Is altitude training a waste of time and money?
The effects of training and, more recently, sleeping at high altitude on athletic performance have been studied in the West for more than 30 years. During that time, these practices have become an almost essential aspect of the preparation of world-class competitors. Yet the evidence base supporting a beneficial effect altitude exposure for sea-level performance remains flimsy at best.
A telling analysis of the benefits of altitude exposure for sports performance was undertaken by François Peronnet and published in a letter the editor of the International Journal of Sports Medicine in 1994(1). He analysed the men’s running speeds corresponding to the world record and 10 best performances per year over 1,500m, 5,000m and 10,000m from 1956 to 1991 Taking 1968 as the watershed marking the interface between the pre- and post-altitude training eras, he reasoned that if altitude training had made a positive contribution performance, the rate of increase of running speeds should be steeper after this point.
The findings were striking. In 1968 the cream the world’s distance running talent trained altitude in preparation for the Mexico City Olympics; despite this, no new world records were set in 1968 or the next four years. In fact, this was the longest period between 1956 and 1991 without a new world record; and furthermore, the rate of increase in world record and annual best performances was slower after 1968 than before. All in all the data suggest that, far from improving endurance performance, altitude training may even have exerted detrimental effect.
Peronnet’s analysis ceased in 1991, and the early 1990s marked another watershed in the altitude debate, which may place a different complexion on the outcome. But before exploring this it would be useful to examine the traditional rationale for high-altitude exposure This rests primarily on the haematological (blood) adaptations that occur in humans to a hypoxic (low oxygen) environment.
Ascent to high altitude is accompanied by a progressive fall in barometric pressure and an accompanying fall in the partial pressure of oxygen. The resulting decrease in arterial oxygen saturation (hypoxaemia) triggers a cascade of physiological disturbances that ultimately result an increase in the production of red blood cells (RBCs), a process known as polycythaemia. The production of RBCs helps to improve the oxygen-carrying capacity of the blood, and hence maximal oxygen uptake (VO2max).
One of the essential precursors to the increase RBC formation is the release of erythropoietic factor (EPO) by the kidney, hence the (illegal) practice of injecting EPO to boost RBC production. While the benefits of this artificially-induced increase in RBC concentration are long established and well documented(2,3), the same is not true for altitude exposure, as I will explain. Early studies of the effects of altitude training focused on the RBC concentration changes and VO2max as their physiological outcome parameters.
VO2max and ‘detraining’ at altitude
As mentioned, high altitude is accompanied by a decrease in the partial pressure of oxygen which, turn, leads to a reduction in the driving pressure for oxygen transport and a corresponding fall in VO2max. The magnitude of this decline is around 5-7% per 1,000m(4). An increase in altitude of as little as 600m has been shown to decrease the performance of cyclists in five-minute cycle power test by 5.9%(5).
Thus, it is easy to see that one confounding influence on the outcome of altitude training is the progressive decline in VO2max,which compromises training intensity, leading to ‘detraining’ at altitude. In the early 1990s, the ‘live-high-train-low’ (LHTL) model was developed to overcome this effect(6). With LHTL, athletes sleep simulated or real altitudes of around 2,500m but train at sea level. So what does the data from the past decade of LHTL tell us about the benefits of altitude exposure for sea-level performance?
The LHTL model was originally predicated on two assumptions:
- That any benefit of altitude exposure was due to an increase in RBC concentration;
- That sleeping at simulated altitude was enough to stimulate RBC production.
Studies from the originators of LHTL appear to support the notion that it boosts RBC production, VO2max and running time trial performance. Levine and Stray-Gundersen went on to compare the effects of LHTL (2,500m and 1,250m), live-high-train-high (LHTH: 2,500m) and live-low-train-low (LLTL: 150m) over a fourweek intervention (7).
