Strength training vs cardio

What the research found.

• Weight training lowered DAS-28 disease activity by 0.69 standard deviations, with the true value running from minus 1.26 to minus 0.11. That came from seventeen trials and 1,010 patients with rheumatoid arthritis. All of them stayed on their usual treatment.

  Wen and Chai, Medicine, 2021

• In that review, the sedimentation rate fell by 0.86 standard deviations and fifty-foot walking time fell by 0.64. Pain came out at minus 0.61. Its range crossed zero, though, so pain didn't reliably improve.

  Wen and Chai, Medicine, 2021

• One trial gave 12 weeks of interval training in rheumatoid arthritis, where VO2 max improved by 3.71 milliliters per kilogram per minute. The comparison group got advice to be moderately active rather than nothing at all. Disease activity showed no clear improvement.

  Bilberg and colleagues, British Journal of Sports Medicine, 2024

• One review compared exercise types across 34 trials and 2,435 patients, and cardio plus weight work came first for morning stiffness. It shortened stiffness by 8.23 minutes. Its ranking score was 100 percent.

  Zhang and colleagues, Journal of Pain Research, 2025

• The 2018 EULAR recommendations looked at how much exercise to advise. They judged the ordinary public health targets fit for people with inflammatory arthritis, and they didn't set any lower target for them. That's more generous than most people expect.

  Rausch Osthoff and colleagues, Annals of the Rheumatic Diseases, 2018

The short answer.

Weight training is the one to start with if you have to pick. It's the only one of the two with evidence for lowering disease activity, and across seventeen trials and 1,010 patients it moved DAS-28 by minus 0.69 and the sedimentation rate by minus 0.86. It improved walking speed as well.

Interval cardio wins on the outcome weight work doesn't touch. Twelve weeks of it raised VO2 max by 3.71 milliliters per kilogram per minute against general activity advice, and in that trial disease activity showed no clear improvement. That isn't a failure, because heart risk is raised in nearly every inflammatory rheumatic condition and fitness is one of the few parts of it you can change.

Neither one reliably improves pain. In the weight training review, pain came out at minus 0.61 with a range that crossed zero, and pain is the outcome most people hope to move. So that's worth knowing before you start, because expecting pain relief and getting fitness instead reads as failure when it isn't.

Side by side.

What each one is tested on

The two kinds of exercise have been studied against different outcomes, which is most of why the comparison gets muddled. Weight training in rheumatoid arthritis has been pooled across seventeen randomized trials holding 1,010 patients, with disease activity as the headline result. Interval cardio has been tested against fitness instead, and its comparison group got general advice to be moderately active rather than nothing at all, which raises the bar it had to clear.

So a straight contest between them barely exists in the research. What exists is good evidence about what each one changes, which is a more useful thing to have anyway. The question worth asking isn't which is better, it's which outcome you want moved.

Both were tested in people already getting usual care, so nothing on this page describes exercise replacing treatment. No trial in either group was built to test that. Read all of it as what to add rather than as what to swap.

What weight work changes

Disease activity is the striking one. Across those seventeen trials, DAS-28 fell by 0.69 standard deviations, with the true value running from minus 1.26 to minus 0.11. That's a real effect on the combined score rheumatologists use to decide whether treatment is working.

The sedimentation rate moved too, falling by 0.86 standard deviations, from minus 1.65 to minus 0.07. So an inflammation marker moved alongside a disease activity score. That pair is more persuasive than either result on its own, because the two measure different things and agreeing with each other is harder than agreeing with yourself.

Function improved as well, with fifty-foot walking time getting better by 0.64 standard deviations. Walking speed is a dull outcome. It predicts a great deal about staying independent over years, though, so it's worth a good deal more attention than it usually gets, and it's the outcome most likely to still be there in twenty years.

What interval cardio changes

Fitness, and by a lot. One 12-week trial in rheumatoid arthritis improved VO2 max by 3.71 milliliters per kilogram per minute, and the comparison group wasn't sitting still, because they were advised to be moderately active for at least 150 minutes a week. So this is planned hard training, measured against the advice most people already get.

Disease activity showed no clear improvement in that trial, and that's information rather than disappointment. Heart risk is raised across inflammatory rheumatic disease, with the inflammation raising part of it and the treatments raising the rest. Fitness is one of the few parts of that risk a person can change directly.

The trial ran in well-controlled disease, in people aged twenty to sixty, and that's the group it describes. So stretching it to somebody in a flare goes beyond what was tested, and so does stretching it to somebody with a lot of joint damage. The safety finding belongs to the group that was studied.

The outcome neither of them reliably moves

Pain is what most people want from exercise, and it's the weakest result here. In the weight training review, pain came out at minus 0.61, with the true value running from minus 1.49 to 0.27, and that range crosses zero. A result that crosses zero can't be told apart from no effect at all.

That doesn't mean exercise never helps anybody's pain. It means the average effect across the pooled trials wasn't reliable, while the effects on disease activity, inflammation, and function all were. Those are different claims about different things, and the difference between them is worth holding onto when somebody quotes one at you.

So don't make pain relief your test of whether exercise is working, because the evidence predicts you'll fail that test. Function, stiffness, and fitness are more likely to reward the effort you put in, so track those instead of the one thing that's least likely to move. Judge it on those and the picture looks quite different.

Where doing both wins

Morning stiffness is the clearest case for doing both. One review covered 34 trials and 2,435 patients and ranked cardio plus weight work first for how long stiffness lasts, shortening it by 8.23 minutes at a ranking score of 100 percent. No other review on this page compares exercise types head to head.

That review compared ten exercise types against each other rather than each one against nothing, and that's the design that answers a which-one question. So its stiffness result is the most direct evidence here that the two add to each other. They don't stand in for each other.

For how much to do, the 2018 EULAR recommendations answered it, judging the public health targets fit for people with inflammatory arthritis and setting no lower target of their own. So that's the advice given to everybody else. It's more generous than most people expect to be told, and nobody is being asked to hold back.

When each one fits.

• Start with weight work if disease activity and function are what you want to change, because that's where its evidence is.
• Put interval cardio first if heart risk is the worry, which covers most inflammatory rheumatic disease, or if fitness rather than joint counts is the goal.
• Do both if morning stiffness is your worst symptom, because doing both ranked first for it across 34 trials.
• Expect neither one to be a reliable pain treatment, and treat pain relief as a welcome extra rather than the reason to start.
• The activity target is the one given to everybody else, which is more generous than most people expect, and it's what the 2018 EULAR recommendations concluded.

Questions patients ask.