Exercise and Inflammation: A Real Effect, an Uncertain Size
Two separate research teams published network analyses this year asking whether exercise lowers chronic inflammation in older adults. One pooled 54 trials, the other 55. They agree that it does. They disagree about nearly everything after that — and neither can tell you the dose.
AGE 77 — The pooled average age in the newest of the two analyses — 54 randomized trials, 2,836 people, and 41 of those trials run in people living with a chronic condition. For once the research is actually about the people reading it.
“Inflammation” is the most profitable word in the supplement aisle. It is printed on turmeric, fish oil, tart cherry, collagen, mushroom powder and about nine hundred teas. It is also a real thing — the slow, low-grade, background version that creeps up with age and tracks with almost every condition we would rather avoid. Three papers below. The first two are about whether training moves it. The third is about what happens when you try to shortcut it with a bottle, and it does not end the way the internet says it does.
Lifting moved the marker
Int. Journal of Environmental Research and Public Health, 2022
Eighteen randomized controlled trials, 539 adults, every one of them 60 or older. The outcome was C-reactive protein — CRP, the blood marker your doctor may already be drawing, and the one most often used to stand in for background inflammation.
Resistance training lowered it: a standardised effect of −0.72, confidence interval −1.06 to −0.38. That interval sits well clear of zero, which is the part that matters. TNF-alpha, a second inflammatory signal, also came down (−0.56). Interleukin-6 only trended downward and did not reach significance.
Two honest limits. First, −0.72 is a standardised effect size, not milligrams per litre — I cannot hand you a number for your own blood test, and anybody who does is inventing it. Second, 539 people is a thin base. The authors also found the CRP reduction held regardless of age, training method, number of exercises, intensity, weekly frequency or duration — which sounds encouraging and is actually a warning. Finding an effect that does not change with dose usually means the data are not fine enough to see the dose.
The newest look — and it argues with the other newest look
Frontiers in Medicine, August 2026
Fifty-four randomized trials, 2,836 people, average age 77, and 41 of those trials in people living with a chronic condition. Published August 2026. It ranked every mode of exercise against every other.
Aerobic training came first for CRP (−0.73, interval −1.17 to −0.29), resistance training second (−0.46, −0.84 to −0.08). Combined training, mind–body work and multi-component programs all had intervals that crossed zero — meaning the data could not rule out no effect at all.
Now the part that complicates my week. A second network analysis published this year in the Journal of the American Medical Directors Association — 55 trials, 3,516 people, searched through September 2025 — concluded the opposite ranking: that resistance training most effectively reduced C-reactive protein, while aerobic training produced the greatest improvements in the other two markers. Two 2026 analyses of overlapping literature, reversed podiums. When that happens, the honest statement is not “lifting wins” or “walking wins.” It is both show an effect and the ranking is unsettled.
And the dose is worse than unsettled. This paper is the only one that tried to model how much exercise is required, and its own authors report that the credible interval crossed zero at every single dose level. They call their thresholds “hypothesis-generating dose regions rather than definitive clinical cut-points,” and rate the overall evidence quality as predominantly low to very low. The programs inside these trials ran anywhere from six weeks to twelve months.
The antioxidant study everyone quotes wrong
The Journal of Physiology, 2014
Here is the logic the supplement aisle runs on: training creates oxidative stress, oxidative stress is inflammation, so take a big antioxidant dose and you will recover better. A Norwegian group tested it properly, twice, and the result is the one you will see quoted in every gym on earth — usually with the second half missing.
Trial one: 32 recreationally strength-trained adults, 10 weeks of heavy lifting four times a week, randomly given either 1,000 mg of vitamin C plus 235 mg of vitamin E daily or a placebo, double-blind. The supplement did interfere with the muscle’s signalling after a session — p38 MAPK phosphorylation was blunted. But on the outcomes a person can feel, the authors write that “the supplementation did not affect the increase in muscle mass” and that it “hampered certain strength increases (biceps curl).” Muscle grew the same in both groups. One strength measure out of several came up short. I am not going to soften their own conclusion: they call it evidence of “tentative long-term negative effects on adaptation to strength training.” Tentative is their word, and it is doing real work in that sentence — but it is not nothing, and you should hear it from me rather than from a headline.
