Tendon rehabilitation loading

Why does resting a damaged tendon completely leave it weaker than carefully loading it while it still hurts?

The instinct is old and feels like common sense: something is damaged, so stop using it. Splint it, rest it, let it mend. For a broken bone that instinct is broadly right.

With a chronically painful tendon it performs badly, and the finding that keeps turning up is the opposite of what it predicts — complete rest tends to leave the tendon less capable than careful loading does, even loading applied while the thing still hurts. So the first question is not what to do about it. It is: what kind of object is a tendon, such that leaving it alone makes it worse?

Picture the tissue. A tendon is mostly type I collagen laid down in dense parallel bundles, with relatively few cells scattered through it and a modest blood supply. It looks inert. It is not. That matrix is continuously turned over — degraded and rebuilt — and the cells doing the rebuilding, the tenocytes, are mechanically sensitive. Strain in the matrix around them is transduced into signalling that changes what they synthesise.

Sit with that for a moment, because it changes the whole question. If the tissue's composition is maintained by an ongoing process, and that process takes its cue from mechanical strain, then there is no such thing as a neutral setting. Removing load is not pausing the tendon. It is sending a signal, and the signal says: less is needed here.

What follows? In controlled immobilisation, tendon loses stiffness and cross-sectional area, and markers of collagen synthesis fall. That is the direct evidence for the mechanism, and it is why "rest" cannot be treated as a null condition.

There is a second thread, arguably the larger one. The muscle the tendon serves atrophies quickly when unused, and the whole limb loses capacity. So the athlete who rests for six weeks returns to the same tasks with less machinery to perform them — the tendon meets the old loads with less protection than before. Rest did not merely fail to build the tendon; it removed the buffer around it.

Now the part that feels most wrong: loading it while it hurts. What is a chronic painful tendon actually like? It was long called tendinitis, on the assumption of inflammation, and that name was largely abandoned because the tissue does not look inflamed in the classical sense. What it shows is a disordered matrix — collagen less organised, more ground substance, new small vessels growing in with nerves alongside them.

If that is the state, then the pain is not a simple gauge of ongoing tissue destruction. And it shows: tendons that look badly abnormal on imaging are common in people with no symptoms at all, and symptoms can resolve while the images stay abnormal. So "it still hurts" is genuinely poor evidence that the load is doing damage. Do you see how much of the intuition rests on assuming pain and damage are the same reading?

Three corrections, and one substantial admission.

The first correction: none of this says load is always good or rest always bad. The tissue's tolerance is finite, and the same evidence that shows loading builds tendon shows that sudden large increases in load provoke it. The claim is about the shape of the relationship, not its direction.

The second: this reasoning concerns chronic tendinopathy, not an acute rupture, which is a different injury with different handling. Lumping them is the commonest way this argument gets misused.

The third: everything above describes averages over studied populations. A trial reporting that a loading protocol beat rest is reporting a group mean with wide spread inside it, and says remarkably little about any particular tendon, sport, or person. This is an explanation of a mechanism, not guidance for anyone's injury.

Now the admission, which matters more than any of them. Why loading helps is genuinely contested. At least five candidates are live: it remodels the matrix; it raises tendon stiffness and load tolerance; it restores the muscle capacity that shields the tendon; it has direct effects on pain processing; and graded exposure reduces avoidance behaviour independently of tissue change. These are not mutually exclusive, and no one has cleanly apportioned them.

Training lore has run ahead of that uncertainty before. For years it was near-doctrine that eccentric contractions specifically — lengthening under load — were the therapeutic ingredient, and elaborate mechanistic stories were built on it. When other heavy loading protocols were compared directly, the special status did not hold up well. The practice survived; the mechanism story attached to it did not. That is worth noticing as a pattern, not just an anecdote.

What would falsify the core mechanotransduction claim? It predicts that removing load measurably degrades the tissue — lower stiffness, smaller cross-section, reduced collagen synthesis — and controlled immobilisation studies find exactly that; had immobilised tendon held its properties, the account would be finished. And here is the refuting observation that has already bitten a narrower version of the story: if loading worked by repairing the damaged matrix, then symptom improvement ought to track improvement on imaging. It largely does not. That kills the tidy "load rebuilds the tissue, so the pain goes" explanation while leaving the broader finding standing — which is exactly the kind of result that should make anyone cautious about the confident version.

• How bone answers the same mechanical question with a different cell system, on a different timescale.
• Why vessel and nerve ingrowth might explain pain that outlives the matrix disorder.

Every term the collection defines is gathered in the glossary.