SYNTH-36 · Guide
Coming back
The reason for the break decides the return, and a few days off is not the emergency the athlete thinks it is
The decision this framework makes
How to restart after a break — and first, which KIND of break it was, because the three answers are genuinely different.
1. Establish why they stopped. Everything follows from it.
Holiday, work, life — a planned or benign gap. REF-374 (B+): detraining is a loss of adaptation to an *insufficient* stimulus, and the review's short-term window is under four weeks — within it, highly trained athletes lose VO2max and blood volume rapidly, and capillary density and oxidative enzymes are already reduced, while strength declines little (though in the studies themselves capillary density did not consistently fall: it held through 84 days off in REF-377 and moved inconsistently across the endurance-athlete studies in REF-1133, B+).The abstracts give no per-day or per-week percentages; the endurance-athlete studies collected in REF-1133 (B+, systematic review of 41 mostly small studies, no pooling, mostly men) do: after a complete stop, VO2max about 4-7% lower at 12-21 days (one 10-day study found no change), about 10% at 5 weeks and 13-20% at about 8 weeks, with no further significant fall after 12 weeks, and time to exhaustion 9% shorter after 2 weeks — each from a single small study (the review reports sample sizes only for a few, e.g. 7 athletes in Coyle 1985). *Inference:* a gap of a few days sits at the very start of that window, so the honest message to a returning athlete is "you have lost some aerobic sharpness, not your fitness" — not "you have lost nothing".The base's own practice, not from a study: treat a gap of under a week as a rest, not as detraining to be repaired; after 7-14 days step back a week and then continue (SYNTH-72) — in recreational marathoners a complete stop of ≥7 days in a build went with a 5-8% slower finish than the same runner's undisrupted marathon, while stops of up to 6 days were the comparison baseline (REF-941, B+, observational, running only; the cause of the stop travels with it).
Illness — the time out can run well past the symptoms (*inference from one COVID-19 cohort; no entry measures a fitness cost of illness beyond the days missed*). REF-376 (B): in 147 elite UK athletes with symptomatic COVID-19 the illness was mostly mild, yet median time loss was 18 days (vs 6 days for ordinary respiratory illness in the same surveillance system) and 27% were still not fully available after 4 weeks; athletes with lower-respiratory features (breathlessness, chest pain, cough, fever) had 3× the risk of prolonged symptoms.Not from a study; the base's own practice: a week lost to a systemic illness earns a phased return of roughly two weeks, not immediate re-entry.
Getting this wrong in either direction is costly: treating a holiday like an illness wastes a fortnight, and treating an illness like a holiday risks a setback (base reasoning, not from a cited entry).
2. The rule that holds across all three
Not from a study; the base's own practice, for a break of about two weeks or more, or after illness or injury (shorter gaps are handled in SYNTH-72): do not replicate pre-break training in the first week back — start at roughly 50-70% of prior volume, and add intensity only once wellness is stable. What the evidence adds:
REF-374 (B+) gives the shape of the loss: within four weeks the aerobic qualities fall (VO2max, blood volume, oxidative enzymes, lactate threshold; capillaries less consistently — REF-377, REF-1133) while strength largely stays — so it is the aerobic side that needs rebuilding. Its own maintenance recipe for a *reduced* period is: keep intensity, cut frequency only a little, cut volume a lot.
Big jumps in weekly load carry a near-term injury cost in healthy runners: REF-623 (B+, 261 recreational runners) — a 20-60% weekly-distance increase produced 22.6 percentage points more injuries at 21 days than a <20% increase (95% CI 0.9-44.3), with no difference left at 8 and 14 weeks; REF-619 (B, 874 novice runners) — a >30% two-week jump showed a non-significant trend toward more distance-related injuries (HR 1.59, 95% CI 0.96-2.66, p=.07). *Inference:* an athlete re-entering at full prior volume after weeks off is making exactly that kind of jump relative to what they have recently done."Rebuild over 2-3 weeks" is the base's own convention, not a number from these cohorts. REF-365 (C) is a reminder that ACWR-style ratio rules have no evidence behind them for injury management — this is judgement, not a formula.
One change at a time — the base's own practice.
