REF-1133 · Barbieri et al., 2024 · Frontiers in Physiology · Paper · PubMed · Free full text

I took 3 weeks off — how much did I lose?

Some aerobic capacity, not most of it: in trained endurance athletes who stopped completely, VO2max was about 4-7% lower after 12-21 days, about 10% after 5 weeks and 13-20% after about 8 weeks (single small studies, mostly men, not pooled). One 10-day study found no change; time to exhaustion was 9% shorter after 2 weeks; little further loss after 12 weeks.

B+moderate evidenceMeasured on trained to elite athletes

Barbieri et al. searched PubMed, Scopus, Web of Science and Medline to July 2023 for studies of endurance athletes that reported how long training stopped or was reduced and gave pre and post values. Forty-one studies met the criteria (endurance-trained by VO2max floors or training history).

They follow Mujika and Padilla (REF-374) in splitting short-term (up to 4 weeks) from long-term (over 4 weeks) detraining, tabulate the ranges per variable (Table 1) and plot each variable against days (Figure 2).

Complete cessation lowers VO2max, mostly through lost blood and plasma volume early and a smaller arterio-venous oxygen difference later; heart rate rises, stroke volume and maximal cardiac output fall, the left ventricle gets lighter and thinner after more than 4 weeks; lactate threshold falls, muscle glycogen drops, RER rises and body mass increases beyond 4 weeks; insulin sensitivity falls within days.

Partial reduction with at least one weekly moderate-to-high intensity session maintained VO2max in the few studies that tested it. There was no quality appraisal and no meta-analysis, so every number below is a single study's result as the review reports it.

Included studies

Search: PubMed, Scopus, Web of Science, Medline and library articles to July 2023; 74 records, 60 after screening, 41 included (19 excluded, mainly for training history or VO2max below the floors).
Population: endurance-trained athletes — performance level 3 (VO2max at least 48 mL/kg/min or 3 L/min for women, 55 mL/kg/min or 4.2 L/min for men), or at least 5 training hours or sessions a week with 2+ years of experience when VO2max was not measured (including masters over 50); elite or Olympic status accepted when training years were not given.Running, cycling, swimming, kayaking, skiing, ski mountaineering, mountain running, rowing, triathlon. The authors state most were male.
Exposure: total cessation or partial reduction, any duration; short-term up to 4 weeks, long-term over 4 weeks. One included case was a masters cyclist stopped by a clavicle fracture.
Synthesis: a data-extraction table (Supplementary Table S1, data on Zenodo), the short- vs long-term range table (Table 1) and a timeline figure (Figure 2) where several data points in one interval are averaged with their maximum and minimum. No risk-of-bias tool, no effect sizes, no pooling.

Key Results

Complete cessation, as reported in the text (single studies; the glycogen row is reduced training)

VariableTime offChangeSource study as cited
VO2max10 daysno changeCullinane 1986
VO2max12 days−7% (runners and cyclists)Coyle 1984 (REF-377; its abstract says −7% over the first 21 days)
VO2max14 days−4.7% (distance runners)Houmard 1992
VO2max15 days−3.7%Chen 2022
VO2max3 weeks−4.18%Martin 1986
VO2max5 weeks−10.1%García-Pallarés 2009
VO2max56 days−13%Coyle 1984/1985 (REF-377 reports 16% below the trained value by day 56)
VO2max2 months−20%; no further significant fall after 12 weeksMartin 1986
Time to exhaustion2 weeks−9%, running economy unchangedHoumard 1992
Blood / plasma volume2-4 weeks−9% / −12%Coyle 1986
Resting plasma volume14 days−5.1 ± 1.9%Houmard 1992
HRmaxshort-term+2 to 9 beats/minseveral
Stroke volume (upright)10 d / 2 wk / 56 d−0.9% / −4.1% / −14%Doherty 2003, Chen 2022, Coyle 1984
Maximal cardiac output5-15 d / 21 dunchanged / −8%, no further fallDoherty 2003, Coyle 1984
Citrate synthase6 d / 10 d / 56 d−5% / −28.6% / −40%Coyle 1984, McCoy 1994
GLUT46 d / 10 d−17% / −33% (one study: unchanged at 14 d)Vukovich 1996, McCoy 1994
Muscle glycogen4 weeks reduced training−18%Madsen 1993
Post-exercise lactate, same submaximal load84 days1.9 → 7.6 mmol/LCoyle 1985
RER, same absolute submaximal load84 days0.93 → 1Coyle 1985
Capillary density12-84 days / 15 daysunchanged / −6%Coyle 1984 / Houston 1979
Max power (Wmax), 2 monthsyoung 19-25 vs older 50-65 cyclists−3.5% vs −8.8%Giada 1998

