Training load and fatigue: what the studies could and could not measure

Overtraining syndrome has no gold-standard test and is a diagnosis of exclusion. A scoping review found 39 eligible studies out of 5,561 results, and a meta-analysis of HRV-guided training found an effect on submaximal physiology and none on performance.

thinPlausible, under-studied. Treat as a hypothesis.Revised 19 Oct 2026

A pair of dark training shoes and a folded pale grey cotton towel on a dark rubber gym floor beside a black stopwatch with a blank face

The measurements everyone trusts most are the ones with the least validation behind them.

Fatigue is the one training variable that is felt before it can be measured. This is a field where the most confident signals — morning heart rate, resting blood pressure, sleep, soreness — have the thinnest validation behind them, and the article exists to make that asymmetry explicit.

There is no test for overtraining

Carrard and colleagues’ scoping review in Sports Health is the document that settles the question. Overtraining syndrome is characterised by a long-term performance decrement after a persisting imbalance between training-related and non-training-related load and recovery — and because there is no gold-standard diagnostic test, it remains a diagnosis of exclusion.

The scale of the search: PubMed, Web of Science and SPORTDiscus from database inception to 4 February 2021, with backward and forward citation tracking, following the Joanna Briggs Institute guidelines and the PRISMA extension for scoping reviews. Level of evidence: 4. Five thousand five hundred and sixty-one results, thirty-nine eligible studies.

The candidate measurements span an enormous range: basal hormone, neurotransmitter and other metabolite levels; hormonal responses to stimuli; psychological questionnaires; exercise tests; heart rate variability; electroencephalography; immunological and redox parameters; muscle structure; and body composition. Three named diagnostic scores emerged — EROS-CLINICAL, EROS-SIMPLIFIED and EROS-COMPLETE.

And the authors’ conclusion is the honest summary of the field: the scores must now be validated in larger samples and within female athletes.

What a wearable actually buys you

Düking and colleagues’ systematic review with meta-analysis, in the Journal of Science and Medicine in Sport 2021, tested whether training guided by heart rate variability beats a pre-planned schedule. Eight studies, 198 participants, nine interventions, Hedges’ g with 95 per cent confidence intervals.

Submaximal physiological parameters: a significant medium-sized effect, g = 0.296 (95% CI 0.031 to 0.562, p = 0.028). Performance: g = 0.079 (95% CI −0.050 to 0.393, p = 0.597). VO2peak: g = 0.171 (95% CI −0.213 to 0.371, p = 0.130). Small and not statistically significant on both.

The conclusion keeps the good news and drops the bad: HRV-guided training had a medium-sized effect on submaximal physiology and only a small, non-significant influence on performance and VO2peak. There were fewer non-responders on performance with HRV-based training. Most HRV-guided interventions also included fewer moderate- and/or high-intensity sessions.

That last detail is the one to hold onto. The wearable changed what got prescribed; nobody has shown the substitution is the reason the physiology moved.

The ratio that has no threshold

The acute:chronic workload ratio is the most widely repeated number in applied sports science, and Jiang and colleagues’ systematic review in the International Journal of Environmental Research and Public Health 2022 explains why it should not be quoted as a threshold. It examined professional male soccer players and concluded that short-term fixture congestion may increase match injury incidence while long-term congestion may have no effect on overall or match injury incidence; that it is impossible to determine conclusively whether any GPS-derived metric — total distance, high-speed running distance, acceleration — is associated with increased injury risk; and that the acute:chronic workload ratio of session rating of perceived exertion may be significantly associated with non-contact injury risk, but that a threshold with a minimum risk of injury could not be obtained.

“There may be a signal but we cannot say where the line sits” is the honest state of a large literature. The 0.8 to 1.3 band you will find quoted is not in this paper.

What the consensus statement does say

The 2018 International Journal of Sports Physiology and Performance consensus statement from the Technical University Munich symposium is the field’s shared vocabulary. Systematic monitoring of recovery and the subsequent implementation of recovery routines aims at maximising performance and preventing negative developments such as underrecovery, nonfunctional overreaching, the overtraining syndrome, injuries or illness.

The sentence doing the real work is the one about variability: because of the inter- and intraindividual variability of responses to training, competition and recovery strategies, a diverse set of expertise is required. Which is the consensus position that “your fatigue chart” is a population statement, not a personal instrument.

The confound nobody puts on the label

Cupka and Sedliak’s mini-review on low energy availability in male endurance athletes is the reminder that fatigue has causes outside training. Low energy availability arises from the inability to cover energy needs for training or normal physiological functions, and it is a different quantity from energy balance, which compares total daily intake with all energy expended regardless of fat-free mass. Insufficient energy consumption affects recovery and adaptation processes and increases the risk of injury or illness. In endurance-trained men it correlates with decreased testosterone levels, decreased bone density and reduced resting metabolic rate.

If you have been training hard for months, feeling flat, and assuming the answer is more recovery, the energy side of the ledger is the first thing to check — and it is measurable in a way that “am I overtrained” is not.

What these studies do not show

They do not show that any single marker can diagnose overtraining syndrome; the scoping review is a map of candidates, level of evidence 4, with a diagnosis of exclusion at the centre of it. The HRV meta-analysis rests on 198 participants across eight studies and did not detect an effect on the outcome most people care about. The soccer review is specific to professional male footballers and cannot be transferred to recreational lifting. And the low-energy-availability review is a mini-review in male endurance athletes, not a diagnostic test for anyone reading this.

This article summarises published reviews and meta-analyses. It is not medical advice and it is not a recommendation for any individual.