Three Weeks Off: What You Lose, and What Waits

Every training plan assumes you will follow it. Life has other commitments.

So the interesting question is not what happens when you train. It is what happens in the three weeks you don't — the flu, the deadline, the holiday where the running shoes stayed in the bag — and what happens on the first day back, when the same easy pace arrives with ten extra heartbeats and legs that seem to be negotiating rather than cooperating.

Almost everyone reads that first session as a verdict: the year is gone. It isn't. What you lose when you stop is remarkably specific, it arrives in a particular order, and the thing that collapses fastest is the thing that matters least.

The first thing you lose is fluid

In 1986, Edward Coyle's group at Washington University took eight endurance-trained men, stopped them for two to four weeks, and measured what happened while they cycled upright. Blood volume fell 9%, from about 5,177 to 4,692 millilitres, driven mainly by a 12% drop in plasma volume. Stroke volume — the amount of blood the heart moves per beat — fell 12%. Heart rate at a submaximal workload rose 11%. Maximal oxygen uptake fell 6%.

Then they did the thing that makes the study worth remembering. They expanded blood volume back to its trained level and measured again. The cardiovascular signature of detraining largely went away. A smaller tank had been limiting how much blood filled the heart between beats, and refilling the tank restored the numbers.

Read that back against your first session. The ten extra beats at your usual pace are not evidence that your aerobic base evaporated in a fortnight. They are substantially a fluid deficit wearing a costume. This is not an invitation to drink your way back into shape — plasma volume expansion is a trained adaptation, built by repeated exercise through hormonal and protein changes rather than by topping up electrolytes. But it does come back quickly, and it comes back first.

Then the curve flattens — and stays flat

Two years earlier the same laboratory ran the longer version: seven endurance-trained subjects, measured 12, 21, 56 and 84 days after they stopped. VO₂max fell 7% over the first three weeks. It then settled, and by 56 days had stabilised 16% below the trained value, where it stayed.

Sixteen per cent, and then nothing. Not a slide back to untrained. The muscle biopsies explain why. Mitochondrial enzyme activity fell with a half-life of about 12 days — fast — but levelled off around 50% above sedentary controls. Capillary density in the trained muscle did not decline at all; after twelve weeks of doing nothing it was still 50% above sedentary. The metabolic machinery was idling. The plumbing had simply stayed.

Seven people is seven people, so it matters that the pattern survived aggregation. A 2022 meta-analysis pooled 21 detraining studies in trained athletes across cessation periods from 10 to 730 days. Up to 30 days off, VO₂max fell an average of 3.9%. Beyond 30 days, 9.4%. And between 30–90 days and beyond 90 days, no significant further difference. The cliff is early. The plateau is long.

One awkward corollary: the fitter you are, the more you drop. Higher-trained athletes showed the larger declines, which is arithmetic rather than injustice — they were falling from a higher perch, and much of what a big VO₂max is made of is the fluid and enzymatic layer that goes first.

Muscle keeps its own calendar

Aerobic fitness is the anxious tenant. Muscle is the one who never answers the door.

A 2022 meta-analysis gathered the resistance-training studies in older adults that also measured what happened after the programme stopped. Training itself produced a large increase in muscle size, as you would expect. Cessation produced a significant loss overall — and then the duration split did something surprising. Between 12 and 24 weeks without training, the loss of muscle size was not statistically significant. Only at 31 to 52 weeks did it become clear.

Six studies, eight groups: thin evidence, and the confidence interval at 12–24 weeks brushes against a real loss rather than excluding it. But the direction contradicts what every gym says out loud. Half a year off is not, on this evidence, a body reset. It fits the wider pattern that the tissues slowest to build are slowest to leave: fluid moves in days, enzymes in weeks, contractile tissue in seasons, connective tissue in years.

Strength holds on even better than size, for a reason worth knowing: a meaningful share of it is not tissue at all but coordination — motor units recruited in the right order at the right rate. Skills degrade more slowly than substances.

The variable you must not cut

All of which raises the practical question nobody asks until the calendar has already collapsed: what is the least you can do and still keep it?

