Article Text

Download PDFPDF

Health risk for athletes at moderate altitude and normobaric hypoxia
  1. Kai Schommer1,
  2. Elmar Menold1,
  3. Andrew W Subudhi2,
  4. Peter Bärtsch1
  1. 1Department of Internal Medicine, Division of Sports Medicine, University Hospital Heidelberg, Germany
  2. 2Department of Biology and Altitude Research Center, University of Colorado, Colorado Springs and Anschutz Medical Campuses, Aurora, Aurora, USA
  1. Correspondence to Professor Peter Bärtsch, Department of Internal Medicine, Division of Sports Medicine, University Hospital, Im Neuenheimer Feld 410, 69120 Heidelberg, Germany; peter.bartsch{at}


  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

View Full Text

Statistics from

Potential health risks for athletes at moderate altitude

Health risks at high altitude include acute mountain sickness (AMS), high-altitude cerebral oedema (HACE) and high-altitude pulmonary oedema (HAPE). These illnesses can occur in young and healthy individuals within the first few days of exposure to high altitude. In addition, high altitude may aggravate pre-existing illnesses or lead to a first manifestation of an illness that was asymptomatic at low altitude and therefore unknown. This can predominantly be a threat to the older accompanying staff that may have undiagnosed cardiopulmonary disease. Our paper focuses on the health of the athletes and does not address the problems that may arise at high altitude from pre-existing disease for which the readers are referred to recently published reviews on high-altitude tolerance of patients with pre-existing disease.1–,4 This applies also to apparently healthy athletes with sickle cell trait who are heterozygous for the haemoglobin S gene and who are at risk for splenic infarctions even at moderate altitudes5 ,6 in addition to their increased risk for sudden death from exertional heat illness and rhabdomyolysis.

Since the risk of acute high altitude illnesses depends predominantly on the altitude, and the time of exposure7 besides rate of ascent and individual susceptibility,8 we focus in this paper on altitudes and types of exposure that are typical for competitions and training of high-level athletes, namely:

  • Competitions mostly at low altitudes and rarely at moderate altitudes between 2000 and 3000 m.

  • Classic altitude training that takes place over 2–3 weeks at altitudes of 2000–2500 m.

  • Living or sleeping for 8–20 h at altitude or in normobaric hypoxia corresponding to an altitude of 2500–3500 m while training at low altitude.

  • Exposures of a few minutes to several hours at normobaric or hypobaric hypoxia corresponding to altitudes of 4000–6000 m at rest or during training.

The above-described settings are not associated with a risk for the development of the potentially lethal lung or brain oedemas of high altitude since HACE usually occurs after several days of continuous exposure to altitudes above 4000 m and HAPE can develop after rapid ascent and a continuous stay of 2–3 days at altitudes above 3000–3500 m.9 Therefore, we focus in this paper on the risk of AMS during prolonged exposures at moderate altitudes or hypoxia that are typical for training and competitions of high-level athletes. Conditions associated with high and extreme altitude ranges are not considered and have been covered in previous reviews.4 ,7 ,9

AMS consists of non-specific symptoms, such as headache, loss of appetite or nausea, insomnia, dizziness and peripheral oedema, which usually occur with a latency of six to eight or more hours after exposure to moderate altitude or hypoxia. Usually, AMS is most prominent after the first night spent at a given altitude and it resolves spontaneously within 1–3 days if no further ascent occurs and exertion is avoided.7 At moderate altitudes AMS is usually mild, transient and does not progress to more severe illnesses such as cerebral or pulmonary oedema. Data obtained in tourists at moderate altitudes in the Rock Mountains show that mild AMS affects about 25% of the population at risk.10

It can be hypothesised that data obtained in tourists are not relevant for athletes who may have a higher risk for AMS since strenuous exercise may exacerbate AMS.11 Although we were asked to write a review we, therefore, present also unpublished data regarding AMS at moderate altitude in settings relevant for athletes, and discuss these results in the context of data available in the literature. Furthermore, we present unpublished data comparing AMS in hypobaric and normobaric hypoxia since many athletes are using normobaric hypoxia as part of their training plan and it has been hypothesised that normobaric hypoxia may cause less AMS than hypobaric hypoxia.12

