REF-1539 · Hansen et al., 2005 · Journal of Applied Physiology · Paper · PubMed

Where did "train twice every second day" come from?

7 untrained men, one leg each way, 10 wk: time to exhaustion rose markedly more in the two-session leg; maximal workload gain reported identical. Knee extension, not a sport.

Main caveat: Read from the abstract. No values for the main performance gain, no p values, no glycogen levels before the second hour.
limited, mixed evidenceMeasured on non-athletes

Seven healthy untrained men trained their two legs on different schedules for 10 weeks, with the same work on each. Day 1: both legs trained for 1 h, then 2 h of fasted rest, then one leg (Low) trained for another hour. Day 2: only the other leg (High) trained for 1 h. The gain in maximal workload was reported as identical.

Time to exhaustion at 90% rose markedly more in the Low leg. Resting glycogen and 3-hydroxyacyl-CoA dehydrogenase (HAD) rose significantly only in Low, and citrate synthase rose in both, more in Low. The authors suggest training twice every second day may be better than daily training.

Our reading: the origin of the train-low claim, which also carries the plainest performance result, in the setting least like a competitive athlete's.

Protocol

What it isolates: the whole package, not one factor. The Low leg's second hour combines two sessions in a day, low glycogen (the study's stated reason), a fasted rest and a full rest day for that leg. The trial does not separate them.
Who: 7 healthy untrained men, each his own control, one leg in each condition.
Training: knee-extensor exercise. Low: 1 h, 2 h fasted rest, 1 h, then a day off. High: 1 h on day 1 with the other leg, then 1 h on day 2, so 1 h daily. Repeated for 10 weeks. Workload and amount reported as equal between legs.
Measures: maximal workload, time to exhaustion at 90%, resting muscle glycogen, HAD and citrate synthase activity.
Funding and interests: PubMed: Research Support, Non-U.S. Gov't; sponsor and conflicts not readable in the abstract record.

Key Results

MeasureLow leg (two sessions every second day)High leg (one session daily)Between legs
Increase in maximal workloadreported identicalreported identicalno difference reported
Time to exhaustion at 90%increased markedlyincreasedmarkedly more in Low (no number in the abstract)
Resting glycogenrose, significantnot significantsignificant only in Low
HAD activityrose, significantnot significantsignificant only in Low
Citrate synthase activityroserosemore pronounced in Low
"Identical" is the abstract's word. No equivalence test is reported in it.
"Significant only in Low" is a within-leg result. The abstract prints no between-leg test for glycogen or HAD, so they are not between-leg findings.
Not given: numbers for the time-to-exhaustion gain and any p value.

Mechanisms

Measured: training enzymes and resting glycogen in the thigh muscle.
Authors' interpretation: low glycogen was the expected route to better adaptation, since it had been shown to raise transcription of training genes. The study was built on that speculation.
Inference: both legs share one body, so blood-borne and whole-body factors are common to both, and a difference between legs points to something local to the muscle. It also means that the design says nothing about whole-body fatigue, fuel or sleep costs of a real double.
Not measured, or not in the abstract: whole-body exercise, VO2max, a race or time trial, training quality in the second hour, the PGC-1 family (see REF-1540), women.

Practical Application

Where did "train twice every second day" come from? 7 untrained men, one leg each way, 10 wk: time to exhaustion rose markedly more in the two-session leg; maximal workload gain reported identical. Knee extension, not a sport.
Good for: understanding where the claim started, and its limit: an isolated muscle in untrained men, with no numbers in the abstract.
Good for: a positive result to set beside the trained-cyclist trials on the bike, which had no between-group performance gain (REF-1096, REF-1097).
Not good for: telling a cyclist or runner to stack sessions for a better adaptation.
Not good for: separating the double from low glycogen, or from the rest day. They come as a package here.
Not from this paper: the base's reading of stacked weeks is in SYNTH-84 and of double threshold days in SYNTH-03.

Limitations

Read from the abstract. No values for the main performance gain, no p values, no glycogen levels before the second hour.
n = 7 and one small muscle group. One-leg knee extension is not cycling or running.
Untrained men only; no women.
Factors bundled: two sessions in a day, low glycogen, fasted rest and more rest for that leg.
Both legs in one body, so the two conditions are not independent of each other's training (inference).
How this entry was checked and changed
Why this level

(from the abstract) — within-subject n = 7 (rubric B: crossover <15), one step lower: key numbers are missing from the abstract, including the time-to-exhaustion gain and any p value, and the exercise is isolated knee extension. Subscription article; no copy in PubMed Central or Europe PMC