REF-451 · Achten & Jeukendrup, 2003 · International Journal of Sports Medicine, Vol. · Paper · PubMed
Maximal fat oxidation during exercise in trained men
FatMax measured in 55 endurance-trained men plus a 10-subject reliability test
Background
Fat oxidation rises from low to moderate intensity and falls from moderate to high intensity. A graded-exercise protocol had recently been developed to locate the intensity of maximal fat oxidation (FatMax); this study tested its reliability and measured FatMax in a large group of endurance-trained men.
Participants
n=55 endurance-trained male subjects performed the graded test once (FatMax determination); a further n=10 performed three graded tests to exhaustion for the reliability analysis.
Protocol
Graded exercise tests to exhaustion on a cycle ergometer, after an overnight fast
Diet and exercise on the day before each test were standardized
Indirect calorimetry; fat oxidation rate plotted against intensity to locate maximal fat oxidation (MFO), FatMax and Fatmin (the intensity where fat oxidation becomes negligible)
Stage length and increments are not in the abstract (removed)
Key Findings
MFO: 0.52 ± 0.15 g/min, reached at 62.5 ± 9.8% VO₂max; the average FatMax intensity was 63% VO₂max
Fatmin: 86.1 ± 6.8% VO₂max
Reliability (n=10): typical error for FatMax and Fatmin was 0.23 and 0.33 L O₂/min, i.e. 9.6% and 9.4%; the CV of the FatMax/Fatmin estimate is 9.0-9.5%
When the 55 subjects were split into two groups by VO₂max, the large spread in FatMax and MFO remained in both — the variation is not explained by fitness level alone
Even within this homogeneous group there was a relatively large inter-individual variation in FatMax and MFO
Practical Application
In 55 endurance-trained men tested fasted on a graded cycle test, maximal fat oxidation was 0.52 ± 0.15 g/min at 62.5 ± 9.8% VO₂max (average FatMax 63%), fat oxidation became negligible at 86.1 ± 6.8% VO₂max, and the test-retest CV for FatMax was 9.0-9.5% (n=10). Meaning: the "~63% VO₂max" average hides a wide individual spread (SD almost 10 percentage points) even in a homogeneous trained group — an individual test, not a generic percentage, tells you where an athlete's fat oxidation peaks.
Lab protocol to determine an athlete's individual FatMax: graded test, overnight fast, standardized previous day
Training prescription (the base's own practice, not tested here): place fat-oxidation-targeted Z2 work at or just below the measured FatMax
Inference, not tested: pacing long events near FatMax spares glycogen
Inference, not tested: FatMax may move with training — re-test periodically rather than assume
Limitations
All male, all endurance-trained — generalizability to females and other levels unknown
Graded stages may not reach true steady state at higher intensities
Indirect calorimetry assumes constant urinary N excretion (not measured)
RER unreliable at higher intensities for substrate estimation
Single fasted test; no untrained comparison group, so "trained vs untrained" differences are not tested here
How this entry was checked and changed
Why this level
regraded 2026-09-27 from A: cross-sectional measurement/test-validation study, n=55 (+10 reliability)
Content correction 2026-09-27
"n=18 well-trained cyclists, VO₂max ~62" -> abstract: n=55 endurance-trained men (FatMax), plus n=10 for reliability (three tests each). VO₂max not stated.
"MFO 0.52 ± 0.16 g/min", "FatMax 64 ± 4% (range 55-75%)", "MFO range 0.18-1.01 g/min" -> abstract: 0.52 ± 0.15 g/min at 62.5 ± 9.8% VO₂max (average 63%); Fatmin 86.1 ± 6.8%; typical error 0.23/0.33 L O₂/min (9.6/9.4%), CV 9.0-9.5%. Ranges removed.
"Trained men show higher MFO than untrained" -> no untrained group in the study; removed. "5-min stages at 35-85% VO₂max", "events >3 h", "misses ~20-30% of athletes" -> not in the abstract; removed or relabelled.
Level A -> B (cross-sectional measurement/test-validation study); added Tier/Sport lines.
Cross-reference descriptions checked against the cited entries' current content; wrong years, wrong entries and claims the cited entry does not make were corrected.
2026-10-01: cross-reference repointed from a merged framework to the one that replaced it (SYNTH-01→31, 05→34, 09→70, 12/15→32, 19→26).