadmill grade.

In an effort to confirm the ability of the race pace protocol to appropriately indicate the V˙O2 response to hypoxia, we compared the present data with that of Wehrlin and Hallén (39). Their analysis of 11 separate studies that measured V˙O2max decline from sea level to various altitudes in elite athletes indicates an average decline in V˙O2max of 7.7% per 1000 m of altitude ascended above sea level. This analysis would suggest that V˙O2max would likely be reduced, on average, by 16.1% between sea level and an altitude of 2100 m. For the men in the present study, the mean reduction between V˙O2max measured in normoxia and the hypoxic race pace V˙O2 (which is assumed to be supramaximal) was 17.7%, and for women, the difference was 15.7%. In addition, we examined HR values at the end of the hypoxic race pace bout (184.0 ± 8.3 beats·min−1) and at V˙O2max in normoxia (186.8 ± 7.3 beats·min−1). Each of the aforementioned values would suggest that the race pace protocol and the V˙O2 data and relationships derived from there would likely be similar and comparable to data derived from separate incremental tests to exhaustion in normoxia and hypoxia. Still, it is important to note that the relationship between SaO2 and the decline in V˙O2max between normoxia and hypoxia has been well established, and the race pace V˙O2 measures obtained in this study simply confirm this relationship. In contrast, the primary outcome of interest for this study was SaO2 and the change in 3000-m performance time between sea level and altitude.

The significant but admittedly loose correlation between SaO2 during race pace running and the slowing of 3000-m time from sea level to altitude extends the established SaO2-ΔV˙O2 relationship to include a measure of exercise performance. Obtaining useful measures of exercise performance, either in the field or in the laboratory, can be confounded by the influence of psychological factors such as motivation, anxiety, strategy, and learning effects (35,37). Each of these and other relevant factors may add considerable noise to measures of performance, which often makes correlations of race performance with less variable physiological measures difficult. Because this study attempted to determine the change in race time between two separate field measures of running performance, the magnitude of "nonphysiologic" variability in correlations involving the Δ3000-m race time variable was likely increased. The fact that the group of subjects was national-class distance runners, each with many years of experience of racing in US collegiate and open track meets, certainly helped to minimize-but not eliminate-the overall influence of psychological factors on within-subject time trial variability. Ultimately, it is important to consider that the SaO2-Δ3000-m race time relationship achieved statistical significance, despite the inherent variability of the performance measures. However, data collected in this study only accounted for roughly 15% of the total variation in the decline in 3000-m time trial performance from sea level to altitude.

The choice of using an oximetry measure of SaO2 versus arterial blood may also have added to the total variability within the SaO2-Δ3000-m race time comparison. An in vitro calibration procedure performed before this study, combined with published comparisons of the HP 47201A ear oximeter with arterial blood gases during exercise (26,31), suggests this model of oximeter is accurate to within 2% of blood gas measures of SaO2. In an attempt to minimize the influence of this small degree of variability in oximetry measures of SaO2, subjects were divided into groups on the basis of SaO2 measures during normoxic race pace exercise (LoSat: SaO2 < 91%; and HiSat: SaO2 > 93%). The 2% SaO2 difference between the border of the two divisions allowed for increased certainty that subjects placed into each of the HiSat and LoSat groups truly had differing SaO2 responses, using an oximetry estimate. Previous studies using oximetry measures of SaO2 and a similar classification strategy (with a 2% SaO2 difference between group divisions) have determined that athletes who demonstrate exercise-induced hypoxemia display a reduced ventilatory response to exercise (20) and a significantly larger reduction in V˙O2max in mild hypoxia (7), compared with normoxemic athletes. In the present study, the LoSat group demonstrated a significantly larger decline in race pace V˙O2 in hypoxia and slowing of 3000-m time at altitude, compared with the HiSat group-lending further support for a potential link between arterial saturation maintenance during heavy exercise and the decline in running performance at altitude.

