REF-1678 · Grivas et al., 2026 · Sports Health · Paper · PubMed

Altitude or hypoxic exposure gave middle- and long-distance runners a small time-trial gain over controls (SMD 0.31, 95% CI 0.02 to 0.60), with no significant change in VO2max (0.32, −0.03 to 0.67), aerobic endurance (0.12, −0.20 to 0.44) or running economy (−0.03, −0.20 to 0.13), while haemoglobin concentration rose (0.77, 0.19 to 1.36)

A meta-analysis of 19 controlled studies in healthy adult runners, read from its abstract

Main caveat: Controlled, not necessarily randomised, studies; altitude and every form of hypoxia pooled together.
moderate evidenceMeasured on trained athletesRead from the abstract only

Grivas and colleagues pooled controlled experimental studies of altitude or hypoxic exposure in healthy adult middle- and long-distance runners, asking what it does to running time-trial performance and to the physiological measures usually reported with it.

They searched four databases and included 19 studies that compared an exposed group with a control group and reported a time trial, VO2max, aerobic endurance, running economy or a haemoglobin-related measure.

In random-effects models, time trials improved with exposure by a small amount that reached significance; the VO2max effect missed significance; aerobic endurance and running economy showed no meaningful effect; haemoglobin concentration rose.

The authors conclude that altitude or hypoxia appears to give small time-trial improvements, while the physiological responses are less consistent across outcomes and protocols.

Participants

Healthy adult middle- and long-distance runners in 19 controlled studies.
Number of runners, sex split, age, training volume and performance level: not in the abstract.

Protocol

Search: Medline, Embase, SportDiscus, Web of Science; dates not in the abstract.
Inclusion: experimental studies in healthy adult middle- and long-distance runners with altitude or hypoxic exposure and a control group, reporting a running time trial, VO2max, aerobic endurance, running economy or a haemoglobin-related outcome.
Analysis: random-effects standardised mean differences with 95% CI.
Not in the abstract: risk of bias, heterogeneity, the hypoxic methods (living, sleeping or training high; natural or simulated), doses, the control conditions, the time-trial distances, how "aerobic endurance" was measured, and how many studies fed each outcome.

Key Results

Outcome (exposure vs control)SMD (95% CI)p
Time-trial performance0.31 (0.02 to 0.60), small, significant< 0.05
VO2max0.32 (−0.03 to 0.67), not significant0.07
Aerobic endurance0.12 (−0.20 to 0.44), not significant0.47
Running economy−0.03 (−0.20 to 0.13), not significant0.70
Haemoglobin concentration0.77 (0.19 to 1.36), significant< 0.001
The time-trial interval runs from almost zero (0.02) to a moderate effect (0.60).

Practical Application

Altitude or hypoxia was associated with a small time-trial gain in distance runners (SMD 0.31, 95% CI 0.02-0.60, 19 controlled studies), with no clear change in VO2max, aerobic endurance or running economy. In standard deviations, not seconds; the lower limit sits near zero.
This fits [SYNTH-06]'s "expect a few per cent at best" in direction and in modesty; the abstract gives no percentage and no dose, so it neither supports nor tests the guide's altitude dose.
Not settled by this design / not from this paper: a 2026 meta-analysis limited to randomised trials in competitive runners found no significant time-trial gain against normoxic training (REF-1676); neither found a significant change in running economy. Which of the pooled methods (natural altitude, sleeping or training in hypoxia) carries the time-trial result is not in the abstract.
Inference, not tested here: the rise in haemoglobin concentration did not come with a clear VO2max change in the same pooled studies, so a blood value after a camp says little on its own about race readiness; the abstract gives no haemoglobin mass result, and a rise in concentration can also come from a fall in plasma volume.

Limitations

Controlled, not necessarily randomised, studies; altitude and every form of hypoxia pooled together.
The time-trial gain is small and its interval nearly touches zero.
Written from the abstract: risk of bias, heterogeneity, sample sizes, doses, methods, the control conditions and publication bias are unknown.
How this entry was checked and changed
Why this level

risk of bias not read (abstract only): the abstract reports no risk-of-bias or certainty assessment, and inclusion asked for a control group, not randomisation; the headline result is time-trial performance, so no marker step