Prevalence of Exercise-Induced Arterial Hypoxemia in Distance Runners at Sea Level


Medicine & Science in Sports & Exercise: May 2017 - Volume 49 - Issue 5 - p 948–954
doi: 10.1249/MSS.0000000000001193
Applied Sciences

Purpose: It has been reported that ~50% of endurance-trained men demonstrate exercise-induced arterial hypoxemia (EIAH) during heavy exercise. However, this often-cited prevalence rate comes from a single study using a cohort of 25 highly trained men who completed maximal cycle ergometry. As arterial oxyhemoglobin saturation (SpO2) during maximal exercise is reported to be significantly lower during treadmill versus cycle ergometry in the same subjects, we hypothesized that the prevalence of EIAH would be greater than previously reported (and commonly referenced) in a larger cohort of highly endurance-trained men during maximal treadmill running.

Methods: Data from 124 highly trained male distance runners (V˙O2max range = 60.3–84.7 mL·kg−1·min−1) were retrospectively examined from previously published studies completed in the Indiana University Human Performance Laboratory. Subjects completed a constant speed, progressive-grade treadmill exercise test to volitional exhaustion, and arterial oxyhemoglobin saturation (SaO2ear) in all subjects was estimated using the same oximeter (Hewlett Packard 47201A).

Results: Using similar inclusion criteria as previously published for highly trained (V˙O2max > 68 mL·kg−1·min−1) and for EIAH (SaO2ear ≤ 91%), 55 of 79 subjects (70%) exhibited exercise-induced arterial desaturation. Across all 124 subjects, 104 (84%) demonstrated at least moderate EIAH (SaO2ear ≤ 93%) during maximal treadmill exercise. SaO2ear was significantly yet weakly correlated with V˙E/V˙O2 (P < 0.01, r = 0.28) and V˙E/V˙CO2 (P < 0.001, r = 0.33) but not with V˙O2max.

Conclusion: These results indicate that the prevalence of EIAH in highly trained men during maximal treadmill exercise at sea level is greater compared with previously suggested data, with exercise mode perhaps playing a factor in the number of athletes who experience EIAH.

1HH Morris Human Performance Laboratory, Department of Kinesiology, Indiana University, Bloomington, IN; 2Department of Health and Kinesiology, Purdue University, West Lafayette, IN; and 3Department of Kinesiology, Kansas State University, Manhattan, KS

Address for correspondence: Robert F. Chapman, Ph.D., Department of Kinesiology, School of Public Health, Indiana University, 1025 E. 7th St., Bloomington, IN 47405; E-mail:

Submitted for publication October 2016.

Accepted for publication December 2016.

Article Outline

The belief that arterial oxyhemoglobin saturation (SaO2) in healthy individuals at sea level remains near resting levels during submaximal and maximal exercise (21,39) has undergone significant examination for the last 60 yr. In the late 1950s and early 1960s, reports emerged documenting substantial reductions in SaO2 in a select number of highly endurance-trained men exercising at very high oxygen uptakes (15,34). However, this notion went largely unverified until work by Dempsey et al. (6) clearly established the occurrence of pulmonary gas exchange limitations in a group of highly trained distance runners. Since then, numerous studies have been published documenting the presence of exercise-induced arterial hypoxemia (EIAH) among otherwise healthy subjects, including endurance-trained men (6,11,27–29), trained and untrained women (9,12,33), and older masters athletes (25,30). EIAH has also been reported to occur during various modes of exercise such as running (4,6,11–13), cycling (8,27,40), swimming (38), and rowing (21).