Limitations in pulmonary gas exchange during heavy exercise in endurance-trained athletes are believed to be due to a combination of factors, including diffusion limitations in the lung (creating a widened alveolar − arterial Po2 difference) and an inadequate hyperventilatory response to exercise (creating a reduced alveolar Po2) (9). Whereas the widened alveolar − arterial Po2 difference is believed to be the primary factor behind arterial oxyhemoglobin desaturation in highly trained endurance athletes (8,27), we found a significant correlation between race pace SaO2 in normoxia and race pace E/V˙O2 in normoxia (r = 0.62, P ≤ 0.01, n = 27). This result suggests that the athletes with the greatest ventilatory response to exercise best defended exercise arterial O2 saturation. Similarly, the ventilatory equivalent for O2 during race pace exercise in normoxia was significantly higher in the HiSat group (E/V˙O2 = 32.8 ± 2.7) versus the LoSat athletes (E/V˙O2 = 27.9 ± 1.8, P ≤ 0.05). Why the LoSat athletes did not exhibit a higher ventilation in normoxia to maximize alveolar Po2 and defend SaO2 is unclear because E was higher in each athlete when the inspirate was switched to 16.3% O2. A reduced ventilatory response during exercise may be related to mechanical constraints on exercise ventilation (4,23) or the metabolic cost associated with exercise hyperpnea (1). Similarly, a blunted chemoreceptor response has been shown to result in a reduced ventilatory drive during heavy exercise (20). In contrast, however, the hypoxic ventilatory response has been shown to be a poor predictor of the magnitude of exercise-induced arterial desaturation in endurance athletes exercising at high work rates (16,21). Whereas the greater ventilatory equivalent for O2 during exercise in the HiSat group may be related to SaO2 maintenance during heavy exercise, it is important to note that increased inspiratory muscle work has been shown to be related to greater locomotor muscle fatigue, particularly during heavy exercise in hypoxia (2,3). At this point, it is unknown if differing ventilatory responses to the 3000-m time trials at sea level and altitude affected both blood flow to the locomotor muscles (via a "metaboreflex" response [17]) and the rate of appearance of peripheral muscle fatigue. In any case, the significance of the ventilatory relationships found in this study by no means excludes the possibility of pulmonary diffusion limitations secondary to other mechanisms, such as ventilation-perfusion mismatching (22), short transit times of the erythrocyte in the pulmonary circulation (8), or the formation of interstitial edema in the alveolar space (25), contributing to the reduced SaO2 measures seen in a large number of our elite athlete cohort.

Taking the correlational and HiSat/LoSat group analyses together, it seems that the maintenance of SaO2 during heavy exercise is an important determinant of V˙O2 and performance at moderate altitude; however, a considerable portion of the variance remains unexplained. Other factors influencing the individual variation in the ability to compensate for ambient hypoxia, such as local control of blood flow, the hypoxic ventilatory response, and the degree of expiratory flow limitations, may also influence the decline in racing performance at altitude to some degree. A multivariate analysis with SaO2 and the measures listed above may help elucidate a more complete mechanistic explanation behind racing performance declines at moderate altitude. However, the inherent noise in measures of running performance may be difficult to overcome. Still, screening SaO2 during heavy or maximal exercise at sea level may serve as a helpful predictive measure to advise athletes who are competing with acute exposure to moderate altitude. For example, five of the seven HiSat athletes (SaO2 > 93%) displayed a slowing of 3000-m time at altitude that was less than the mean of the overall athlete cohort (n = 26, Δ3000-m time = 48.5 s). Similarly, five of the seven LoSat athletes (SaO2 < 91%) had a larger slowing of 3000-m time at altitude versus the group mean. Therefore, although SaO2 and our group designations are not absolute predictors of the magnitude of performance decline, 71% of the time, the HiSat/LoSat divisions did correctly predict whether an athlete might experience performance declines more or less than the group mean. For an elite athlete, for whom just a few seconds over 3000 m can encompass several places in the order of finish, screening exercise SaO2 at sea level may give an important indication of whether his/her individual performance decline with acute altitude exposure might be more or less than that of the average athlete.

Several interesting sex differences were evident in the mean group responses and correlational relationships. In the group of women, the relationship between normoxic race pace SaO2 and the change in race pace V˙O2 in hypoxia (r = −0.48) did not reach significance, which might have been because of the small sample size of women in this study (power analysis indicates four additional women subjects are needed to show statistical significance at α = 0.05 with r = −0.48). It is important to note that the women did not demonstrate as wide of a range in SaO2 measures during race pace exercise (range = 90.9-95.4) as the men (range = 85.9-94.5), which may have contributed to the failure of establishing significant correlations involving SaO2 within the cohort of women athletes. In our data, the group of nine women did have a significantly higher mean SaO2 during race pace exercise (2.1%) than the men. In contrast, evidence suggests that highly trained female distance runners may be more susceptible to pulmonary gas exchange limitations during heavy exercise than male athletes with similar relative V˙O2max values (18,19). A work by Harms et al. (19) suggests that within females, for every 1% reduction in SaO2 below resting levels, V˙O2max is reduced ∼2%. Compared with men, this reduction in V˙O2max is between two (28) and four (7) times greater for a 1% reduction in SaO2, suggesting that women may experience a greater racing performance decline at altitude than men. However, our data do not suggest the presence of a sex difference in performance decline at altitude in national-class endurance athletes because the group of women and men demonstrated almost identical mean increases in 3000-m race time at altitude (48.6 vs 48.4 s). Interestingly, however, it should be noted that in terms of percentage of total race time, this level of performance impairment trends slightly smaller in the group of women than in the men (8.3% vs 9.6%, P = 0.08), which is opposite of the expected response. Whereas our cohort of women athletes did not demonstrate the same degree of oxyhemoglobin desaturation during exercise as the men, larger groups of highly trained women endurance athletes will need to be studied to confirm the presence or absence of sex-specific relationships between SaO2 maintenance during heavy exercise and the degree of V˙O2max and performance decline at altitude.