What protects fitness through a low-volume period is a small dose of quality (the first two studies below are reduced-volume/transition periods in athletes who kept training, the third in novices; applying them to a return after a layoff is inference): REF-695 (B, narrative review) — speed-endurance training at near-maximal intensity maintained oxidative capacity and improved 30 s-4 min performance in trained athletes *despite* a reduced training volume; REF-166 (B, 13 well-trained cyclists) — one HIT session every 7-10 days on top of low-intensity riding during an 8-week transition gave a very likely larger power at 4 mmol/L (between-group 10.6%, 90% CI ±8.2%) and 40-min power (12.4%, ±7.6%) than low-intensity only.
In untrained young women just out of a 12-week strength-and-intervals block, the same full session once a week held leg strength, quadriceps size and maximal aerobic power for 12 weeks, while once a fortnight gave back 4.4%, 5.9% and 9.0% of the post-training values — about half of the size and aerobic-power gains (our arithmetic) — and stopping gave back almost all (REF-1136, B, groups of 10, 10 and 7).
So the return is not "easy for a fortnight" — it is reduced volume with the quality reintroduced deliberately, once the athlete is well. *Inference:* for a benign gap, one short sharp session in the first week back is consistent with the evidence; after illness it waits for the gate in §3.
3. Illness specifically
Fever is the red line (REF-609, C): fever raises fluid losses and metabolic demand, cuts strength and endurance through muscle catabolism, strenuous exercise during a febrile illness can worsen it, and no consensus recommendation supports returning before the fever resolves; training resumes gradually once fever and dehydration have cleared. This is a red line, not a judgement call.
The neck check — described, not tested, in the REF-609 full text from expert guidelines (above-the-neck symptoms: mild to moderate exercise for 10-15 min, continue if symptoms do not worsen) (corrected 2026-10-12); the base's own version: symptoms above the neck with no fever → easy Z1-2, cut intensity and volume, reassess daily and abort if it worsens. Fever or any below-the-neck symptom → stop. The base waits 24-48 h fever-free without medication before resuming, and schedules nothing hard for about a week after that (also base practice; the REF-376 entry lists the same line as general guidance that is not from its cohort).
Graded return after anything systemic, with an explicit watch for red flags — resting tachycardia, breathlessness, chest pain, palpitations, unusual fatigue → medical review to exclude myocarditis.This is the base's own practice and general return-to-play guidance carried in REF-609 and REF-376; the REF-376 cohort reported no cardiac outcomes; what it shows is that chest features lengthen the illness: the lower-respiratory phenotype (18% of athletes) tripled the risk of prolonged symptoms. Say this out loud rather than coaching around it.A sustained shift over weeks in several nightly signals away from the athlete's own baseline, together with unexplained fatigue — especially in an older athlete or one with known heart disease — is a reason for a medical check, not a lighter week (REF-1057, C, a single case report analysed in hindsight: a 76-year-old with advanced structural heart disease).
The readiness signal the base uses on the way back is heart rate at easy intensity (base practice from unverified return-to-play guidance, not a study finding). The REF-376 entry carries a 5-stage graduated return (light aerobic <70% HRmax for 2+ days → moderate <80% HRmax → skill/complex → full) from a separate return-to-play consensus, not the cohort — the entry does not name that consensus and has not checked it against a source, so the stage numbers are base practice, not evidence: progress only symptom-free with a normal HR response, drop back 1-2 stages if symptoms recur.Elevated HR for the same easy effort means not ready, regardless of how the athlete feels about the calendar.
A spell in bed is not a rest week. No study has measured bed rest in real athletes; in exercise-trained men, upright endurance fell within 3 days of strict bed rest (VO2max −16.5% in 3 days on an upright test; −3.9% in 7 days on supine tests in another laboratory), and the authors recommend at least 2-4 weeks of progressive rehabilitation, strength included, after up to 28 days in bed — possibly longer when illness or injury is the reason (REF-1134, B, seven small studies of healthy volunteers; the recovery figure is from sedentary and recreational people).Two weeks of strict bed rest took 5% of young men's thigh muscle, 7% of the quadriceps compartment (REF-1137, B, n = 12).Not from a study; the base's own practice: an athlete who spent days in bed starts the return in §2 at its cautious end, with strength work in it, and after a week or more in bed rebuilds over the authors' 2-4 weeks rather than the default 2-3; how much the illness itself adds is still not measured (all volunteers were healthy).