The LHTL and LHTH groups both showed increases in VO2max (5%) that were ‘related’ to the increase in their RBC volume (9%). However, only 14% of the variation in VO2max could be explained by the change in RBC volume. Performance in a 5,000m running time trial improved only in the LHTL group, but again less than half (42%) of the variation in performance could be explained by the change in VO2max.
Altitude simulation with hypoxic tents
In this study, altitude exposure was achieved by physically travelling up a mountain. More recently, the LHTL methodology has been implemented with the aid of so-called altitude houses, or hypoxic sleeping tents. In both, the athlete sleeps at a simulated altitude of around 2,500m.
In their 2001 review, Hahn et al retrospectively analysed data from a series of six studies conducted in the Australian Institute of Sport altitude house in Canberra(8). In four of the studies they examined the physiological adaptations stimulated by the LHTL methodology. They observed an increase in serum EPO of 80% after the first one to five nights of exposure, but this declined to a level that was not significantly different from baseline or the control situation by the end of the intervention. Similarly, there was no significant increase in either reticulocyte formation (immature RBCs) or RBC mass, and VO2max tended to decrease by comparison with baseline.
The only positive outcome across the four studies was that after a period of up to 23 days of LHTL there was a trend towards improved performance in exercise tasks lasting around four minutes. The researchers concluded that any benefits to performance of LHTL were unlikely to be due to an increase in RBC mass or VO2max.
It is most likely that differences in the outcome of studies focusing on RBC mass and VO2max are due to methodological factors. For example, Levine and Stray-Gundersen(7) noted an increase in RBC volume, but this was measured using a dye-injection technique that is now thought to overestimate RBC volume. In later research, using an isotopic labelling technique that is widely acknowledged to be the most reliable method of estimating RBC volume, a 13-day exposure to an altitude of 4,300m resulted in no change(9).
Where changes in VO2max are concerned, an analysis of 17 studies examining the effect of hypoxia on sea-level VO2max revealed a mean effect of +0.3% for hypoxia-exposed subjects, compared with -0.4% for control subjects(9). The researchers concluded that there was very little evidence for an effect of hypoxic exposure on VO2max and that differences between studies were probably due to biological variability and the random error of measurement.
Thus, if exposure to hypoxia does lead to performance improvements, it seems unlikely that these are due to improvements in the oxygen transport system. However, recent evidence is pointing to an unexpected benefit of LHTL. A number of studies have demonstrated a lower oxygen cost of steady-state cycling(11,12,13) and running (14,15) after exposure to hypoxia.
For example, in their study of 22 elite distance runners, Saunders et al compared the influence of a 20-day programme of LHTL, live-moderatetrain- moderate (LMTM: 1,500-2,000m) and LLTL (600m) (15). The oxygen cost of running at three submaximal speeds was on average 3.3% lower after LHTL than after the other two interventions. There were no significant differences between or within groups for minute ventilation, heart rate, respiratory exchange ratio (R), or haemoglobin mass. Since there was also no difference in lactate concentration in the LHTL group, the lower oxygen cost of running could not be explained by an increase in anaerobic metabolism.
Similarly, the absence of a change in R failed to lend support to a proposed mechanism for improved mechanical efficiency following LHTL, ie increased utilisation of carbohydrate. The researchers were unable to offer any alternative explanation for their observations, but this was a carefully conducted study suggesting that improvement in mechanical efficiency in response to altitude exposure justifies further investigation. In high-performance athletes, variations in race performance are often attributable solely to differences in their mechanical efficiency, especially in running (16).
To date, only one study has examined the influence of LHTL on ‘anaerobic’ performance. Nummella and Rusko observed an improvement in 400m running performance (0.8%) after a 10- day LHTL intervention in eight athletes, by comparison with controls who undertook identical training but remained at sea level throughout (17). The performance improvement was accompanied by increases in running speed at a range of lactate concentrations, as well as a reduction in perceived exertion.