We conclude that the degree of pulmonary gas exchange limitations during exercise, already known to contribute to the magnitude of the decline in V˙O2max at altitude, also contributes to the extent of the decline in 3000-m race performance at moderate altitude in highly trained distance runners. Athletes who can best maintain SaO2 during race pace running at sea level demonstrate less performance impairment at altitude, compared with athletes with similar levels of V˙O2max but more reduced measures of SaO2 during heavy exercise. Decrements in 3000-m race performance at moderate altitude are similar between elite men and women distance runners. We propose that screening SaO2 during heavy exercise at sea level may be a useful tool in predicting which athletes will experience the largest and smallest declines in distance running performance with acute exposure to moderate altitude.

This work was supported by US Olympic Committee grant SST97-ATH-007, a grant from USA Track & Field, and institutional support from Presbyterian Hospital of Dallas.

The results of this study do not constitute endorsement by the American College of Sports Medicine.


1. Aaron EA, Seow KC, Johnson BD, Dempsey JA. Oxygen cost of exercise hyperpnea: implications for performance. J Appl Physiol. 1992;72:1818-25.
2. Amann M, Romer LM, Pegelow DF, Jacques AJ, Hess CJ, Dempsey JA. Effects of arterial oxygen content on peripheral locomotor muscle fatigue. J Appl Physiol. 2006;101:119-27.
3. Amann M, Romer LM, Subudhi AW, Pegelow DF, Dempsey JA. Severity of arterial hypoxaemia affects the relative contributions of peripheral muscle fatigue to exercise performance in healthy humans. J Physiol. 2007;581:389-403.
4. Babb TG. Ventilatory response to exercise in subjects breathing CO2 or HeO2. J Appl Physiol. 1997;82:746-54.
5. Blomqvist G, Johnson RL Jr, Saltin B. Pulmonary diffusing capacity limiting human performance at altitude. Acta Physiol Scand. 1969;76:284-7.
6. Buick FJ, Gledhill N, Froese AB, Spriet L, Meyers EC. Effect of induced eyrthrocythemia on aerobic work capacity. J Appl Physiol. 1980;48:636-42.
7. Chapman RF, Emery M, Stager JM. Degree of arterial desaturation in normoxia influences V˙O2max decline in mild hypoxia. Med Sci Sports Exerc. 1999;31(5):658-63.
8. Dempsey JA. Exercise-induced imperfections in pulmonary gas exchange. Can J Sport Sci. 1987;12:66s-70s.
9. Dempsey JA, Hanson PE, Henderson KS. Exercise induced arterial hypoxemia in healthy persons at sea level. J Physiol (Lond). 1984;355:161-75.
10. Derchak PA, Stager JM, Tanner DA, Chapman RF. Expiratory flow limitation confounds ventilatory response during exercise in athletes. Med Sci Sports Exerc. 2000;32(11):1873-9.
11. Dill DB, Adams WC. Maximal oxygen uptake at sea level and at 3,090-m altitude in high school champion runners. J Appl Physiol. 1971;30:854-9.
12. Ekblom B, Huot R, Stein EM, Thorstensson AT. Effect of changes in arterial oxygen content on circulation and physical performance. J Appl Physiol. 1975;39:71-5.
13. Gastin PB. Energy system interaction and relative contribution during maximal exercise. Sports Med. 2001;31:725-41.
14. Gavin TP, Stager JM, Derchak PA. Ventilation's role in the decline in V˙O2max and SaO2 in acute hypoxic exercise. Med Sci Sports Exerc. 1998;30(2):195-9.
15. Gore CJ, Hahn AG, Scroop GS, et al. Increased arterial desaturation in trained cyclists during maximal exercise at 580 m altitude. J Appl Physiol. 1996;80:2204-10.
16. Guenette JA, Diep TT, Koehle MS, Foster GE, Richards JC, Sheel AW. Acute hypoxic ventilatory response and exercise-induced arterial hypoxemia in men and women. Respir Physiol Neurobiol. 2004;143:37-48.