Although it is generally believed that exercise-induced oxyhemoglobin desaturation occurs mostly in individuals who reach high levels of maximal oxygen uptake (V˙O2max) (7,27), this phenomenon is not universally demonstrated in all highly endurance-trained individuals. Work by Powers et al. (27) reported that 13 (52%) of 25 men with V˙O2max > 68 mL·kg−1·min−1 demonstrated EIAH, defined in their study as arterial O2 saturation ≤91% (measured using pulse oximetry) during maximal exercise. This prevalence rate became a well-established standard within the literature and has been cited >100 times since it was published in 1988 (7,24,31,36). However, this 52% prevalence rate for EIAH has been characterized as being “at best a guess” by one review article (7) and is based on a single study using a cohort of 25 men who performed maximal cycle ergometry. Interestingly, arterial O2 saturation at V˙O2max is reported to be significantly lower during treadmill running versus cycle ergometry exercise in the same subjects (11,40), perhaps because of differences in muscle mass utilization, maximal cardiac output and maximal oxygen uptake between the two exercise modes (10,35). Decreased erythrocyte pulmonary transit time, secondary to the high cardiac output produced by highly trained endurance athletes with extreme metabolic demand, has been strongly linked to incomplete pulmonary gas exchange and EIAH (6,7,18,27).

Another major factor contributing to EIAH in highly trained athletes is inadequate hyperventilation during maximal exercise (7,13). Lower pulmonary ventilation and ventilatory equivalents for O2 and CO2 (V˙E/V˙O2 and V˙E/V˙CO2, respectively) demonstrated during running compared with cycling within the same subjects (11,32,40) likely indicate a more pronounced inadequate hyperventilatory response to running, thus a greater chance of developing some critical magnitude of arterial oxyhemoglobin desaturation during maximal running exercise. Therefore, it is reasonable to expect that the prevalence of EIAH would be higher during treadmill exercise, compared with cycle ergometry exercise, due to the higher cardiac outputs that can be achieved running versus cycling and/or due to the differences in ventilatory patterns between the exercise modalities (10,11,40).

The purpose of this study was to determine the prevalence of EIAH in a large cohort of highly endurance-trained men during treadmill exercise. Our primary hypothesis was that the overall prevalence rate of EIAH, defined using previously published criteria, in highly endurance-trained men during maximal treadmill exercise would be greater than the 52% previously reported during maximal cycle ergometry (27). In addition, we examined ventilatory parameters indicative of inadequate hyperventilation (V˙E/V˙O2 and V˙E/V˙CO2) and their correlation with degree of EIAH.

Back to Top | Article Outline


Back to Top | Article Outline


Data were retrospectively examined from research studies completed in the Indiana University Human Performance Laboratory (elevation 230 m) between 1993 and 2007, with the primary outcomes for several of these studies published elsewhere (3–5,11,13,39,40). Criteria for inclusion consisted of (a) healthy, nonsmoking male subjects between the ages of 18 and 40 yr; (b) completion of a maximal graded test to exhaustion while running on a treadmill; and (c) V˙O2max ≥ 60 mL·kg−1·min−1. In all cases, subjects gave written informed consent before testing, and all protocols and procedures for testing were approved by the Institutional Review Board of Indiana University.

Back to Top | Article Outline


With advancement in measurement techniques and increasing awareness and interest in pulmonary gas exchange limitations during exercise in recent decades, terminology concerning this topic has become more precise. Consequently, discrepancies in the literature exist regarding the use of terms such as hypoxemia and desaturation, causing some confusion and adding difficulty when comparing between studies completed in different eras. Currently, the term hypoxemia is properly used to describe a reduction in arterial O2 pressure when blood gases are measured directly, whereas desaturation refers to a reduction in arterial oxyhemoglobin saturation often estimated noninvasively by pulse oximetry. Many previous studies, several of which are referenced in this manuscript, have used the term EIAH with defining criteria based on some magnitude of arterial oxyhemoglobin saturation, and no blood gas values measured. To avoid confusion in reference to these previous works, we will use the term arterial hypoxemia (or EIAH) to refer to reductions in arterial oxyhemoglobin saturation below some criterion value, independent of the method of measurement.

