Sea-level exercise performance following adaptation to hypoxia: a meta-analysis
51 studies, six protocols; in subelite athletes natural live-high-train-low gave +4.2% (±2.9%) and brief intermittent artificial LHTL +2.6% (±1.2%) in maximal endurance power; in elite athletes only natural LHTL showed a possible +4.0% (±3.7%), and placebo, nocebo and training-camp effects explained part of the gains
Scope
The first meta-analysis of how performance and linked physiological measures change after adapting to one of six hypoxia protocols, natural or artificial: LHTH (live high, train high); LHTL (live high, train low); artificial LHTL where the daily hypoxia is long and continuous (8-18 h), brief and continuous (1.5-5 h) or brief and intermittent (<1.5 h); and artificial LLTH (live low, train high).
51 qualifying studies, giving 11-33 estimates per protocol for power output and up to 20 for VO2max and other mediators.
Models adjusted for study characteristics (altitude level, days of exposure); weighting limited to sample size because inferential statistics were poorly reported. Smallest worthwhile effect on performance set at 1%.
Key Results — maximal endurance power output (controlled studies)
| Protocol | Subelite athletes | Elite athletes |
|---|---|---|
| Natural LHTL | likely +4.2% (90% CL ±2.9%) | possible +4.0% (±3.7%) |
| Artificial brief intermittent LHTL | very likely +2.6% (±1.2%) | unclear |
| Artificial long continuous LHTL | possible +1.4% (±2.0%) | unclear |
| LHTH | unclear +0.9% (±3.4%) | unclear |
| Artificial brief continuous LHTL | unclear +0.7% (±2.5%) | unclear |
| LLTH | unclear +0.9% (±2.4%) | unclear |
Part of these effects was mediated by substantial placebo, nocebo and training-camp effects with some protocols.
"Enhanced" protocols (study characteristics set to their most favourable values) produced clear effects with all protocols in subelite athletes (3.5-6.8%), but in elite athletes only with LHTH (5.2%) and LHTL (4.3%).
Key Results — VO2max and mediators
Subelite: VO2max increase very likely with LHTH (+4.3%, ±2.6%).
Elite: a reduction in VO2max was possible with LHTH (−1.5%, ±2.0%); changes with other protocols unclear.
Erythropoietic and other physiological mediators gave little additional insight into mechanisms.
Authors' summary: for both elite and subelite athletes, natural LHTL is at present the most effective protocol for raising endurance performance; in subelite athletes some artificial protocols also work; VO2max plus placebo, nocebo and training-camp effects are the likely mediators.
Practical Application
A meta-analysis of 51 altitude/hypoxia studies found the clearest sea-level gains in maximal endurance power in subelite athletes with natural live-high-train-low (+4.2%, 90% CL ±2.9%) and with brief intermittent artificial LHTL (+2.6%, ±1.2%); in elite athletes only natural LHTL showed a possible gain (+4.0%, ±3.7%) and everything else was unclear, and part of the effect was attributed to placebo, nocebo and training-camp effects. So natural LHTL is the best-supported protocol, the expected gain is a few percent with wide uncertainty, and the better trained you are the less the artificial alternatives are shown to do.
Inference, not tested: for an elite athlete the possible VO2max reduction with LHTH (−1.5%) is a reason to keep quality sessions at low altitude if you go high — the LHTL logic.
Limitations
2009 data set; later trials are not included and the entry no longer summarises any update.
Weighting by sample size only (poor reporting of inferential statistics in the source studies).
Wide confidence limits on nearly every estimate; several protocol effects are "unclear".
Placebo, nocebo and training-camp effects were substantial and could not be fully separated from the hypoxic dose.
How this entry was checked and changed
Content correction 2026-09-27
All "2023 update" content (Bayesian network meta-analysis; efficacy ranking LHTL natural +2.5-3.5%, simulated +1.5-2.5%, LHTH +1-2%, IHT +0-1%, IHE ≈ placebo; 2000-2500 m for 3-4 weeks; Hb mass +3-5% returning in 2-3 weeks; 2025 IHT meta-analysis) -> not in the 2009 abstract this entry's DOI points to; removed. Journal line now names the 2009 paper only.
"Peak Hb effect 1-2 days post-descent", "performance window 2-3 weeks", "race 1-3 weeks after the block", "heat acclimation +5-8% plasma volume", "event at 1500-2500 m" -> not from the paper; removed or pointed to SYNTH-06 / REF-001 as house practice.
Added the abstract's actual estimates by protocol and athlete level (e.g., natural LHTL +4.2% subelite / +4.0% elite; brief intermittent artificial LHTL +2.6%; LHTH VO2max +4.3% subelite, −1.5% elite) and the placebo/nocebo/camp-effect finding. Sport "running" -> general.