17. Harms CA, Babcock MA, McClaran SR, et al. Respiratory muscle work compromises leg blood flow during maximal exercise. J Appl Physiol. 1997;82:1573-83.
18. Harms CA, McClaran SR, Nickele G, Pegelow D, Dempsey JA. Exercise induced arterial hypoxaemia in highly trained women athletes. J Physiol (Lond). 1998;507(Pt 2):619-29.
19. Harms CA, McClaran SR, Nickele G, Pegelow D, Nelson WB, Dempsey JA. Effect of exercise-induced arterial O2 desaturation on V˙O2max in women. Med Sci Sports Exerc. 2000;32(6):1101-8.
20. Harms CA, Stager JM. Low peripheral chemoresponsiveness and inadequate hyperventilation contribute to exercise-induced hypoxemia. J Appl Physiol. 1995;79:575-80.
21. Hopkins SR, McKenzie DC. Hypoxic ventilatory response and arterial desaturation during heavy work. J Appl Physiol. 1989;67:1119-24.
22. Hopkins SR, McKenzie DC, Schoene RB, Glenny RW, Robertson HT. Pulmonary gas exchange during exercise in athletes. I. Ventilation-perfusion mismatch and diffusion limitation. J Appl Physiol. 1994;77:912-7.
23. Johnson BD, Saupe K, Dempsey JA. Mechanical constraints on exercise hyperpnea in endurance athletes. J Appl Physiol. 1992;73:874-86.
24. Lawler J, Powers SK, Thompson D. Linear relationship between V˙O2max and V˙O2max decrement during exposure to acute hypoxia. J Appl Physiol. 1988;64:1486-92.
25. Marshall BE, Teichner RL, Kallos T, Sugerman HJ, Wyche MQ Jr, Tantum KR. Effects of posture and exercise on extravascular lung water volume in man. J Appl Physiol. 1971;31:275-9.
26. Poppius H, Viljanen AA. A new oximeter for assessment of exercise-induced arterial desaturation in patients with pulmonary diseases. Scan J Respir Dis. 1977;58:279-83.
27. Powers SK, Lawler J, Dempsey JA, Dodd S, Landry G. Effects of incomplete pulmonary gas exchange on V˙O2max. J Appl Physiol. 1989;66:2491-5.
28. Powers SK, Martin D, Cicale M, Collop N, Huang D, Criswell D. Exercise-induced hypoxemia in athletes: role of inadequate hyperventilation. Eur J Appl Physiol Occup Physiol. 1992;65:37-42.
29. Robinson S, Edwards HT, Dill DB. New records in human power. Science. 1937;85:409-10.
30. Saltin B, Strange S. Maximal oxygen uptake: "old" and "new" arguments for a cardiovascular limitation. Med Sci Sports Exerc. 1992;24(1):30-7.
31. Smyth RJ, D'Urzo AD, Slutsky AS, Galko BM, Rebuck AS. Ear oximetry during combined hypoxia and exercise. J Appl Physiol. 1986;60:716-9.
32. Snell PG, Mitchell JH. The role of maximal oxygen uptake in exercise performance. Clin Chest Med. 1984;5:51-62.
33. Stenberg J, Ekblom B, Messin R. Hemodynamic response to work at simulated altitude, 4,000 m. J Appl Physiol. 1966;21:1589-94.
34. Stray-Gundersen J, Chapman RF, Levine BD. "Living high-training low" altitude training improves sea level performance in male and female elite runners. J Appl Physiol. 2001;91:1113-20.
35. Tenenbaum G. Theoretical and practical considerations in investigating motivation and discomfort during prolonged exercise. J Sports Med Phys Fitness. 1996;36:145-54.
36. Terrados N, Mizuno M, Andersen H. Reduction in maximal oxygen uptake at low altitudes: role of training status and lung function. Clin Physiol. 1985;5:75-9.
37. Turner PE, Raglin JS. Variability in precompetition anxiety and performance in college track and field athletes. Med Sci Sports Exerc. 1996;28(3):378-85.
38. Wagner PD. Determinants of maximal oxygen transport and utilization. Annu Rev Physiol. 1996;58:21-50.
39. Wehrlin JP, Hallén J. Linear decrease in V˙O2max and performance with increasing altitude in endurance athletes. Eur J Appl Physiol. 2006;96:404-12.
40. Young AJ, Cymerman A, Burse RL. The influence of cardiorespiratory fitness on the decrement in maximal aerobic power at high altitude. Eur J Appl Physiol Occup Physiol. 1985;54:12-5.


©2011The American College of Sports Medicine