Another discrepancy often encountered in the literature is the complimentary use of SaO2 and SpO2. Traditionally, SaO2 was used as a global abbreviation for arterial oxyhemoglobin saturation, regardless of the methods used to measure or estimate arterial saturation levels. However, it has now become common to distinguish between values for arterial oxyhemoglobin saturation that are measured using blood gases (SaO2) and estimates of this measure obtained via pulse oximetry (SpO2). Although we acknowledge that some studies referenced in this article have used direct arterial measurements to determine arterial oxyhemoglobin saturation, in this study, we have estimated the level of arterial oxyhemoglobin saturation using a noninvasive technique. Therefore, throughout this article, SpO2 will be used when referring to arterial oxyhemoglobin saturation data from other studies where pulse oximetry was used or where the measurement techniques are unclear. Because of the uniqueness of the oximeter used in our study, which is not precisely a pulse oximeter (see more details below), SaO2ear will be used when referring to our arterial oxyhemoglobin saturation results.

Back to Top | Article Outline

Graded exercise test protocol

Each subject completed a maximal graded exercise test on a motor-driven treadmill (Quinton, Bothell, WA). Although the protocol used was not standardized across all subjects, all graded exercise tests involved running at a constant velocity (9.7–16.0 km·h−1) with the slope of the treadmill increasing either 1% each minute or 2% every 2 min until volitional fatigue. Criteria for a valid assessment of V˙O2max included the following: 1) a heart rate (HR) of ±10% of age-predicted maximum (220 − age), 2) a respiratory exchange ratio of ≥1.10, and 3) a plateau (≤150 mL increase) in O2 consumption (V˙O2) with an increase in workload (20). If at least two of these three criteria were met, the highest V˙O2 recorded for a 60-s period was chosen as the subject's V˙O2max. The subject's pulmonary ventilation (V˙E) during the same period was used in the analysis of maximal ventilation (V˙Emax), and V˙E/V˙O2 and V˙E/V˙CO2 were also recorded during the corresponding 60-s epoch.

Ventilatory and metabolic variables were continuously measured and monitored during exercise using a computer-interfaced, open-circuit, indirect calorimetry system. The expired side of a two-way, large-bore non-rebreathing valve (Hans Rudolph 2700, Kansas City, MO) was connected to a 5-L mixing chamber. Fractional concentrations of O2 and CO2 were determined from either (a) a continuous sample of dried expired gas at a rate of 300 mL·min−1 using an Applied Electrochemistry S-3A oxygen analyzer and CD-3A CO2 analyzer (Ametek; Thermox Instruments, Pittsburgh, PA) or (b) mass spectrometry (Marquette RMS-100, Milwaukee, WI). The analyzers or mass spectrometer were calibrated with a gas of known concentration in the physiological range before and after each test. Ventilation was determined on the inspired side by (a) a turbine-based electronic flowmeter (Model VMM-2; Sensormedics, Anaheim, CA), (b) a dual-thermistor flow probe (Torrent 1200; Hector Engineering, Ellettsville, IN), or (c) a pneumotachometer (Hans Rudolph 3813, Kansas City, MO). Each of these devices was calibrated using a Tissot spirometer and pulsatile flow. The previously mentioned variables, as well as SpO2, were continuously measured and monitored with a data acquisition control system (WorkBench PC 2.0; Strawberry Tree, Sunnyvale, CA; or DASYLab 10.0, National Instruments, Norton, MA) sampling at 40 Hz.

Back to Top | Article Outline

SaO2ear measurement

For all subjects, arterial oxyhemoglobin saturation was estimated using the same ear oximeter (model 47201A; Hewlett-Packard [HP], Waltham, MA). The oximeter was calibrated using an internal protocol before each test according to the manufacturer's instructions. Before exercise, the pinna of the subject's right ear was cleaned with alcohol and vigorously rubbed for >30 s to increase blood flow. 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.

The “older” HP47201A 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 the calculation of SaO2ear values that are satisfactorily independent of factors such as movement of the earpiece on the ear, skin pigmentation, or the waveform of the pulse. Therefore, because this device does not depend on pulse detection and thus should not be considered a pulse oximeter, we will refer to measurements obtained using this oximeter as SaO2ear.