Short-term (0-4 weeks) vs long-term (over 4 weeks) ranges, Table 1

VariableShort-termLong-term
VO2max−25.5% to +2.4%−16% to −11.3%
HRmax+1% to +5%+1% to +1.5%
Qmax−8% to +4.4%−10% to −9%
Stroke volume−12% to +3.4%−14% to −13%
Left ventricular mass−19.5% to −1.8%"−23.8% < x < 19.5%" as printed
Body mass+0.3% to +8.7%+1.6% to +11.3%
a-v O2 difference−1.3% to +2%−7% to −4%
Capillary density−6.3% to +7.1%−5.6% to +2.6%
Insulin+7% to +39%not available

Partial reduction (as reported): VO2max appeared maintained after 4 weeks of low-volume, moderate-to-high intensity training (Madsen 1993, McConell 1993); one weekly HIT session on reduced low-intensity volume over an 8-week transition maintained it while low-intensity training alone did not (Rønnestad 2014, see REF-166); a weekly sprint session in a 3-week transition made no significant VO2max difference against low-intensity only (Almquist 2020, see REF-165); elite kayakers lost 5.6% of VO2max after 5 weeks of reduced volume at moderate intensity (the primary abstract, REF-950, gives −4.8%); one weekly 35-min HIT session at 95% HRmax over "a more extended period" of reduction still left triathletes' time to exhaustion at 75% VO2max 21% lower (Madsen 1993).

Mechanisms

Measured (in the included studies): plasma volume fell 12% and blood volume 9% within 2-4 weeks, and the early VO2max loss coincided with lower stroke volume and a rise in maximal and submaximal heart rate.
Measured: oxidative enzymes fell fast (citrate synthase −28.6% in 10 days; succinate dehydrogenase with a half-life of 12 days) while maximal a-v O2 difference fell only after about 3-8 weeks.
Authors' interpretation: the drop in VO2max in the first 4 weeks is mainly the loss of blood and plasma volume; the further loss between the 3rd and 12th week is attributed to a smaller a-v O2 difference, probably fewer mitochondria, with blood flow and capillary transit time not excluded.
Measured: after 84 days, the same submaximal load produced post-exercise lactate of 7.6 instead of 1.9 mmol/L and an RER of 1 instead of 0.93 — a shift toward carbohydrate.
Authors' interpretation: masters athletes may keep some adaptations from lifelong training through detraining; the only direct comparison they cite (Giada 1998) found older cyclists lost MORE maximal power over 2 months (−8.8% vs −3.5%), without matched in-season training.
Inference, not tested here: losses carried by plasma volume should return faster than those carried by mitochondria; the review measures no retraining time course.

Practical Application

I took 3 weeks off — how much did I lose? Some aerobic capacity, not most of it: in trained endurance athletes who stopped completely, VO2max was about 4-7% lower after 12-21 days, about 10% after 5 weeks and 13-20% after about 8 weeks (single small studies, mostly men, not pooled). One 10-day study found no change; time to exhaustion was 9% shorter after 2 weeks; little further loss after 12 weeks.
If you can still train a little, keep one hard session a week. In the few reduced-training studies, VO2max held when at least one weekly moderate-to-high intensity session stayed in a reduced week (Rønnestad 2014, see REF-166); low-intensity riding alone did not hold it, and one weekly session did not hold triathletes' time to exhaustion at 75% VO2max (−21%). The maintenance rules for general populations are in REF-587.
Expect heart rate to run higher than usual at the same pace or power in the first weeks back. Upright submaximal heart rate was 2.6% higher after 21 days off and 10-11% higher after 56-84 days (Martin 1986, measured); the authors attribute it to lost blood volume. The base's return rules are in SYNTH-36.
Weight gain is clearest after about a month off. The text says body mass and composition did not change in the short-term studies, yet Table 1 lists +0.3% to +8.7% for up to 4 weeks; beyond 4 weeks the range is +1.6% to +11.3%.
Not from this paper: how fast the losses come back. The review gives no retraining time course; its "return to fitness" section recommends low-volume HIIT from other literature.

