Journal of Applied Physiology

Counterpoint: Positive effects of intermittent hypoxia (live high:train low) on exercise performance are not mediated primarily by augmented red cell volume

Christopher J Gore, Will G Hopkins

To engage in this debate we will address the following questions: what is the change in performance after adaptation to living high and training low (LHTL); what physiological mechanisms could be responsible; what is the evidence that a change in red cell volume (RCV) is one such mechanism; and what is the evidence for other mechanisms?

Change in endurance performance after LHTL.

The smallest worthwhile change in performance time for elite middle-distance runners is ∼0.5% (15, 16). Controlled trials of LHTL via either terrestrial or artificial hypoxia have consistently revealed greater increases in endurance performance, typically ∼1% (10-12, 19, 22, 27). However, none of the studies was performed blind, and the placebo effect may be of similar magnitude (4). Substantial correlations between individual responses in performance and changes in physiology would provide assurance that the performance change is not entirely a placebo effect (14), but the evidence for such correlations is still unclear (see below). Meanwhile we assume that LHTL produces a physiologically mediated enhancement in endurance performance averaging approximately ≤1%.

Mechanisms for LHTL-mediated enhancement of endurance performance.

Exercise tests performed at intensities greater than V̇o2max last <10 min and are powered partly by anaerobic mechanisms, whereas longer exercise is powered essentially by the aerobic system (7). If LHTL enhances only anaerobic mechanisms, the enhancement of performance would decline to zero for tests lasting >10 min. The observed effects of LHTL are as follows: 45 s, 0.8% (22); 4 min, 1.0% (10); 9 min, 1.1% (27), 1.5% (12) and 1.8% (11); and 17 min, 1.3% (19). Although uncertainty in these estimates precludes firm conclusions, it would appear that LHTL affects mainly the aerobic system.

Di Prampero (5) realized that aerobic power at intensities below V̇o2max is the product of three components: V̇o2max, the fraction of V̇o2max representing exercise intensity (Vo2fracmax), and exercise economy (power per unit of Vo2). It follows that percent changes in these components add up to the percent change in endurance performance, apart from measurement error. Furthermore, changes in performance could be due to changes in any component and to more fundamental physiological effects underlying it.

Evidence for the role of RCV.

An increased RCV would enhance performance by increasing V̇o2max via increases in maximum cardiac output (from increased total blood volume) or oxygen-carrying capacity (from increased hemoglobin concentration). Levine and colleagues (3, 19) reported increased RCV of 5–8% after terrestrial LHTL, but we believe that such large changes are more likely to be artifacts of measurement error than physiological adaptations to moderate altitude (9). Indeed, RCV change in the various studies appears to be directly proportional to measurement error! In studies of artificial LHTL at the Australian Institute of Sport, where the carbon-monoxide method is used to measure hemoglobin mass rather than RCV, the measurement error is ∼2% and observed changes in hemoglobin mass are consistent with little or no real change (1, 2, 24). Moreover, artificial LHTL results in little evidence of an increase in reticulocytes, despite transient increases in erythropoietin concentration (1). Nevertheless, the real changes in hemoglobin mass or RCV in some studies may be sufficient to account for the changes in performance.

A correlation between changes in individuals’ RCV and performance after LHTL would represent additional circumstantial evidence for the role of RCV. There are no reports of such correlations, presumably because the correlations were statistically nonsignificant. Lack of significance could be due to masking of a substantial correlation by large measurement error, but in our view it is more likely the correlations are truly small or trivial.

In the absence of a clear direct relationship between RCV and performance after LHTL, Levine and coauthors focused on the role of V̇o2max. Again, changes in mean V̇o2max and correlations between individuals’ changes in V̇o2max and performance are less than compelling. There is a wide range in the mean change in V̇o2max after LHTL (23), and uncertainty in the estimates makes the range even wider. A change in V̇o2max sufficient to explain the performance change after LHTL (∼1%) is therefore possible in many studies, and we accept that these studies represent supporting but not sufficient evidence that a change in RCV is the mechanism. However, in most studies, the true change in V̇o2max could have been trivial or negative, which would necessitate some other mechanism. In the only report of a correlation between changes in V̇o2max and performance after LHTL, Levine and Stray-Gundersen (19) stated that “…the close correlation between the increase in V̇o2max and improvement in 5000-m time…argues strongly that this is the key adaptation during altitude training.” The correlation was indeed strong (0.63) for the pooled data of subjects in all three groups (LHTL, live high-train high, and control). However, performance did not improve posttreatment for the latter two groups, and the correlation was smaller for the LHTL subjects (0.51, recalculated from their Fig. 6). Most of this correlation was due to one subject, who ran 12% slower in the posttest. In any case, V̇o2max in an incremental test often does not show a plateau with endurance athletes (6); V̇o2max itself may therefore be modified by voluntary effort, so at least part of the correlation between changes in V̇o2max and time-trial time could be due to the placebo effect.

Researchers have also investigated the relationship between changes in RCV and V̇o2max for evidence of the role of RCV in performance enhancement. Levine and Stray-Gundersen (19) reported a correlation of 0.37. There have been reports of correlations of 0.70 after 24 days of LHTL (28) but only 0.04 after 46 days of artificial LHTL, with measurement of hemoglobin mass rather than RCV (25). We conclude that in some studies there is evidence consistent with an increase in V̇o2max due to increases in RCV, but the extent to which these changes contribute to performance enhancement is unclear.

Evidence for other mechanisms.

We suspect that exercise economy is the component of the di Prampero model most likely to mediate effects of LHTL. Improvements in economy of 3–6% have been observed after various hypoxic interventions with athletes (8, 17, 18, 21, 24, 25), although correlations with change in performance have not been reported. A switch to a more economic mode of oxygen utilization is a teleologically appealing adaptation to a shortage of oxygen in tissues (13), and a suitable regulatory system involving various intracellular changes mediated by hypoxia inducible factor (20, 26) exists in most cells.

LHTL researchers have neglected the Vo2fracmax component of endurance performance and its surrogate, lactate anaerobic threshold. Although it is unclear why adaptive responses to hypoxia would include changes in this component, its role needs to be clarified experimentally. Finally, even if an increase in the V̇o2max component mediates the effect of some LHTL protocols on endurance, the underlying mechanism need not be an increased RCV; teleologically and physiologically plausible alternatives could involve changes in cardiovascular regulation that result in increased muscle blood flow during intense exercise.

In conclusion, the quality and quantity of published data are insufficient to elucidate the mechanism of the effect of LHTL on performance, but improvements in economy seem more likely than increases in RCV. Future studies should attend to methodological issues, including double-blind designs, smaller errors of measurement for performance and putative physiological mechanism variables, and measurement of economy and Vo2fracmax, in addition to V̇o2max.