This model oximeter has been reported to be accurate across a wide range of SaO2ear, with arterial SaO2 values >75% underestimated by <2% (37). 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 arterial oxyhemoglobin saturation 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% SpO2, the two independent measures never differed by more than 1.9% and were tightly correlated (r = 0.99) (3). In addition, the ear oximeter output was verified versus blood samples analyzed with an ABL3000 blood gas analyzer (Radiometer) with similar results. For SaO2ear ranging from 65% to 98%, the ear oximeter and the 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% (8). However, despite the unique nature of this oximeter and our validation work, conflicting evidence exists on the validity of pulse oximeters during exercise. Some studies report an underestimation of SpO2, which can result in an overestimation of the prevalence of EIAH (2,26), whereas other studies suggest that those devices are valid predictors of SaO2 over a wide range of arterial oxyhemoglobin saturation even during maximal exercise (22,37). In addition, large variations between different devices exists (37), making it difficult to compare results of different studies, particularly between the oximeter used in our study and the modern pulse oximeters that are not as accurate and sensitive as the HP47201A oximeter.

Back to Top | Article Outline

Data analysis

Arterial desaturation was defined using two set of criteria: 1) using the definitions published by Dempsey and Wagner (7) of non-EIAH (SaO2ear > 95%), mild EIAH (SaO2ear of 93%–94.9%), moderate EIAH (SaO2ear of 88%–92.9%), and severe EIAH (SaO2ear < 88%) as defined elsewhere and 2) as SaO2ear ≤ 91% (EIAHPowers) to match the definition used in the study of Powers et al. (27) examining the prevalence of EIAH. To mirror the Powers et al. (27) subject grouping of “elite highly trained endurance athletes,” only subjects meeting the V˙O2max criteria of V˙O2max > 68 mL·kg−1·min−1 were used in this analysis (i.e., using SaO2ear ≤ 91% as a cutoff for EIAHPowers). To test for differences in the dependent variables (V˙O2max, V˙Emax, V˙E/V˙O2 and V˙E/V˙CO2) between the EIAH groups, MANOVA were performed, and where appropriate a Tukey's post hoc test was used. Correlations between SaO2ear and metabolic variables (V˙O2max, V˙Emax, V˙E/V˙O2 and V˙E/V˙CO2) were evaluated using Pearson correlations. Values presented are mean ± SD unless otherwise stated. The alpha for all comparisons was set at P ≤ 0.05.

Back to Top | Article Outline


A total of 124 subjects met the inclusion criteria and were included in the statistical analyses. Subject characteristics can be found in Table 1.

Back to Top | Article Outline

EIAH criteria as defined by Dempsey and Wagner (7)

Metabolic and ventilatory data during maximal exercise utilizing four levels of EIAH criteria are listed in Table 2. Of the 124 subjects, 104 met the criteria of moderate (n = 71, 57%) or severe (n = 33, 27%) EIAH. While V˙O2max, V˙Emax, and V˙E/V˙O2 were similar between EIAH levels, there was a significant main effect of EIAH level for V˙E/V˙CO2 (F = 5.11, P < 0.01). Specifically, V˙E/V˙CO2 at maximal exercise was significantly lower in the moderate and severe EIAH groups compared with the mild-EIAH group (P < 0.01 and P < 0.05, respectively).

Back to Top | Article Outline

EIAH criteria as defined by Powers et al. (27)

Metabolic and ventilatory data during maximal exercise using the subject grouping and EIAH criteria of Powers et al. (27) can be found in Table 3. Of the 79 subjects with V˙O2max > 68 mL·kg−1·min−1, 55 (or 70%) demonstrated SaO2ear ≤ 91% during maximal exercise. No significant differences between EIAHPowers and non-EIAHPowers groups were found in V˙O2max or V˙Emax. However, the EIAHPowers group demonstrated significantly lower values for V˙E/V˙O2 (P < 0.01) and V˙E/V˙CO2 (P < 0.001) during maximal exercise.

