Exercise

Ankylosing spondylitis and exercise

What exercise does in ankylosing spondylitis, which kinds have been compared against each other, and why the biggest effect is also the least certain one.
Quick answerExercise lowers disease activity in ankylosing spondylitis, and a network analysis of 48 trials ranked aquatic stretching first on the BASDAI, a patient-reported disease activity score scored 0 to 10, at minus 1.42 points. Land aerobic exercise came second at minus 0.94, and land stretching third at minus 0.49. The biggest effect there is also the least certain one.
Read the intervals, not the ranking, because they're telling you two different things. Aquatic stretching tops the table at minus 1.42 BASDAI points, on an interval ranging from minus 2.51 to minus 0.33, so the evidence is consistent with an effect three times the size of land aerobic exercise or with a third of it. Land aerobic exercise is at minus 0.94, from minus 1.41 to minus 0.47, which is a narrower spread around a more dependable number. A ranking built out of few trials tends to put the least studied option on top.

What the research found.

  • A network meta-analysis pooled 48 studies and 3,140 participants with ankylosing spondylitis, comparing exercise types against each other. Aquatic stretching lowered the BASDAI, a patient-reported score of disease activity scored 0 to 10, by 1.42 points, on a 95 percent confidence interval from 2.51 to 0.33. That's the largest effect in the analysis and the least certain one of the three.

    Kong and colleagues, PeerJ, 2025

  • In that analysis land aerobic exercise lowered BASDAI by 0.94 points, on an interval from 1.41 to 0.47. Land stretching lowered it by 0.49 points. The smaller effect is the one measured more tightly, which is the pattern worth noticing across the whole table.

    Kong and colleagues, PeerJ, 2025

  • A separate meta-analysis and meta-regression of randomized trials in axial spondyloarthritis looked at function rather than at disease activity. Exercise therapy improved the BASFI, a patient-reported functional index also scored 0 to 10, by a weighted mean difference of 0.49 points. BASFI asks how easily you manage everyday movements like bending, reaching, and standing, so that figure is about function rather than about how much pain you're in on a given day.

    Zhang and colleagues, Archives of Physical Medicine and Rehabilitation, 2025

  • The 2018 EULAR recommendations judged public health physical activity targets applicable to people with inflammatory arthritis. They didn't set a reduced target for them, which is a more generous position than many people expect to find. That's a guideline position rather than a trial result.

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

  • A randomized trial of aqua stretching and aqua Pilates in ankylosing spondylitis reported significant improvement in pain, at a P value of 0.0001. Its abstract doesn't report the group sizes anywhere in it. So the size of that improvement is unknown, and a very small P value says nothing about how large a difference was.

    Gandomi and colleagues, BMC Sports Science, Medicine and Rehabilitation, 2022

The movements, and the trial behind each block.

This is the best evidenced exercise tool on the site, because axial spondyloarthritis has been studied harder than the rest of these diseases put together. Five tabs, each one built from the trials that tested that modality, with the figures given where the analysis reports them. The demonstration videos are her own, and not one of them loads until you press the button. The trials below score their results on a handful of standard patient-reported measures, each scored roughly 0 to 10, and a fall in any of them is the improvement to look for. BASDAI covers disease activity, BASFI covers everyday function like bending and reaching, and BASMI covers how far your spine physically moves. ASDAS is newer and folds a blood marker in alongside what you report yourself.

One meta-analysis and meta-regression pooled twenty randomized trials covering 1,670 patients with axial spondyloarthritis. Function on the BASFI improved by 0.49 points, on a 95 percent interval from 0.65 to 0.32, and spinal mobility on the BASMI improved by 0.49 on an interval from 0.87 to 0.11. Disease activity improved on both the measures used for it, at 0.78 on the BASDAI and 0.44 on the ASDAS.

Fitness and endurance moved further than the disease scores did. Peak oxygen consumption rose by 3.16, on an interval from 1.37 to 4.94, and the six-minute walk test improved by 27.64 metres on an interval from 12.04 to 43.24. Pain improved by a standardized mean difference of 0.47 and fatigue by 0.49, while chest expansion and both inflammatory blood markers showed no significant benefit at all.

That last clause is the honest limit on what any of this shows. Symptoms, function, mobility, and fitness improved, and the sedimentation rate and the C-reactive protein did not move. Exercise trials cannot blind the people in them either, so the patient-reported outcomes here come with the expectations of people who knew which group they were in.

Bodyweight only, and this tier builds the five patterns that get loaded in the tier above it. Morning stiffness is the usual obstacle to training early, and five minutes of walking before the first set deals with most of it. Pair this tab with the spinal mobility tab rather than choosing between them, because they do different jobs.