4. Injury and postpartum
Gate every step by symptoms, not by a percentage — not from a study; the base's own practice: resume only after pain-free walking, start walk-run intervals, progress volume before intensity, add roughly one variable at a time.The pain rule (also base practice, the same number as SYNTH-73): ≤3/10, settling by the next morning, with no week-to-week upward trend — back off if it trends up. None of the base's load-progression cohorts (REF-619, REF-623) tested a return-to-run protocol.
A leg taken off loading loses force fast and gets it back with load. This is not a holiday break, where strength declines little (§1): ten days with one leg suspended, walking on crutches, cost young men 29% of quadriceps force against 4% of muscle volume, and nine sessions of 3 × 10 at 70% 1RM over 21 days restored the force (REF-1142, B, n=11, uncontrolled; the retraining was planned at twice the disuse).Not from a study; the base's own practice: once the injury allows, put load back on the limb rather than waiting for walking alone to restore it; a quick rise in muscle size on a scan afterwards was probably partly water in that study (authors' hypothesis).
Postpartum is its own protocol (REF-569, B — Delphi consensus, 42/47 statements agreed): a period of relative rest, then gradual increases in duration and intensity, starting with a walk-run protocol and incorporating strength training, all individualised; modify against musculoskeletal or pelvic-floor symptoms, sleep, mental health, lactation and energy availability.Do not restart with continuous mileage. Not from a study; the base's own practice (the consensus says only "pelvic symptoms"): pelvic-floor symptoms (leaking, heaviness, pressure) trigger regression and a pelvic-health referral — never "push through".
Practical Application
Not from a study; the base's own practice: after about two weeks or more off, or after illness or injury, first week back at 50-70% of prior volume with hard work limited to one short quality session once well, rebuilt over 2-3 weeks; after illness the gate is heart rate at easy effort, and fever or any below-the-neck symptom means stop, not easy. The evidence behind the shape of it: aerobic qualities fall within four weeks of insufficient stimulus while strength holds (REF-374); a 20-60% weekly jump costs injuries in the next three weeks (REF-623); COVID-19 time loss ran to a median 18 days even when mild, against 6 days for ordinary respiratory illness (REF-376, elite cohort); fever means no training (REF-609).
Ask why they stopped before saying anything else. The answer changes the plan more than the length of the gap does.
Reassure on short gaps, honestly. Under a week (the base's own cut-off; REF-941 used stops of up to 6 days as its baseline and priced only stops of ≥7 days, in marathon builds), the message is "you have lost a little aerobic sharpness not your fitness" — REF-374 places the fast losses inside the first four weeks, so a few days is a rest, not a rebuild. That conversation prevents the panic block that actually does damage.
When asked how much went, give the numbers with their weight. After a complete stop in trained endurance athletes, single small studies put VO2max about 4-7% down at 12-21 days, about 10% at 5 weeks and 13-20% at about 8 weeks (REF-1133, unpooled); after strict bed rest the fall starts within 3 days, and the authors advise at least 2-4 weeks of progressive rehabilitation after up to 28 days (REF-1134, exercise-trained men, not athletes).One masters case adds a 12-week point: a 53-year-old long-distance triathlete lost 7% of absolute VO2max (9% per kg) and 10% of peak cycling power in 12 weeks off, and 12 weeks of retraining (from about 50% of usual volume to full volume by week 7, with more high-intensity work and more cycling than before) ended 5% above his pre-break VO2max (REF-1191, C+, n = 1, one test per time point, VO2max on a bike only).
Reintroduce quality sparingly — one hard session every 7-10 days (REF-166 tested that frequency in a transition period, not session size and not a return from a break; REF-695 shows quality holding up under reduced volume but does not test dose size) — one hard session per 7-10 days on low volume protected threshold and 40-min power in REF-166; that is what protects the top end while volume recovers.
Never coach through fever, chest symptoms or breathlessness (REF-609; red flags are base practice) — refer out.
Log the restart in the athlete file so the next block reads as a rebuild rather than as an unexplained dip.
What NOT to do
Do not prescribe a "catch-up" week — the missed sessions are gone and cannot be recovered, and re-entering above prior load is the kind of jump REF-623 associates with more injuries at 21 days in healthy runners (REF-619: a non-significant trend only) (base reasoning; REF-374 does not address catch-up training).
Do not apply a fixed percentage ramp to an injury return (base practice; the percentage cohorts are in healthy runners).