Using data from the full cohort of 124 subjects, SaO2ear was significantly yet weakly correlated with V˙E/V˙O2 (P < 0.01, r = 0.28) and V˙E/V˙CO2 (P < 0.001, r = 0.33) (Fig. 1); however, no correlations were found between SaO2ear and either V˙O2max (Fig. 2) or V˙Emax.

Back to Top | Article Outline


The primary finding of this study is that the prevalence of EIAH during maximal treadmill exercise is as high as 70% when using the same criteria for EIAH (SaO2ear ≤ 91%) and fitness level (highly trained endurance athletes, V˙O2max > 68 mL·kg−1·min−1) as suggested by Powers et al. (27). Furthermore, when EIAH was defined using the criteria of Dempsey and Wagner (7), 84% (n = 104) of highly trained athletes experienced SaO2ear values <93% (moderate and severe EIAH), and this level of arterial desaturation was associated with lower ventilatory equivalents for CO2, indicating that an inadequate hyperventilatory response may be associated with the development of EIAH.

In determining the prevalence of EIAH during maximal treadmill exercise within an endurance-trained athletic population, the key advantages of this study are the large sample size (n = 124), all using the same (and arguably most reliable) oximeter (HP47201A) to estimate SaO2ear. Because of practical difficulties with the measurement of arterial gases, and/or using highly trained/elite athletes as subjects, many studies compromise on the method with which SaO2 is assessed, the study's sample size, or the training status of the subjects. The most commonly cited prevalence rate of EIAH in healthy male endurance athletes is 52%, as reported by Powers et al. (27) nearly 30 yr ago in a study that included 25 highly trained cyclists. With a cohort of highly endurance-trained athletes that is nearly five times larger, we have demonstrated that the prevalence of EIAH during running is not comparable with that previously reported for cycling, and as hypothesized, EIAH is much more prevalent during maximal treadmill running than cycle ergometry. This finding is consistent with others who reported lower values of SpO2 (11,40) and partial pressure of arterial O2 (PaO2) (6,32) during maximal running compared with cycling within the same subjects.

Evidently, EIAH prevalence in young, elite endurance male athletes depends on exercise mode. However, this prevalence rate is likely even higher when considering other factors such as sex and age. Specifically, females and older masters athletes are expected to experience higher rates of EIAH than males and younger athletes, respectively (9,12,25,30,33). We therefore believe that a prevalence rate of 52% should not be used collectively to characterize EIAH in all populations and for every exercise mode; rather, each group of athletes and mode of exercise should be considered separately. This will also help to identify and characterize the mechanisms of EIAH in each group, as the underlying causes of EIAH are likely different depending on the population and mode of exercise (9,17,25,40). In conclusion, it is important to realize that the prevalence of EIAH is not a set percentage, applicable to all, and future studies should determine the occurrence of EIAH in each subpopulation.

Back to Top | Article Outline

Causes of EIAH

The higher prevalence of EIAH during maximal treadmill running compared with cycling can be partially explained by the differences in hemodynamic and ventilatory patterns/responses between the exercise modes (7,19,31). The metabolic demand in highly trained athletes during maximal running requires a very large cardiac output that is thought to reduce erythrocyte pulmonary transit time and lead to incomplete pulmonary gas exchange (10,14,18). As a result, the widening of the alveolar–arterial (A-a) O2 difference and the existence of EIAH are more likely to occur during this mode of exercise in highly trained athletes (17,19). Furthermore, because running requires activating a larger proportion of total muscle mass than cycling (1), a lower mixed venous PO2 also contributes to a reduction in PaO2 and therefore to the widening of (A-a) O2 difference and a greater degree of arterial O2 desaturation (7,17). Another potential consequence of such a high cardiac output and the associated increase in pulmonary blood flow is further enlargement of the (A-a) O2 difference because of the development of interstitial pulmonary edema and diffusion limitations (7,16).