The researchers speculated that the improved 400m performance might have been due to an increase in muscle-buffering capacity. This suggestion is supported by the observations of Gore et al, who observed an increase in the muscle-buffering capacity of six cyclists/triathletes after a 23-night LHTL intervention (11). However, in another study that same group detected no change in muscle-buffering capacity in 29 cyclists/triathletes following a 20-night programme of LHTL (18). The researchers suggested that the discrepancy might be due to the slightly higher simulated altitude used in their first study (3,000 v 2,650m).
Interestingly, the second study did identify a significant lowering of blood lactate concentration during exercise at 85% VO2max. However, this could not be ascribed to a change in the abundance of lactate transport proteins (monocarboxylate transporters MCT1 and MCT4), which remained unchanged. The researchers suggested that these transporters were already up-regulated by the subjects’ ordinary training and, further, that the overall pattern of response of the lactate system was consistent with a muscle cellular adaptation that reduced the rate of lactate production and facilitated oxidative energy provision.
Intermittent hypoxic training
Thus, there is some preliminary evidence that LHTL may also be beneficial for athletes whose competitive events rely heavily upon anaerobic metabolism.
An alternative method of utilising hypoxia to improve exercise performance is to train high and live low (THLL), also known as intermittent hypoxic training (IHT). This methodology has a long history in the former Soviet Union, going back as far as the 1930s(19). The Soviets were interested in the effects of acute exposure to hypoxic environments on early aviators who flew in open cockpits. In its modern guise, IHT is based upon the rationale that exercising in hypoxia enhances muscle adaptations to training.
In a meticulous review of the research output of scientists working in the former Soviet Union, Serebrovskaya recently provided access to a wealth of data that was not previously accessible to Western scientists(19). Of particular interest in the context of athletic performance was data on local muscle adaptations in response to IHT. For example, Serebrovskaya suggests that the Soviet data provides evidence of a range of mitochondrial adaptations that increase the efficiency of oxygen utilisation in the production of ATP (adenosine triphosphate – the body’s universal energy donor).
The western literature examining the effects of training in moderate hypoxia (IHT) is far less abundant than that examining the LHTL methodology. An early study used an elegant unilateral exercise model, in which 10 subjects exercised one leg while breathing room air and the other while breathing a gas mixture containing 13.5% oxygen (about 3,250m) for 30 minutes at a time, three days per week over an eight-week period(20).
They observed an increase in single leg VO2max and oxidative enzyme activities in both legs, but there was a greater increase in the activity of the mitochondrial aerobic enzyme citrate synthase in the hypoxically trained leg. There were also trends for increased activity of succinate dehydrogenase (another aerobic enzyme) and phosphofructokinase (an important regulator enzyme for glycolysis) in the hypoxically trained leg. Increased activity of these enzymes are normally associated with improved aerobic and anaerobic capacity following training. The greater improvements in the hypoxically-trained leg suggest that the normal training response may be enhanced by hypoxia.
Contradictions in the literature
The finding of an increase in oxidative enzyme activity is unique to training in hypoxia, since LHTL has not been shown to induce such changes. However, a subsequent study failed to replicate the finding of a greater effect of training in hypoxia on citrate synthase activity, although this may have been due in part to the use of a lower simulated altitude (2,500m)(21).
Changes in anaerobic performance were noted in a later study involving eight triathletes who trained under hypoxic conditions (2,500m) for two hours per day for 10 days, and a matched control group undertaking the same training at sea level(22). The hypoxically-trained athletes showed improvements in performance during a Wingate Anaerobic test (mean power, peak power and time to peak power) and VO2max that were not observed in the control group. In slight contrast, a study on non-endurance-trained subjects observed greater improvements in VO2max in hypoxically-trained subjects, but only when measured under hypoxic conditions(23).
Once again, the picture for IHT is clouded by contradictions within the literature. For example, the the most enthusiastic proponents of LHTL(7,24) have also published a study showing no additional benefit of IHT above normoxic training(25). This study compared the effects of IHT and normoxic training on high intensity performance in 16 swimmers and noted no differences in time trial performance over 100m or 400m between the two groups.