Do not treat a ten-day holiday as detraining to be repaired (base practice); but do not tell the athlete nothing was lost either — REF-374 puts the fast aerobic losses inside the first four weeks.
Do not let an athlete train through fever (REF-609) or below-the-neck symptoms (base practice).
Do not restart with intensity because "volume takes too long" — that ordering is backwards in every one of the three cases (base practice; the evidence supports *small* doses of quality — one HIT session per 7-10 days in REF-166, with REF-695 showing speed-endurance quality works under reduced volume — not a full-intensity restart).
Do not read a poor first week back as lost fitness; it is the expected shape of a return (base practice).
Do not sell a planned long break as a way to come back stronger. The one case that finished above baseline (REF-1191, n = 1) retrained with more high-intensity work and more cycling than before the break and was tested on a bike; its authors cannot exclude that the changed training explains the gain.
Do not read "mild COVID-19" as "back in a week" — median time loss in REF-376 was 18 days (elite COVID-19 cohort; ordinary respiratory illness in the same surveillance system: median 6 days).
Limitations
The detraining evidence is old and broad (REF-374 is a 2000 narrative review spanning populations). Its direction is well established; REF-1133 adds single-study percentages for endurance athletes (unpooled, no quality appraisal, mostly small studies of men from the 1980s-90s, with some numbers that differ from the primary papers), and the 50-70% start-back figure and the one-week cut-off are the base's own conventions, not measured optima.
[REF-376] is an observational elite cohort of 147 athletes from the 2020-21 COVID period, medically supervised; its time-loss figures may over-state what a mild non-systemic illness costs an amateur, and the graduated return framework in its entry comes from a separate consensus that the cohort did not test.
[REF-609] is level C. It supports "fever = stop" and "resume gradually once fever and dehydration have resolved"; the neck check is a widely taught triage heuristic that the paper's full text states from expert guidelines but does not test (corrected 2026-10-12) — which is exactly why the red-flag referral sits beside it.
The load-progression cohorts ([REF-623], [REF-619]) are in healthy recreational and novice runners in a build, not in athletes returning from injury; the symptom-gating rules for injury are base practice and the specific pain thresholds have not been tested in cycling.
How this entry was checked and changed
Why this level
for the prescriptions (house practice) on B-level background evidence — regraded 2026-09-27 from B+ and 2026-10-01 from B: the detraining direction is B+ (REF-374), but every re-entry number here (start-back percentage, rebuild duration, "a week ill → two weeks back") is house practice, not from a cited entry; the illness cohort is B, the load-progression cohorts B/B+, the febrile-illness review C
Content correction 2026-09-27
"REF-374: breaks of ≤10 days are minimal concern and often beneficial; start at 50-70% of prior volume; do not replicate pre-break training; no catch-up week" -> REF-374's corrected entry carries none of these; it gives direction only (fast aerobic losses within < 4 weeks in the highly trained, strength holds) and a maintenance recipe (keep intensity, cut frequency a little, cut volume a lot). The ≤10-day cut-off and 50-70% are relabelled as the base's own practice.
"REF-376: a week lost to illness earns a two-week phased return" -> removed from the entry; the cohort (147 elite athletes, B) shows median time loss 18 days vs 6, 27% unavailable at 4 weeks, lower-respiratory phenotype RR 3.0; the two-week rule is relabelled base practice; the easy-HR gate is attributed to the 5-stage consensus framework carried in the entry, not the cohort.
"REF-623 (B+): rigid percentage rules fail; progress by symptoms; pain ≤3-4/10; walk-run" -> REF-623 is a healthy-runner cohort: a 20-60% weekly jump gave 22.6 pp more injuries at 21 days than <20%; symptom-gating rules relabelled base practice.
"REF-609: neck check; red flags → exclude myocarditis" -> REF-609 supports fever = stop and gradual resumption once fever and dehydration resolve; the neck check and red-flag list are labelled the base's own practice in the entry (corrected 2026-10-12: the REF-609 full text does describe the neck check, untested).
Level B+ -> B (re-entry numbers are house practice; supporting entries regraded to B/C).
Sync after the full-base pass: "REF-569: increase duration before intensity" -> the consensus says gradual increases in duration and intensity (no ordering); the pelvic-floor symptom examples and referral rule labelled the base's own practice, as in the corrected REF-569 entry.