Trial two ran the same doses through 11 weeks of endurance training in 54 young men and women, three to four sessions a week. Mitochondrial markers inside the muscle were blunted — and VO2 max rose by 8 ± 5% in the supplement group and 8 ± 5% in the placebo group, with running performance not measurably different either.
So what it actually shows. Mega-dosed antioxidants change what happens inside the cell. They did not change what happened to the person. “Vitamin C ruins your gains” is not what these papers found. Two more limits worth saying out loud: these were young adults — recreationally trained in the first trial — not people in their seventies — I am not allowed to hand you their result and pretend it is about you — and 1,000 mg is roughly eleven times the daily requirement. This says nothing about vitamin C in an orange, or a normal multivitamin.
Matt's take
Read this first if you take anything for an inflammatory condition. Rheumatoid arthritis, polymyalgia rheumatica, inflammatory bowel disease, an autoimmune diagnosis, long-term steroids, a biologic, a prescribed anti-inflammatory — nothing in this article is a reason to reduce, delay or stop any of it. “Exercise is anti-inflammatory” is a statement about a blood marker moving in a group average over months. It is not a substitute for a medication that is holding a disease down, and I do not want anyone reasoning their way off a prescription because of a newsletter. That is your rheumatologist’s call, full stop.
And one specific supplement warning. High-dose vitamin E is not harmless: it can increase bleeding risk, which matters if you take a blood thinner, take daily aspirin, or have surgery of any kind coming up. High-dose fish oil and turmeric or curcumin sit in the same category. If you take any of them, tell your physician — and tell me, so I know what is in the picture.
Here is the whole thing in one line: the only anti-inflammatory intervention in this article with randomized trials behind it is the training itself. The bottles have the word. The work has the data.
Now the part that costs me something to write. I cannot give you a dose. I would love to tell you that three sessions a week for twelve weeks lowers your CRP by some specific amount, because that is a fantastic thing to sell. The best dose–response analysis anyone has done on this — study two, published in August 2026 — reports that its credible interval crossed zero at every dose it modelled, and calls its own numbers hypothesis-generating. The one before it found the effect held regardless of frequency, intensity or duration, which means it could not see a dose either. Two 2026 papers cannot even agree on whether walking or lifting comes first. Any trainer, clinic or supplement brand handing you a precise anti-inflammatory prescription is making up the precision. What the evidence supports is direction, in older adults, measured in months. That is less than I would like to promise you and it is what is true.
If you’re at a desk all day: the version of this that applies to you is not exotic. You are not going to out-supplement ten hours of sitting, and the trials that found anything ran for months, not weeks. The lever you control is showing up in the same week in October that you showed up in September. In New England the sun sets before 6:45 in mid-October and before 4:30 by Christmas — that is the thing that quietly ends training years, not a missing capsule. Protect the appointment.
If you’re over 70, or managing a long-term condition: study two is genuinely about you — average age 77, and 41 of its 54 trials were run in people with a chronic condition, which is unusual and makes it worth more than most research. Its finding that aerobic work ranked at least as high as lifting is not an argument against in-home strength training. It is an argument for the walk as well. If you already strength train, the cheapest addition available to you is not a purchase — it is a walk on the days you are not lifting, at whatever pace and distance your body and your doctor allow.
On the aisle, practically: I am not telling you to throw anything out. Food that is genuinely associated with lower inflammation is boring and familiar — fish, olive oil, vegetables, beans, nuts — and that pattern has far better evidence than any pill extracted from it. Where I do push back is on paying $40 a month for the word on the label while skipping sessions. That is buying the sticker instead of the thing.
I am a strength coach, not your physician. If you have an autoimmune or inflammatory diagnosis, kidney or liver disease, a bleeding disorder, or you are on a blood thinner or any prescribed medication, their instructions outrank every word above.
Train this at home
Forged Home Fitness brings one-on-one training to your home in Wellesley, Weston, Newton, Needham, Groton and the surrounding towns, built around your joints, your history and your schedule. Book a free in-home evaluation.
This article first ran in the Forged Home Fitness weekly newsletter. It is general fitness education, not medical advice, and does not replace your physician.