While V˙O2max and maximal cardiac output are higher during running than cycling in the same subjects, minute ventilation is either similar (40) between exercise modes or lower (11,17,32) during running. Accordingly, for any given level of V˙O2 or V˙CO2, ventilation is also lower during running compared with cycling (11,32,40). Interestingly, SpO2 has also been reported to be lower during maximal treadmill running than cycling in the same subjects (11,40), suggesting a possible cause and effect relationship between ventilatory drive and EIAH. Our finding that V˙E/V˙O2 and V˙E/V˙CO2 were lower at V˙O2max in the EIAHPowers group (SaO2ear < 91%) compared with non-EIAHPowers supports the notion that EIAH could be partially caused by an inadequate hyperventilatory response. Further strengthening this argument is the finding that PaCO2 is higher, PaO2 lower, and the (A-a) O2 difference larger during high-intensity running compared with cycling (17,32).

Although several studies support the role of inadequate hyperventilation in the development of EIAH (6,13,23), this topic remains equivocal in the literature (29,40). Harms and Stager (13) suggested that up to 50% of the variability in SpO2 is due to lower V˙E/V˙O2 and V˙E/V˙CO2 in hypoxemic compared with normoxemic group of healthy men. One factor that is thought to play a role in the occurrence of inadequate hyperventilation and the subsequent development of EIAH is low chemoresponsiveness, as demonstrated by a blunted hypoxic and hypercapnic drive in those who experience more severe arterial hypoxemia (13). In addition, mechanical constraints of the respiratory muscles (i.e., expiratory flow limitation [EFL]) and respiratory muscle fatigue should also be considered as factors contributing to the failure to adequately increase ventilation during strenuous exercise in athletes (6,7,16,23). Putting those ideas together, a significant relationship between the hypoxic ventilatory response (HVR) and SpO2 has been demonstrated only in athletes without EFL (8). This finding indicates that the presence of EFL “masks” the importance of HVR in the development of EIAH, and therefore EFL should be considered (i.e., controlled for) when investigating the causes of EIAH. Furthermore, this suggests that the prevalence of EIAH and the mechanisms contributing to its development are specific not only to sex, age, fitness level, and mode of exercise but also to EFL. Unfortunately, we did not assess EFL and HVR in the full cohort of 124 subjects, making it impossible to conclude whether it is low chemoresponsiveness (i.e., HVR) or mechanical constraints (i.e., EFL) that cause an inadequate hyperventilatory response during maximal exercise. In addition, which of those factors ultimately contributes more to the development of EIAH (as suggested by Derchak et al. [8]) cannot be determined from our data.

If highly trained athletes are believed to be more susceptible to experience EIAH, it would therefore be reasonable to expect that high V˙O2max values would be related to lower SaO2 (or SpO2) levels. Indeed, when examining a large range of fitness levels, a correlation between those variables does exist (12,28). However, this relationship is highly variable (7,12,16), and not all studies, including our own, have been able to demonstrate a correlation between V˙O2max and SpO2 or SaO2ear (Fig. 2). One possible reason for this discrepancy is that most studies include a specific population, such as endurance-trained individuals, with a small range of V˙O2max values, and therefore it is harder to detect trends and correlations. For example, despite a significant correlation between V˙O2max and either PaO2 or (A-a) O2 difference over a wide range of V˙O2max values, a review by Hopkins et al. (16) demonstrated a variation of nearly 60 mm Hg in PaO2 and a fivefold difference in (A-a) O2 difference in subjects with V˙O2max between 70 and 80 mL·kg−1·min−1.

Back to Top | Article Outline


With a large sample size of 124 athletes, we have demonstrated that 84% of highly trained men experience moderate or severe EIAH (SaO2ear < 93%) during maximal treadmill exercise. Moreover, when implementing more strict inclusion criteria for fitness level (V˙O2max > 68 mL·kg−1·min−1) and for EIAH (SaO2ear ≤ 91%) as published by others (27), 70% of subjects experienced EIAH—greater than the previously reported and often-cited prevalence rate. Exercise mode appears to be a key factor in the high number of athletes who experience EIAH during maximal running exercise. However, it should be emphasized that other factors such as sex and age, and perhaps the presence of EFL and/or low chemoresponsiveness, also appear to influence the prevalence of EIAH, and therefore, a set percentage should not be used collectively to characterize EIAH. Rather, future studies should identify EIAH prevalence and the specific mechanisms leading to its occurrence in each subpopulation and mode of exercise separately.