A 2021 review in the Journal of Strength and Conditioning Research assembled the reduced-training studies to answer exactly that. For endurance, performance held for up to 15 weeks when frequency dropped to as little as two sessions a week, or when session volume was cut by a third to two thirds — down to 13–26 minutes — provided exercising heart rate stayed where it was. For strength and muscle size, the numbers are almost rude: up to 32 weeks maintained on one session a week and one set per exercise. In one of the studies, maximal strength continued to improve on that dose.

One variable ran through every result. Intensity is the load-bearing wall. Frequency and volume are decoration you can strip almost to nothing.

Which is precisely backwards from how people actually triage a busy month. The hard session is the one that gets cancelled — it needs preparation, recovery, a certain mood — while the easy volume, the pleasant hour that costs nothing to start, survives. That is the wrong amputation. Cut the hour, keep the twenty hard minutes. It is the same logic that makes a single weekly session more defensible than it sounds, and the reason the weekly-minutes question is less interesting than the intensity question.

Age is the tax

Every reassuring number above comes with an asterisk, and the asterisk is age.

In that same review, full maintenance on one weekly set held in roughly 20–35 year olds — and not in 60–75 year olds, who appeared to need up to two sessions a week and two to three sets to hold muscle size, with intensity still preserved. The minimal dose is not a fixed quantity. It rises as you do.

A 2025 meta-analysis in the Scandinavian Journal of Medicine & Science in Sports looked at what remains after older adults stop a supervised programme: 15 randomised trials, 21 intervention arms, 787 participants, training periods of 8 to 43 weeks and cessation periods of 4 to 36 weeks. Functional capacity retained a medium, significant protective residue — around 0.9 in effect-size terms. Agility, walking ability, standing ability and stair climbing were all measurably preserved. Gait speed and balance came out positive but not significant, which is a real limit and not a rounding error.

Two moderators tell the story. The larger the gain during the programme, the larger the residue after it. And the older the participant, the less of it persisted. Training banks something durable; the interest rate falls with the years. That is an argument for building the deposit earlier and larger, not for treating any of it as permanent — the same reason general activity does not substitute for structured training as the decades accumulate.

Coming back is not starting over

The most useful thing in Coyle's biopsies is what wasn't there: a return to baseline. After twelve weeks of complete inactivity, his subjects' muscle still had the capillary bed and roughly half the enzymatic surplus that training had built. They were not untrained people. They were trained people running on a smaller fluid volume with quieter mitochondria.

That is the physiological reason a comeback feels dreadful for a fortnight and then abruptly stops feeling dreadful. You are not rebuilding a network. You are re-pressurising one. Nobody has run the definitive trial on exactly how much faster retraining is than training, and anyone who quotes you a precise ratio is guessing. But the asymmetry is not in doubt, and it is the single most underrated fact about interruption.

What to do with three weeks off

If the break has already happened, judge the first two sessions by effort rather than by pace or watts. Your heart rate at a familiar intensity will read high; that is a smaller blood volume, not a smaller you, and it resolves in days rather than months. Chasing the old pace against an elevated heart rate is how a layoff turns into an injury. Watch the trend across a week, not the number on Tuesday — a morning resting heart rate is a better instrument here than any single workout, with the usual caveat that a wearable measures fewer things than it displays.

Do not retest anything for at least a fortnight. A VO₂max estimate or a strength test taken on a detrained body will hand you a number that is 6% pessimistic for no useful reason, and if you rebuild your training zones from it you will spend the next month training too easily.

And if the break is still ahead of you — a trip, a project, a surgery you have a date for — do not plan a pause. Plan a reduction. One hard session per quality per week, full load, minimal volume, and let everything else go without ceremony. On the evidence above, that holds most of what a year of work bought you, for months, on an hour of it.

The bottom line

You lose fitness in a strict order, and the order is kinder than it feels. Blood volume goes first, within a fortnight, taking your submaximal heart rate hostage and returning it almost as fast. VO₂max falls around 4% in the first month, closer to 9% after that, and then largely stops. Muscle size in older adults showed no clear loss up to 24 weeks off. Capillaries did not budge in twelve weeks of nothing.

What actually erodes a training history is not a three-week gap. It is a three-week gap that becomes a habit, because the first session back felt like proof that nothing had been kept. It had. Nearly all of it, in the tissues that took longest to make.

Reduce rather than stop, and if you can only protect one variable, protect the hard one.