Data on AMS obtained in settings relevant for athletes

Methods of assessing AMS

The severity of AMS is evaluated by clinical examination and is quantified by using the Lake Louise (LL) scoring protocol13 or the AMS-C score of the Environmental Symptom Questionnaire (ESQ).14 For the LL self-assessment-scoring protocol the subjects answer questions about the severity of headache, gastrointestinal symptoms, fatigue, light headedness or dizziness and insomnia. A score of 0–3 points (0 = no symptoms, 1 = mild symptoms, 2 = moderate symptoms and 3 = severe symptoms) is assigned to each item. The sum of all points yields the LL self-assessment score. Subjects are considered to have AMS if they have a LL score >2 points composed of headache and one additional symptom. If a medically trained person is on site the LL self-assessment score can be supplemented by adding a clinical score that gives 0–3 points for peripheral oedema (0–3 points) and 0–4 points for mental status (4 points for coma) and for ataxia (4 points for inability to stand).

The AMS-C score of the ESQ is composed of 11 items, which are graded from 0 (not present) to 5 (extremely severe). The factorial weight of each item is given in parentheses: light-headed (0.489), headache (0.465), dizziness (0.446), feeling faint (0.346), dim vision (0.501), off-coordination (0.519), feeling weak (0.387), sick to stomach (0.347), loss of appetite (0.413), feeling sick (0.692) and feeling hung-over (0.584). To obtain the AMS-C score, the sum of all item scores multiplied by the respective factorial weight is multiplied by 0.1926 to reduce the total score to a range 0–5. Individuals are considered to have AMS if they have an AMS-C score ≥0.70 points. The cut-off value of ≥0.70 for AMS-C identifies those who report feeling sick with a sensitivity of 60% and a specificity of 99%.14 An AMS-C score ≥0.70 is equivalent to an LL score >4.15

AMS with rapid exposure to 1950 m followed by exercise between 1950 and 3200 m

Ascent to moderate altitude followed rapidly by intense physical exercise is a typical setting for athletes in high-altitude training camps or with training or competitions in alpine skiing. Therefore, we report data obtained in 15 healthy, mostly well-trained members of the Division of Sports Medicine at the University Clinic of Heidelberg (age 44.2, female 8, male 7) who drove from Heidelberg to Adelboden (1350 m) by car in 5 h and ascended by cable car to Engstligenalp (1950 m) were they stayed 2 days for skiing and ski touring up to altitudes of 2600–3200 m. They filled out complete ESQ for the calculation of AMS-C in the two mornings preceding the travel and on the two mornings at 1950 m. Although these data were not obtained in high-level athletes, we consider them relevant for competitions and training in Alpine skiing because of a similar setting with regard to altitude exposure and exercise.

Because a definition of AMS based on the LL score identifies subjects with less severe disease, we tried to estimate LL scores based on the answers given in the ESQ. Thus, for headache, we used question 2 of the ESQ, for gastro-intestinal symptoms questions 24 and 52, for dizziness/light-headedness questions 1 and 4–6, for fatigue questions 56 and 57, and for sleep question 58. A score of 1 or 2 in the ESQ was rated as 1 in the LL score, a score 3 or 4 as 2, and a score of 5 as 3, respectively. When several items of the ESQ related to one symptom of the LL score, the highest score of any of these items was considered. Values are reported as means±SD. The scores were compared by a non-parametric analysis of variance (Friedman test) and the incidence of AMS by the exact Fischer test.

LL scores and the incidence of AMS based on this score were not significantly different between days at low and days at moderate altitude (table 1). Nevertheless, four subjects (27%) had AMS defined by the LL score >2 including headache after the first night and two (13%) had AMS after the second night spent at 1950 m. When the more rigorous criterion of AMS-C ≥ 0.7 is applied, none of the subjects had AMS.

