Inflammation

The question this branch answers

Training produces an inflammatory and oxidative response. A large recovery industry exists to reduce it, on the shared premise that the response is a cost to be minimized.

The interesting finding across this branch is not that the premise is wrong. It is that in most cases the intervention is not doing what the premise says it does. Cold water does not measurably reduce inflammation inside the muscle, and it costs you something anyway. High-dose antioxidants move the signaling proteins and leave the outcome roughly where it was. Soreness, the sensation people use as their proxy for the whole process, does not scale with the damage it is standing in for.

So the question this branch answers is not "is inflammation good or bad." It is narrower and more useful: when you take a recovery intervention, what is changing?

What is settled

The response carries information. Current reviews treat post-exercise inflammation as integral to adaptation rather than as an obstacle, provided it is regulated and resolves on schedule (Peake et al., 2017). This is a shift from the older framing, and it is the reason blunting interventions are worth examining at all.

Suppression and resolution are different things. An inflammatory response that is switched off early is not the same as one that completes and clears. Most of what the recovery industry sells is aimed at the first. The evidence that matters is about the second.

Direct measurement is scarce and it disagrees with the background story. Where researchers have taken muscle biopsies rather than measuring surrogates, the results have repeatedly failed to confirm the mechanisms stated in the introductions of the papers citing them. A narrative review that tallied this literature for resistance training found most studies reporting no effect on inflammatory markers, several reporting a larger response, and one a reduction (Petersen and Fyfe, 2021).

Population matters more than the marketing admits. The same intervention reads as a cost for one training goal, as neutral for another, and as a clinical necessity in a third context. This is not hedging. It is the shape of the evidence.

The measurement is harder than the question. A recurring obstacle across this branch is that the available tools cannot cleanly separate the things being argued about. Ultrasound and magnetic resonance imaging (MRI) cannot distinguish a larger muscle from a swollen one. Circulating markers are not the same as what is happening inside the tissue. Soreness is a report, not a measurement. Several of the disputes below are, underneath, disagreements about what an instrument was capable of seeing.

Timing is the least-studied variable and often the most important one. Almost every trial in this branch applied its intervention immediately after training, because that is the convenient and conventional design. The question most readers have is whether separating the intervention from the session avoids the cost. That question has been asked far less often than it has been answered.

Where it splits

Cold water immersion. The assumption underneath the ice bath is that cold reduces inflammation. Measured in human muscle, it does not, and the growth cost that has been observed appears to travel through a different route. Soreness relief is real and separately supported. (→ Inflammation 01)

Delayed-onset muscle soreness. The most-used feedback signal in training, measuring something other than what it is read as measuring. Whether damage contributes to growth at all is actively contested, on the record, between named researchers. (→ Inflammation 02)

Antioxidant supplements. The clearest case of a signal-level finding being reported as an outcome-level one. The cellular markers move; the muscle and the aerobic capacity mostly do not. And the blunting that does occur appears to track dose, breadth and timing more than the compound, a pattern no trial has yet tested directly, which makes it a narrower claim than the category name suggests. (→ Inflammation 03)

Anti-inflammatory drugs. The cleanest test of the premise, because those drugs unambiguously do suppress inflammation. Of eleven training trials, two found more muscle growth with the drug, one found less and eight found no difference, with too many differences between them to say why. (→ Inflammation 04)

A pattern runs through all four. In each case the popular claim was built from a mechanism, and in each case the direct human measurement arrived later and came out weaker than the mechanism promised.

That is not a scandal, and it is not a case of anyone being dishonest. It is how mechanistic reasoning normally fails. A mechanism is a story about why something should work, and a good one is informative about where to look. What it cannot do is tell you the size of the effect, or whether the effect survives contact with a whole organism doing a whole training program. The gap between those two things is where most of this branch lives.

The practical consequence is that the four entries above reach different verdicts despite starting from the same premise. If the premise were doing the work, they would agree.

What this branch does not cover

Cold exposure as a daily habit. Deliberate cold exposure away from training is a separate literature with a separate population and separate outcomes. One finding from it is worth flagging here because it cuts against the popular framing. In a meta-analysis of randomized trials in the general population that specifically excluded pre- and post-exercise immersion, cold immersion raised inflammation immediately and one hour afterward, and reduced stress twelve hours later. Effects on immune function were not significant. Sleep quality and quality of life improved; mood did not (Cain et al., 2025).

Acute injury and heat illness. Cooling in the management of an acute injury, and cooling in the treatment of exertional heat stroke, are different problems with different evidence bases. Conclusions from the training literature do not transfer to them, in either direction.

Nutrition and sleep. The two interventions with the strongest claim on recovery are not in this branch, because they are not framed as inflammation management. Each has its own. (→ Sleep 01: Sleep and muscle; Nutrition sub-hub)

How to use this branch

If you came here to find out whether you should stop doing something, the answer in every case below is conditional on a goal you have not told us. We have tried to write the conditions out rather than pick one for you.

If you came here to find out what is true, the most defensible summary is this. Post-exercise inflammation is part of the adaptive process rather than a waste product. Most of the interventions marketed as suppressing it do not measurably suppress it. Some of them carry a real cost anyway, by a route that has nothing to do with inflammation. And the sizes involved are smaller than the confidence with which they are usually stated.