This tier matches the moderate-intensity protocols in the trials feeding the pooled analysis above. Progress here once the bodyweight work is comfortable and your morning stiffness is manageable rather than merely tolerable. Finding your working weight takes one honest test: if the heaviest single squat you can do with good form is fifty pounds, your working weight is thirty to thirty-five, and it needs retesting every six to eight weeks.

The anchor trial used supervised high-intensity training, combining cardiovascular and strength work three times a week for twelve weeks. At these loads, train two or three times a week with full recovery between sessions, and use a rack with the safety bars set. Read the note below before you decide these figures are evidence about you.

Where the evidence doesn't exist

Most trials in this disease did not separate people by how far it had progressed. Radiographic and non-radiographic disease, the degree of fusion, and baseline spinal mobility are all left pooled together. So this evidence applies most clearly to a mobile or early spine, and much less clearly to a fused one, and significant fusion is a reason to build this with a physiotherapist rather than from a page.

One network and dose-response meta-analysis compared twelve exercise interventions across thirty-two randomized trials and 1,757 participants. Aerobic exercise combined with Pilates came out best for disease activity on the BASDAI and for spinal mobility on the BASMI. The dose-response relationships in that analysis were non-linear, which means each outcome has its own effective range rather than more always being better.

One randomized trial tested Pilates on its own in fifty-five patients, an hour three times a week for twelve weeks. Function on the BASFI improved in the Pilates group at week twelve and again at week twenty-four, and did not change significantly in the controls. Between the two groups, the Pilates group was better at week twenty-four, and its authors call it the first clinical study of Pilates in this disease.

One network meta-analysis of forty-eight studies and 3,140 participants ranked water-based stretching first on three of its four outcomes. Disease activity on the BASDAI improved by 1.42 points, on an interval from 2.51 to 0.33, and function on the BASFI by 1.74 on an interval from 2.45 to 1.04. Quality of life on the ASQoL improved by 3.67, on an interval from 6.17 to 1.18.

Read those intervals. Don't read the ranking, because the top-ranked option is also the least certainly measured one. Land aerobic exercise improved the BASDAI by 0.94 on an interval from 1.41 to 0.47, which is a narrower spread around a slightly smaller number. Fewer trials produce wider intervals and higher point estimates more often than people expect, so a pool is worth using if you enjoy it, and not because a ranking says so.

Three randomized trials in the corpus tested water-based programs directly. One compared aqua Pilates and aqua stretching against routine treatment in forty patients over six weeks, and both water arms improved pain, function, walking speed, quality of life, and spinal range while the control arm improved in not one of them. Two others tested aquatic programs against no supervised exercise and against a home exercise booklet, and this site holds no results from either.

Two or three times a week in warm water, at thirty to thirty-four degrees Celsius, which is the temperature the trial used. Warm water reduces muscle spasm and morning stiffness, which are the two things most likely to stop a land session before it starts. Buoyancy takes load off the spine while the water itself supplies the resistance.

Spinal mobility on the BASMI improved by 0.49 points across the pooled trials, on an interval from 0.87 to 0.11. The network analysis of thirty-two trials put aerobic exercise combined with Pilates top for that outcome. The 2018 EULAR recommendations judged the general public activity targets applicable to people with inflammatory arthritis rather than setting a reduced target for them.

This tab is not a modality and it is not competing with the others. It is the daily maintenance that preserves whatever range you have today, done every morning against overnight stiffness and again before bed. Fifteen to twenty minutes, and the cost of skipping it is measured in years rather than in sessions.

Fatigue improved across the pooled trials by a standardized mean difference of 0.49, on an interval from 0.71 to 0.27. Long stretches of sitting make the stiffness worse in a disease that is already producing it. So breaking the sitting up is a reasonable thing to do here even though the trial on it has never been run.

The nearest evidence comes from a different disease and it transfers on reasoning rather than on proof. One study measured seven days of movement in sixty-one patients with rheumatoid arthritis and found that sedentary time, and stretches of twenty minutes or longer, both correlated with a higher ten-year cardiovascular risk score. Both associations held whatever the person's moderate to vigorous activity added up to.

What the evidence covers

Exercise in ankylosing spondylitis has been studied more than in most rheumatic conditions, and most of those studies are small ones. The largest synthesis is a network meta-analysis of 48 studies and 3,140 participants, which compared exercise types against each other rather than each of them against no exercise. That design answers the question people ask first, which is which kind to do.