Table 1

AMS scores after rapid ascent to 1950 m in Alpine skiers

Assessment of AMS during a high-altitude training camp at 2100 m

We were able to obtain data on AMS in a unique group of eight junior German world class swimmers who were not acclimatised to moderate altitude. They trained in the Sierra Nevada at 2300 m and lived for 21 days at 2100 m. The athletes filled out the complete ESQ in the morning of days 2–5, 12 and 21 of the training camp. Day 1 was the day of arrival at 2100 m. The AMS-C and LL scores were calculated and analysed as described in the paragraph two of the preceding section. Data are reported as mean values±SD.

LL scores and AMS-C during a training camp (table 2) were comparable to the values obtained in the ski tourists shown in table 1. There was no significant difference in each of the two scores between days at the training camp. During the first few days, at which AMS may occur, only one person had mild symptoms that fulfilled the criterion AMS by the LL score, while, based on the AMS-C score, only one subject fulfilled the criterion AMS on day 21, a time at which these symptoms cannot be attributed to AMS any more.

Table 2

AMS scores during a training camp at 2100-2300 m

Comparison of AMS in normobaric and hypobaric hypoxia

Many athletes are using normobaric hypoxia as part of their training plan. Since it has been hypothesised that normobaric hypoxia may cause less AMS than hypobaric hypoxia12 we report data on AMS from two studies that involved a 16-h exposure at an equal ambient pO2 in a normobaric or hypobaric environment. We obtained LL scores using the LL questionnaires and AMS-C scores from an abbreviated version of the ESQ,16 to which the question on sleep quality was added in 30 non-acclimatised subjects (age 26.3±5.8 years, 13 female and 17 male) after a 16-h exposure in a normobaric hypoxia room at an FIO2 0.12 resulting in an ambient pO2 of 90 mm Hg. This exposure took place at 100 m, where subjects entered the hypoxia room with an FIO2 of 0.12. We compared these data with the same scores obtained in 32 non-acclimatised subjects (age 32.0±9.3, 13 female and 19 male) after a stay of 16 h in the Margherita hut (4559 m, ambient pO2 ≈90 mm Hg). The latter subjects had participated in a previously reported study17 and they had all ascended to 4559 m within 24 h after transportation by cable car to 2900 m on foot, including an overnight stay at 3661 m. Although a hypoxia corresponding to an altitude of 4559 m is much higher than an altitude used for living or training by athletes, it is, however, comparable to the degree of hypoxic used for training in hypoxia while living in normoxia. Furthermore, we hypothesised if no significant difference between normobaric and hypobaric exposure occurred at 4500 m this should also be the case at lower altitudes.

Scores were calculated as mentioned above. Data are reported as mean values±SD. Overall scores and scores of each symptom were compared between groups by Mann Whitney U Rank Sum Test and incidence of AMS between groups by a χ2 test. Statistical significance was set at p < 0.05 (two-sided).

The incidence of AMS was not different between normobaric and hypobaric hypoxia with 60% vs 63%, respectively (p=0.95) when AMS was defined as LL score > 4 and 67% vs 63%, respectively (p=0.94) when AMS was defined as AMS-C ≥ 0.7. The incidence of AMS was 90% in normobaric and 75% (p = 0.23) in hypobaric hypoxia when AMS was defined as LL score > 2 (including headache). The self-assessment LL score was not significantly different between normobaric and hypobaric hypoxia with 6.73±1.16 and 6.50±1.85, respectively (p=0.41). There was also no significant difference for the total LL score: 7.07±1.49 and 7.70±2.49, respectively (p=0.38). Furthermore, the comparison of the pattern of the scores of individual symptoms was not significantly different between normobaric and hypobaric hypoxia as shown in figure 1 with the exception of peripheral oedema that only occurred in hypobaric hypoxia and that might be attributable to the exercise of ascending and/or hypobaria.

Figure 1

Mean values±SD of individual items of the Lake Louise score: 1, headache; 2, gastrointestinal; 3, fatigue/weakness; 4, dizziness; 5, sleep; total SA, total score for self-assessment; 7, mental status; 8, ataxia; 9, peripheral oedema; SA±C, total score for self-assessment and clinical findings. p-Values are given for comparison of individual items between normobaric and hypobaric exposure. All p values except for those stated in the figure were >0.20.