The discrepancy between the results of this study and those reported in the previous paragraph is probably due to differences in the magnitude and duration of the IHT stimulus and outcome measures between studies. In the swimming study, subjects trained for only eight weeks, at a lower simulated altitude (2,500m) and with less frequent (three times per week) and much shorter bouts of hypoxic exposure (around 12.5 min of high-intensity exercise and 10 min of rest between repetitions and sets). Previous studies with positive outcomes from IHT have typically trained subjects in hypoxia for periods of at least 30 minutes, five days per week(22,23).
To date, the theoretical rationale favouring IHT is probably stronger than the evidence from studies directly assessing its influence on physical performance. There appears to be some evidence of adaptations at muscle level, including increases in the activities of oxidative enzymes, mitochondrial volume and capillary length. This is supported by evidence of an increase in the gene expression of a range of factors with the potential to influence muscle metabolism and, possibly, performance(26).
This latter study compared the effects of training in hypoxia (3,850m) and normoxia at two training intensities for 30 minutes at a time, five days per week over a period of six weeks. The most potent stimulus was high intensity hypoxic training, which resulted in sub-cellular changes that could theoretically improve the muscles’ performance. In fact, though, there was no evidence of any functional benefit from these changes, since neither VO2max nor peak power output during the incremental test differed significantly between the groups.
So there you have it. The literature tends to support the idea that the live-high-train-low model has beneficial effects on mechanical efficiency, but probably not (as has been assumed for many years) on systemic oxygen transport. By contrast, intermittent hypoxic training appears to elicit muscle biochemical and structural adaptations that may, under some conditions, result in improvements in VO2max and anaerobic performance. Since both LHTL and IHT appear to exert their influence on performance via the muscles, it seems sensible that future studies should focus on muscular mechanisms. At this time, however, the jury is still out on the role of hypoxia as an ergogenic aid.
What does this mean for athletes and coaches? In practical terms, those considering altitude training should be cautious about investing a large amount of time and money in a practice that will, at best, yield only minor benefits and at worst may even be detrimental to sea-level performance.
If you are really committed to giving it a try, the cheapest and safest option would be to use a sleeping tent.
If, however, you are obliged to compete at high altitude, some acclimatisation will be essential. In this case, a 4-6-week block of inspiratory muscle training before your acclimatisation trip will help to overcome the huge increase in respiratory effort sensation that occurs in everyone attempting to train above 1,500.
- Int J Sports Med 15(6):335-6, 1994
- Med Sci Sport Exerc 16:256-262, 1984
- Haematologica 85:564-572, 2000
- Aviat Space Environ Med 69:793-801, 1998
- Eur J Appl Physiol 75:136-143, 1997
- Med Sci Sport Exerc 23:S25, 1991
- J Appl Physiol 83:102-112, 1997
- Comp Biochem Physiol 128:777- 789, 2001
- Sports Medicine 31:533-557, 2001
- J Appl Physiol 81:636-642, 1996
- Acta Physiol Scand 173:275- 286, 2001
- J Appl Physiol 89:1189-1197, 2000
- Can J Appl Physiol 26: 143- 156, 2001
- High Alt Med Biol 4:291-304, 2003
- J Appl Physiol 96: 931-937, 2004
- Med Sci Sports Exerc 32(6):1130-4, 2000
- J Sport Sci 18:411-419, 2000
- J Appl Physiol 96: 517-525, 2004
- High Alt Med Biol 2:205-221, 2002
- Med Sci Sports Exerc 29:238-243, 1997
- Eur J Appl Physiol 85:486-490, 2001
- Eur J Appl Physiol 84:283-290, 2001
- Int J Sports Med 22:579-585, 2001
- J Appl Physiol 91:1113-1120, 2001
- J Appl Physiol 94:733-743, 2003
- J Appl Physiol 91:173-182, 2001