A separate systematic review, meta-analysis, and meta-regression of randomized trials in axial spondyloarthritis reports exercise therapy improving the BASFI functional index by a weighted mean difference of 0.49 points. It also assessed disease activity, spinal mobility, and walking capacity. Both papers are from 2025 and they end up in broadly one place, which is that exercise helps and that the size of the help depends on what you choose to measure.

The third document in this set is a guideline rather than a trial. The 2018 EULAR recommendations for physical activity judged public health activity targets applicable to people with inflammatory arthritis rather than setting a reduced target for them. That's a position on how much, and it predates the trials above rather than resting on them.

Which kind of exercise, and why the ranking misleads

The network analysis puts aquatic stretching first for disease activity, where it lowered BASDAI by 1.42 points. Land aerobic exercise comes second at 0.94, and land stretching third. Read as a ranking and nothing else, that says get yourself into a pool as soon as you can.

Read as intervals, it says something a good deal more careful. The aquatic stretching estimate ranges from 2.51 to 0.33, so the evidence is consistent with an effect three times the size of land aerobic exercise or with an effect a third as large. Land aerobic exercise ranges from 1.41 to 0.47, which is a much tighter spread around a slightly smaller number.

That difference in width is usually about how many trials contributed rather than about how good the exercise is. Fewer trials produce wider intervals and, often enough, higher point estimates as well. A ranking built out of a thin evidence base tends to put the least studied option on top, which is worth knowing before you rearrange your week around a pool.

What one point on BASDAI means

BASDAI is a questionnaire scored from 0 to 10, covering fatigue, spinal pain, peripheral joint pain and swelling, areas of local tenderness, and morning stiffness in both severity and duration. Almost every trial in this condition uses it as the disease activity outcome, which is why the numbers above are all expressed in BASDAI points. You'll see it on your own clinic letters, so the scale is worth recognizing.

A change of roughly one point is generally treated as clinically meaningful. That puts land aerobic exercise just under the line at 0.94 and aquatic stretching above it at 1.42, with the caution about intervals still applying to both of them. Land stretching at 0.49 is a real effect that most people would struggle to notice on its own.

That doesn't make the smaller effects worthless. Exercise improves function, pain, and cardiovascular risk alongside disease activity, and those four don't all move together. The disease activity score is one outcome among several rather than the whole account of what exercise does.

Weeks and months, not decades

No study in this evidence base ran long enough to say anything about structural progression. Ankylosing spondylitis can cause new bone formation and spinal fusion over years, and trials lasting weeks to months can't detect a change in that. Exercise wasn't shown to accelerate it and it wasn't shown to slow it, so both of those are absences rather than findings.

Every trial tested exercise alongside usual care rather than in place of it. So nothing here tells you what happens if you exercise instead of taking treatment, because that study has never been run. Exercise is worth doing for what it demonstrably improves, and the trials are clear about what that is.

One trial of aqua stretching and aqua Pilates reported significant improvement in pain, at a P value of 0.0001, and its abstract gives no group sizes. A very small P value tells you a difference is unlikely to be chance. It tells you nothing at all about how big the difference was, so that trial supports the direction and adds nothing to the magnitude.

Questions patients ask.

Which kind of exercise works best?

A network analysis of 48 trials ranked aquatic stretching first for disease activity, at minus 1.42 BASDAI points, ahead of land aerobic exercise at minus 0.94 and land stretching at minus 0.49. The ranking is less settled than it sounds, because the option on top has the widest confidence interval of the three. Land aerobic exercise has the more dependable number, and it's the one most people can do without a pool.

How much should I be doing?

The 2018 EULAR recommendations judged general public health physical activity targets applicable to people with inflammatory arthritis, rather than setting a reduced target. That means the guidance given to everybody else applies to you, which is more generous than most people expect. How you get there is a conversation for you and your physiotherapist, who can see what your spine does and what it won't do yet.

Will exercise make my spine fuse faster?

Nothing in this evidence base suggests that exercise does that. The trials measured disease activity, function, spinal mobility, and pain, and exercise improved all four of them. No trial here ran long enough to establish an effect on structural progression in either direction, because these studies lasted weeks and months rather than decades. That's a question the research hasn't answered rather than a warning about exercise itself, and it's worth stating as one.

Does swimming really beat everything else?

Not on the strength of the evidence here. Aquatic stretching came top of the network ranking on an interval ranging from minus 2.51 to minus 0.33, which is consistent with a very large effect or a fairly small one. Fewer trials produce wider intervals and higher point estimates more often than people expect. If you've got access to a pool and you enjoy it, that's a good enough reason to use it.