The AMS-C score was not significantly different between normobaric and hypobaric hypoxia with 1.49±0.67 and 1.63±0.84, respectively (p 0.91). Furthermore, the comparison of the pattern of the scores of individual symptoms including the question for the quality of sleep was not significantly different between normobaric and hypobaric hypoxia as shown in figure 2.

Figure 2

Mean values±SD of individual item scores of the acute mountain sickness (AMS)-C score. Numbers correspond to those given in the complete Environmental Symptom Questionnaire. 1, light-headed; 2, headache; 4, dizzy; 5, faint; 6, vision; 7, coordination off; 19, feel weak; 24, sick to stomach; 52, lost appetite; 53, feel sick; 54, feel hung-over. In addition, the score of the item could not sleep (question 58) and the total AMS-C score are shown. p > 0.20 for comparison of all individual items and the total AMS-C score between normobaric and hypobaric exposure.

AMS with prolonged continuous exposure at moderate altitude

Olympic Games are usually held considerably below the altitude of 2000 m. The highest altitude so far was 2260 m in 1968 in Mexico City for the summer games. With few exceptions, most of the Olympic villages in the winter games and the sites of competition (with the exception of Alpine skiing) are also below 2000 m. Classic high-altitude training is usually carried out between at altitudes 2000 and 2500 m. The data reported in this paper that were obtained in the ski tourists at 1950 m touring between 2600 and 3200 m and in elite athletes in a training camp at 2100 m with a training altitude of 2300 m demonstrate that symptoms of AMS, which are typically most prominent on day 2 or 3, are minor and not more severe or frequent than on the two control days at low altitude. They are also not different from non-specific symptoms that accompany intense physical activity during a training camp as shown by equally high scores on days 2–5 with days 12 and 21 when athletes are acclimatised and symptoms cannot be attributed to AMS more.

The prevalence of AMS defined as >2 symptoms in conference attendees travelling in 1 day by car to 2100 m was 25%.18 This is comparable to the 27% incidence of AMS using a similar definition in our ski tourists on the first morning at 1950 m. It should be noted, however, that no subject in the latter study had AMS when using the more rigorous definition of AMS-C ≥ 0.70. It is very likely that the mild symptoms of AMS are in part attributable to the stress of travelling and possibly also to a higher than normal intake of alcohol on such occasions. Interestingly, it was shown in children that the stress of travel and change of environment during a science camp at sea level increased symptoms associated with AMS, such that 21% of the children had ‘AMS’.19 This not-altitude-specific effect accounts for about 5% of AMS in adults.18

The percentage of athletes that fulfils the criterion for AMS is low, even when the cut-off score of an LL score of >2 including headache is applied. Only one athlete (day 21) was above the criterion score for AMS-C, which is more rigorous and was shown to identify subjects who consider themselves sick. These data fit well with the observation that the first symptoms attributable to hypoxia occur in resting individuals after an exposure of 3–9 h between an altitude of 2100 and 2400 m without affecting the occurrence of AMS defined by AMS-C ≥ 0.70.20 Our data suggest that physical exercise associated with a training camp of elite athletes at these altitudes has little impact on the prevalence of AMS and discomfort throughout the altitude exposure.

Although the number of subjects in the reported studies on physically active individuals at moderate altitude is small and we cannot exclude statistical type II error, our data fit well with previously published findings in less active subjects. Therefore, we conclude, that AMS is not a problem for the vast majority of athletes at altitudes between 2000 and 2500 m. Mild symptoms may occur during the first few days of exposure and are best prevented and treated by reducing activity during the first few days at a moderate altitude.4

AMS with sleeping in hypoxia

This form of exposure is present when athletes train at low altitude or near sea level and sleep or live at moderate altitude. Usually, altitude exposure particularly for sleeping is gradually or stepwise increased from 2500 to 3000 or even 3500 m to allow for acclimatisation and particularly for avoiding disturbance of sleep. The exposure to hypoxia often takes place in a normobaric environment.