The authors disclose that no funding was received for this work. No conflict of interest is declared. The results of the present study do not constitute endorsement by the American College of Sports Medicine. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.

Back to Top | Article Outline


1. Bijker K, De Groot G, Hollander A. Differences in leg muscle activity during running and cycling in humans. Eur J Appl Physiol. 2002;87(6):556–61.
2. Brown DD, Knowlton RG, Sanjabi PB, et al. Re-examination of the incidence of exercise-induced hypoxaemia in highly trained subjects. Br J Sports Med. 1993;27(3):167–70.
3. Chapman RF, Emery M, Stager J. Degree of arterial desaturation in normoxia influences V˙O2max decline in mild hypoxia. Med Sci Sports Exerc. 1999;31(5):658–63.
4. Chapman RF, Stager JM. Caffeine stimulates ventilation in athletes with exercise-induced hypoxemia. Med Sci Sports Exerc. 2008;40(6):1080–6.
5. Coyle MA, Stager JM. Exercise-induced hypoxemia is ameliorated by nedocromil sodium and diphenhydramine HCL. Med Sci Sports Exerc. 2001;33(5):S58.
6. Dempsey JA, Hanson P, Henderson K. Exercise-induced arterial hypoxaemia in healthy human subjects at sea level. J Physiol. 1984;355:161.
7. Dempsey JA, Wagner PD. Exercise-induced arterial hypoxemia. J Appl Physiol. 1999;87(6):1997–2006.
8. Derchak PA, Stager JM, Tanner DA, et al. Expiratory flow limitation confounds ventilatory response during exercise in athletes. Med Sci Sports Exerc. 2000;32(11):1873–9.
9. Dominelli PB, Foster GE, Dominelli GS, et al. Exercise-induced arterial hypoxaemia and the mechanics of breathing in healthy young women. J Physiol. 2013;591(12):3017–34.
10. Faulkner JA, Roberts DE, Elk RL, et al. Cardiovascular responses to submaximum and maximum effort cycling and running. J Appl Physiol. 1971;30(4):457–61.
11. Gavin TP, Stager JM. The effect of exercise modality on exercise-induced hypoxemia. Respir Physiol. 1999;115(3):317–23.
12. Harms CA, McClaran SR, Nickele GA, et al. Exercise-induced arterial hypoxaemia in healthy young women. J Physiol. 1998;507(Pt 2):619–28.
13. Harms CA, Stager JM. Low chemoresponsiveness and inadequate hyperventilation contribute to exercise-induced hypoxemia. J Appl Physiol (1985). 1995;79(2):575–80.
14. Hermansen L, Ekblom B, Saltin B. Cardiac output during submaximal and maximal treadmill and bicycle exercise. J Appl Physiol. 1970;29(1):82–6.
15. Holmgren A, Linderhold H. Oxygen and carbon dioxide tensions of arterial blood during heavy and exhaustive exercise. Acta Physiol Scand. 1958;44(3–4):203–15.
16. Hopkins SR. Exercise induced arterial hypoxemia: the role of ventilation-perfusion inequality and pulmonary diffusion limitation. In: Roach RC, Wagner PD, Hackett PH, editors. Hypoxia and Exercise. Springer; 2006. pp. 17–30.
17. Hopkins SR, Barker RC, Brutsaert TD, et al. Pulmonary gas exchange during exercise in women: effects of exercise type and work increment. J Appl Physiol (1985). 2000;89(2):721–30.
18. Hopkins SR, Belzberg AS, Wiggs BR, et al. Pulmonary transit time and diffusion limitation during heavy exercise in athletes. Respir Physiol. 1996;103(1):67–73.