Limitations

No appraisal and no pooling. The review used PRISMA for the search but assessed no risk of bias and calculated no effect sizes; Table 1 reports ranges, Figure 2 joins points from different cohorts (single studies, or averages where several fall in one interval) with lines, so neither is a measured trajectory.
The evidence underneath is thin and old. Most VO2max time points come from single small uncontrolled pre-post studies (Coyle 1984, Houmard 1992, Martin 1986; the review gives sample sizes only for a few, e.g. 7 athletes in Coyle 1985); the authors note that athletes rarely agree to stop for long and that most participants were male.
Internal inconsistencies. Table 1 gives short-term VO2max as low as −25.5% (a single day-28 point in Figure 2) while the text reports −3.7% to −7% for 12-21 days and −4.18% at 3 weeks; the long-term range (−16% to −11.3%) omits the text's −20% at 2 months; the left-ventricular mass long-term range is printed as "−23.8% < x < 19.5%" and the type IIA range is inverted; the Garcia-Pallares kayak result is cited to Martin 1986; Figure 2's hormone panel reaches about +100% while Table 1 gives insulin up to +39% and cortisol up to +83%; the text says body mass and composition did not change in the short-term studies, yet Table 1 lists +0.3% to +8.7% for up to 4 weeks.
Numbers differ from the primary source. The text gives Coyle 1984's VO2max loss as −7% at 12 days and −13% at 56 days; the Coyle abstract (REF-377) gives −7% over the first 21 days and −16% at day 56. The review gives the kayakers' VO2max loss with reduced training as −5.6%; the García-Pallarés abstract (REF-950) gives −4.8% (and −10.1% for complete cessation). We quote the review and flag the difference.
Total cessation and reduction are mixed. The extraction table classifies each study, but Figure 2 does not say whether partial-reduction studies are among its points.
Return to fitness is not reviewed. The section on it cites general HIIT literature, not retraining studies after detraining.
Funding: Italian Ministry of Health (Ricerca Corrente, IRCCS MultiMedica); no conflicts declared.
How this entry was checked and changed
Why this level

systematic search of four databases, but no risk-of-bias assessment, no pooling and no registered protocol; the included studies are small, mostly uncontrolled pre-post designs (the VO2max time course rests on single small studies, e.g. Coyle 1984 with 7 athletes, most from the 1980s-90s), mostly men; several numbers in the text, Table 1 and Figure 2 disagree (see Limitations); read from the full text

Full text (checked 2026-10-15)
74 studies were discovered, 60 remained, "only 41 of these met the inclusion criteria" (full text).
Short-term "less than or equal to 4 weeks"; long-term "longer than 4 weeks" (full text).
Inclusion: performance level 3, female "≥48 mL min–1 · kg–1", male "≥55 mL min–1 kg–1", otherwise "≥5 endurance training hours or sessions per week with at least 2 years of experience" (full text).
Coyle et al. "reported a −7% decline of VO2max in a group of runners and cyclists just 12 days later training interruption" (full text).
"14 days of training cessation, caused −4.7% losses"; "15 days of total stop (−3.7%)"; "no change in VO2max after 10 days" (full text).
"up to 3 weeks, there is a decrease of 4.18%"; "After 5 weeks there is a decrease of 10.1%"; "following 56 days a decrease of 13%"; "after 2 months a decrement of 20%"; "No further significant decrements are found after 12 weeks" (full text).
"a 9% reduction of blood volume, attributed to a 12% decline in plasma volume"; "5.1 %± 1.9% decrease in resting plasma volume following 14 days" (full text).
HRmax "increases by 2-9 beats per minute"; stroke volume "after 10 days −0.9%", "−4.1% after 2 weeks", "−14% after 56 days"; cardiac output "After 21 days, a decline of 8%" (full text).
"a 9% decrease in Time To Exhaustion (TTE), with no observed changes in running economy"; triathletes' TTE "reducing it by 21%" with one weekly HIT session "(35 min at 95% HRmax)" (full text).
RER "from 0.93 of the trained state to 1, after 84 days"; lactate "from 1.9 to 7.6 mmol/L" after 84 days; glycogen "decrease of muscle glycogen by 18%" after 4 weeks of reduction (full text).
"A 5% decrease in CS was observed following 6 days", "a 28.6% CS decrease ... after 10 days", "after 56 days, the CS reduction was equal to 40%"; GLUT4 "reduced by 17% after 6 days" and "by 33% in the first 10 days"; succinate dehydrogenase "half-life of 12 days" (full text).
Capillarization "unchanged ... after 12–21-56–84 days"; Houston "a 6% decrease after 15 days" (full text).
Elite kayakers "a loss of–5.6% in VO2max after 5 weeks of reduced training volume at moderate intensity" (full text).
Giada: "(−3.5% vs. −8.8%, respectively)" for Wmax in younger vs older cyclists after 2 months (full text).
Table 1 VO2max "−25.5% < x < +2.4%" short-term and "−16% < x < −11.3%" long-term; body mass "+1.6% < x < +11.3%" long-term (full text).
Funding: "Italian Ministry of Health–Ricerca Corrente–IRCCS MultiMedica"; no conflict of interest declared (full text).