It was recently suggested that hypobaric hypoxia is associated with more severe AMS compared with normobaric hypoxia.12 The comparison of the incidence, severity and the pattern of symptoms caused by normobaric and hypobaric hypoxia at the same ambient pO2 of 90 mm Hg (equivalent to an altitude of about 4500 m) provided in this paper (figure 1) shows no significant difference between the two modalities. It is, however, not clear as to how the protective factor of a slower ascent over 20 h on the mountain compares with the possibly enhancing effect of an exposure to normobaric hypoxia within seconds or the possibly enhancing effect of a moderate exercise of climbing.11 Nevertheless, the striking similarities of the symptom patterns, of the severity and of the incidence of AMS in normobaric and hypobaric hypoxia suggest that an exposure to an equal level of hypoxia over 16 h has the biggest impact and overrides the potentially modifying effects of exercise and speed of exposure.

On the basis of the data at real altitude we can expect that a daily hypobaric exposure over up 20 h beginning at an equivalent altitude of 2000–2500 m with a gradual increase to 3000–3500 m will not cause significant symptoms of AMS. This conclusion is extended to a comparable exposure in normobaric hypoxia, since we found no clinically relevant difference between an exposure in normobaric and hypobaric hypoxia. Our assumption is confirmed by a study that found no AMS in well-trained athletes (VO2max about 60 ml/kg/min) spending 18–24 nights over 11–16 h daily in normobaric hypoxia starting at a hypoxia equivalent to 2500 m and increasing it by 500 m every 6–12 days.21 There was only a minor disturbance of sleep throughout the study that could not be attributed to hypoxia since it also occurred in the control group, a finding that was confirmed in a further study.22

In summary, we conclude that sleeping or living in normobaric or hypobaric hypoxia up to 3500 m has no negative effects on the health of athletes when the altitude is gradually or stepwise increased above 2000–2500 m at a rate of about 100 m/day.

Health risks with short exposures to severe hypoxia

There are three major ways of applying short term hypoxia to athletes:

  1. Exposure of one to a few hours at rest, often with increasing altitude up to 5000 m with the intention to increase the endogenous erythropoietin production.23 ,24 Normobaric and hypobaric modalities of exposure are used for this purpose.

  2. Training sessions that are usually undertaken in normobaric hypoxia equivalent to altitudes of 4000–5000 m.25

  3. Repeated intermittent brief exposures over a few minutes at rest with severe normobaric hypoxia equivalent to altitudes of up to 6000 m.26 ,27

All these modalities use a level of hypoxia that could induce AMS, HACE or HAPE. However, it takes about 6–8 h for the development of AMS and HACE or HAPE occur only after continuous exposures of at least 2 days above threshold altitudes of 4000 and 3000–3500 m, respectively. Therefore, these brief hypoxic exposures used by athletes cannot induce acute high-altitude illnesses. When brief exposures of such levels of hypoxia are applied to young healthy athletes there is also no danger of inducing acute cardiac arrhythmias or ischaemia.28 This is, however, a concern, when such methods are applied to an older population in which silent pre-existing cardiac disease, predominantly coronary artery disease, may be present.


AMS is mild, transient and affects at the most 25% of a tourist population at risk at moderate altitudes at which athletes may compete or train. Data presented in this paper demonstrate that more intense physical activity associated with high-altitude training or mountaineering does not increase prevalence or severity of AMS at these altitudes. These conclusions can also be extended to the use of normobaric hypoxia, as data in this paper suggest that the severity of AMS is not significantly different between hypobaric and normobaric hypoxia at the same ambient pO2. Furthermore, high-altitude cerebral or pulmonary oedema do not occur at these altitudes and intermittent exposure to considerably higher altitudes (4000–6000 m) used by athletes for hypoxic training are too short to cause acute high-altitude illnesses. Even moderate altitude between 2000 and 3000 m can, however, exacerbate cardiovascular or pulmonary disease or lead to a first manifestation of undiagnosed illness in older people that may belong to the accompanying staff of athletes. Moderate altitudes may also lead to splenic infarctions in healthy athletes with sickle cell trait.

What this study adds

  • Those who are more prone to develop mild acute mountain sickness at moderate altitude should reduce physical during the first few days of acclimatisation. The risks associated with heterozygous haemoglobin S at moderate altitude need to be addressed by further studies.


We thank Mrs I Slater for the secretarial work and Dr B Friedmann-Bette for obtaining ESQ during a training camp of elite junior athletes.


View Abstract


  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.