19. Hopkins SR, Harms CA. Gender and pulmonary gas exchange during exercise. Exerc Sport Sci Rev. 2004;32(2):50–6.
20. Howley ET, Bassett DR, Welch HG. Criteria for maximal oxygen uptake: review and commentary. Med Sci Sports Exerc. 1995;27(9):1292–301.
21. Kyparos A, Riganas C, Nikolaidis MG, et al. The effect of exercise-induced hypoxemia on blood redox status in well-trained rowers. Eur J Appl Physiol. 2012;112(6):2073–83.
22. Martin D, Powers S, Cicale M, et al. Validity of pulse oximetry during exercise in elite endurance athletes. J Appl Physiol (1985). 1992;72(2):455–8.
23. McClaran SR, Harms CA, Pegelow DF, et al. Smaller lungs in women affect exercise hyperpnea. J Appl Physiol (1985). 1998;84(6):1872–81.
24. McKenzie DC. Respiratory physiology: adaptations to high-level exercise. Br J Sports Med. 2012;46:381–4.
25. Mucci P, Prioux J, Hayot M, et al. Ventilation response to CO2 and exercise-induced hypoxaemia in master athletes. Eur J Appl Physiol Occup Physiol. 1998;77(4):343–51.
26. Norton L, Squires B, Craig N, et al. Accuracy of pulse oximetry during exercise stress testing. Int J Sports Med. 1992;13(7):523–7.
27. Powers SK, Dodd S, Lawler J, et al. Incidence of exercise induced hypoxemia in elite endurance athletes at sea level. Eur J Appl Physiol Occup Physiol. 1988;58(3):298–302.
28. Powers SK, Lawler J, Dempsey JA, et al. Effects of incomplete pulmonary gas exchange on V˙O2 max. J Appl Physiol (1985). 1989;66(6):2491–5.
29. Powers SK, Martin D, Cicale M, et al. Exercise-induced hypoxemia in athletes: role of inadequate hyperventilation. Eur J Appl Physiol Occup Physiol. 1992;65(1):37–42.
30. Préfaut C, Anselme F, Caillaud C, et al. Exercise-induced hypoxemia in older athletes. J Appl Physiol (1985). 1994;76(1):120–6.
31. Prefaut C, Durand F, Mucci P, et al. Exercise-induced arterial hypoxaemia in athletes: a review. Sports Med. 2000;30(1):47–61.
32. Rice AJ, Scroop GC, Thornton AT, et al. Arterial hypoxaemia in endurance athletes is greater during running than cycling. Respir Physiol. 2000;123(3):235–46.
33. Richards JC, McKenzie DC, Warburton D, et al. Prevalence of exercise-induced arterial hypoxemia in healthy women. Med Sci Sports Exerc. 2004;36(9):1514–21.
34. Rowell LB, Taylor HL, Wang Y, et al. Saturation of arterial blood with oxygen during maximal exercise. J Appl Physiol. 1964;19(2):284–6.
35. Saltin B, Astrand P-O. Maximal oxygen uptake in athletes. J Appl Physiol. 1967;23(3):353–8.
36. Sheel AW, Romer LM. Ventilation and respiratory mechanics. Compr Physiol. 2012;2:1093–142.
37. Smyth RJ, D'Urzo AD, Slutsky AS, et al. Ear oximetry during combined hypoxia and exercise. J Appl Physiol (1985). 1986;60(2):716–9.
38. Spanoudaki S, Maridaki M, Myrianthefs P, et al. Exercise induced arterial hypoxemia in swimmers. J Sports Med Phys Fitness. 2004;44(4):342.
39. 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 (1985). 2001;91(3):1113–20.
40. Tanner DA, Duke JW, Stager JM. Ventilatory patterns differ between maximal running and cycling. Respir Physiol Neurobiol. 2014;191:9–16.


© 2017 American College of Sports Medicine