What does BASDAI measure?

It's a questionnaire scored from 0 to 10, covering fatigue, spinal pain, joint pain and swelling, areas of local tenderness, and morning stiffness in both severity and duration. Most trials in this condition use it as their disease activity outcome. A change of about one point is generally treated as clinically meaningful, which puts the aquatic and land aerobic figures on either side of that line.

Does exercise replace my medication?

No, and every trial in this evidence base tested exercise alongside usual care rather than instead of it. So nothing here tells you what happens if you stop treatment, and that study has never been run. Ankylosing spondylitis can cause structural damage that doesn't reverse, and exercise is worth doing for what it demonstrably improves, which is disease activity, function, and pain.

References.

  1. Kong L; Yu C; Wang C et al. Comparative efficacy of different exercise interventions in patients with ankylosing spondylitis: a systematic review and network meta-analysis. PeerJ. 2025;13:e20336. 10.7717/peerj.20336Systematic review and network meta-analysis of 48 studies published between 2002 and 2024
  2. Zhang M; Liang Z; Tian L et al. Effects of Exercise Therapy in Axial Spondyloarthritis: A Systematic Review, Meta-analysis, and Meta-regression of Randomized Trials. Archives of Physical Medicine and Rehabilitation. 2025;106:113-123. 10.1016/j.apmr.2024.06.005Systematic review
  3. Rausch Osthoff A; Niedermann K; Braun J et al. 2018 EULAR recommendations for physical activity in people with inflammatory arthritis and osteoarthritis. Annals of the Rheumatic Diseases. 2018;77:1251-1260. 10.1136/annrheumdis-2018-213585EULAR recommendations for physical activity in people with inflammatory arthritis and osteoarthritis
  4. Gandomi F; Soufivand P; Ezati M et al. The effect of Aqua Stretching exercises and Pilates on pain, function and spine posture in patients with ankylosing spondylitis: a randomized controlled trial. BMC Sports Science, Medicine and Rehabilitation. 2022;14. 10.1186/s13102-022-00577-0Randomised controlled trial with three arms
  5. Ye G; Gao L; Liu C et al. The effectiveness of different exercise mode interventions in improving disease activity in patients with ankylosing spondylitis: a network and dose-dependent meta-analysis. Frontiers in physiology. 2026;16:1715944. 10.3389/fphys.2025.1715944Network and dose-response meta-analysis of 32 randomized controlled trials
  6. Sveaas SH; Dagfinrud H; Berg IJ et al. High-Intensity Exercise Improves Fatigue, Sleep, and Mood in Patients With Axial Spondyloarthritis: Secondary Analysis of a Randomized Controlled Trial. Physical therapy. 2020;100:1323-1332. 10.1093/ptj/pzaa086Secondary analysis of a multicentre RCT
  7. Altan L; Korkmaz N; Dizdar M et al. Effect of Pilates training on people with ankylosing spondylitis. Rheumatology international. 2012;32:2093-2099. 10.1007/s00296-011-1932-9Randomized
  8. Dundar U; Solak O; Toktas H et al. Effect of aquatic exercise on ankylosing spondylitis: a randomized controlled trial. Rheumatology international. 2015;34:1505-1511. 10.1007/s00296-014-2980-8RCT
  9. Fernández García R; Sánchez Sánchez LdeC; López Rodríguez MdelM et al. Efectos de un programa de ejercicio físico y relajación en el medio acuático en pacientes con espondiloartritis: ensayo clínico aleatorizado. Medicina clinica. 2015;145:380-384. 10.1016/j.medcli.2014.10.015Randomized single-blind trial
  10. Fenton SAM; Veldhuijzen van Zanten JJCS; Kitas GD et al. Sedentary behaviour is associated with increased long-term cardiovascular risk in patients with rheumatoid arthritis independently of moderate-to-vigorous physical activity. BMC musculoskeletal disorders. 2018;18:131. 10.1186/s12891-017-1473-9Cross-sectional analysis of baseline data from 61 patients in the Physical Activity in Rheumatoid Arthritis study

This page gathers the published research on this subject into one place. The studies behind it were published between 1989 and 2026, and every figure links to the paper it came from. Those studies were peer reviewed. This summary of them was not. Dr. Sarah Luebker is reviewing these pages one at a time and has not reached this one yet, so it carries no medical review date and nothing here is her opinion or her advice to you. Each page gets updated as she reaches it. It is here in the meantime because the science is worth having in one organized place that is easy to find and easy to read. Talk to your own clinician before acting on any of it.