Residual sync (REF-695 re-checked): "in endurance-trained athletes" -> trained athletes (the review spans running, cycling and team sports); "small, sharp doses (REF-695)" -> the small-dose evidence is REF-166 (one HIT per 7-10 days); REF-695 shows quality maintained oxidative capacity and improved 30 s-4 min performance under reduced volume but does not test dose size.
Level parentheticals synced to the cited entries' current Level lines: REF-374 A -> B+.
Level-line note: REF-374 is B+, not A-level.
Resync 2026-09-28: re-read against REF-365 (journal line added) and REF-569 (full-text muscle list); the postpartum and ACWR claims stand.
Verification 2026-10-01: every quoted number from REF-374, REF-376, REF-619, REF-623, REF-166, REF-695, REF-609, REF-569, REF-365 matches the snapshot entries, and all level/n parentheticals match the head blocks.
Verification 2026-10-01: the medical core 'fever = stop' is supported (REF-609, C). The neck check, >=24 h fever-free, red-flag list and myocarditis referral are correctly labelled base practice (corrected 2026-10-12: the REF-609 full text describes the neck check, untested).The NOT-labelled medical items are: the easy-HR gate and 5-stage HRmax numbers (unverified, unnamed consensus), 'illness costs more than the days missed', 'return becomes a relapse', and the COVID 18-day figure generalised to 'mild illness'.
Verification 2026-10-01: SYNTH-30 §4 step 5 ('illness or fever ... No training, no exceptions') conflicts with the neck check here; one of the two has to change — not fixed by a pair in this file because the conflicting line is in SYNTH-30.
Verification 2026-10-01: REF-376's unverified guidance also says 'no hard training within about 7 days of fever resolution'; SYNTH-36 does not carry that line (not a contradiction, but the stricter rule lives only in the entry).
2026-10-01: aligned with SYNTH-72 and SYNTH-73. Short-gap cut-off moved from 'about ten days' to 'under a week', with 7-14 days = step back a week (SYNTH-72): REF-941 uses stops of up to 6 days as its baseline and prices complete stops of ≥7 days in marathon builds, so the more conservative line is the one both cards' evidence supports.The 50-70% restart is now scoped to breaks of about two weeks or more, illness or injury (SYNTH-72 handles shorter gaps). Pain ceiling ≤~3-4/10 -> ≤3/10 (house rule, same as SYNTH-73). Return after fever: 24-48 h fever-free, then nothing hard for about a week (base practice, the assistant's illness rule).
Revision 2026-10-03: added one labelled sentence to the red-flag bullet in §3: a sustained multi-signal shift from the athlete's own baseline over weeks plus unexplained fatigue, especially in an older athlete or with known heart disease, means a medical check, not a lighter week (REF-1057, Lautman 2026, C, single case report, hindsight analysis). No rule, default or level changed.
Revision 2026-10-15: added REF-1133 (Barbieri 2024, B+, full text: single-study VO2max time course in endurance athletes) to §1 holiday, §2 and Practical Application; REF-1134 (Spiering 2023, B, full text: bed rest) and REF-1137 (Fuchs 2025, B, full text: thigh muscle −5% in 14 days of bed rest) as a new §3 bullet; REF-1136 (Mpampoulis 2024, B, full text: once a week vs once a fortnight) to §2.The REF-374 capillary line is kept and qualified: capillaries held for 84 days in REF-377 and moved inconsistently in REF-1133. Limitations updated. Defaults, gates and level unchanged.
Revision 2026-10-15: added REF-1142 (Franchi 2025, B, n=11, uncontrolled, full text: 10 days of one-leg unloading cost 29% of quadriceps force, restored by nine sessions of 3 × 10 at 70% 1RM over 21 days) to §4 and Cross-References. Prescriptions and level unchanged.
Revision 2026-10-22: added REF-1191 (Zanou 2024, C+, single case, full text): 12 weeks off cost a 53-year-old long-distance triathlete 7% of absolute VO2max (4.58 → 4.27 L/min; 9% per kg) and 10% of peak power; 12 weeks back (full volume by week 7) ended 5% above baseline — to the 'how much went' Practical bullet, a What NOT bullet (a planned break is not a way to come back stronger: the retraining mix changed) and Cross-References. Defaults, gates and level unchanged.