Ankylosing spondylitis and environment

What the research found.

• One study of 1178 patients looked at how much people smoked rather than only whether they did. Against non-smokers, smoking for at least 15 years came with an odds ratio of 1.70 for active disease, ranging from 1.06 to 2.73. At 15 cigarettes a day it was 1.75, from 1.08 to 2.82, and at 15 pack-years it was 1.97, from 1.06 to 3.67.

  Zhang and colleagues, Clinical Rheumatology, 2018

• That paper also pooled the earlier studies on this question. Current smokers scored higher on the disease activity index than non-smokers, at a standardized mean difference of 0.34 from 0.18 to 0.48. Former smokers scored 0.10, from 0.01 to 0.19, which is much the smaller of the two.

  Zhang and colleagues, Clinical Rheumatology, 2018

• One study followed 230034 adults in South Korea. Fine particles raised the rate of new rheumatoid arthritis, at a hazard ratio of 1.74 from 1.06 to 2.86. For ankylosing spondylitis that study found no positive link at all, and that held in both of its models.

  Park and colleagues, Rheumatology, 2021

• One Taiwanese study matched 584 patients starting a biologic against 2336 who didn't. Carbon monoxide in the year before came with an odds ratio of 8.57, ranging from 2.02 to 36.32. Nitrogen dioxide in that model ran the other way, at 0.23 from 0.11 to 0.50.

  Kao and colleagues, Arthritis Research and Therapy, 2023

• A Cochrane review searched for trials of stopping smoking in inflammatory joint disease. It found two of them, with 57 smokers between them, and both were in rheumatoid arthritis. It found no trial at all in ankylosing spondylitis, and neither of those two trials measured disease activity in the first place.

  Roelsgaard and colleagues, Cochrane Database of Systematic Reviews, 2019

The one exposure with real numbers

Most environment pages on this site struggle to find sources at all, and this one turns out not to. It has a clear answer for smoking and a muddle for everything else. That difference is worth naming before any of the figures start, because it changes how much weight each of them holds, and because it's an unusual shape for a page here.

The smoking research in ankylosing spondylitis is decent by the standards of this subject. It's also asking one particular question rather than the one people usually mean. Knowing which question it answers changes what you can do with the figures, which is why it's worth spelling out here.

It asks how smokers who have this disease do compared with non-smokers who also have it. It doesn't ask whether smoking gives you the disease in the first place. Those are different questions, and only the first of them has been answered here, which is the distinction the rest of this page rests on.

How much you smoke is the finding

One study looked at 1178 patients, and it didn't only ask whether they smoked. It asked how long they'd smoked, how many cigarettes a day, and how many pack-years. All three of those tracked disease activity.

Smoking for at least 15 years came with an odds ratio of 1.70 for active disease, where the true value falls between 1.06 and 2.73. That range doesn't touch 1, so the link is a real one. At 15 cigarettes a day the figure was 1.75, from 1.08 to 2.82, and at 15 pack-years it was 1.97, from 1.06 to 3.67.

The risk climbed across each range rather than jumping at one threshold, and that last part is the useful bit. There's no line below which smoking stopped counting at all. Less was associated with lower risk rather than with no risk.

What happens to people who stopped

That paper also pooled the earlier studies on this question, comparing three groups of people against each other. Those three were current smokers, former smokers, and people who had never smoked. Current smokers scored higher on the disease activity index than the people who had never smoked at all, which is the comparison the rest of this section turns on.

The standardized mean difference between those two groups was 0.34, ranging from 0.18 to 0.48. For the function index the difference was 0.35, on an interval from 0.16 to 0.55. Former smokers scored a good deal closer to non-smokers, with a disease activity difference of only 0.10, from 0.01 to 0.19.

That's about a third of the current-smoker figure, and it's the most encouraging thing on this page. Read it carefully all the same, because it compares groups of people and never follows anybody through the act of quitting, which is a different study design and a weaker one. It's a reason for hope and it isn't proof that quitting moves the number.

No trial has tested quitting

This is the part that most pages on this subject skip over entirely. A Cochrane review went looking for trials of quitting, covering the whole inflammatory joint disease group rather than this one diagnosis on its own. It found two trials in all of that literature, with 57 smokers between the pair of them.

Both of those two trials were in rheumatoid arthritis, and it found no trial at all in ankylosing spondylitis. Neither of them measured disease activity, which happens to be the thing the review had set out to assess in the first place. The reviewers rated the certainty of the evidence very low, and they were right to.

So the honest position on all of this runs like this. Smokers with this disease do worse than non-smokers, and former smokers do better than current smokers. Stopping smoking has never been proved to be what makes the difference, which is a question the research hasn't answered rather than a negative result about quitting.

The air pollution muddle

Here the evidence pulls two ways at once, which is why this section exists at all. The biggest study on the question says no, and the smaller ones say yes without agreeing with each other about anything. That combination is the whole of what can honestly be reported.

One study followed 230034 adults living across three separate South Korean cities. Fine particles raised the rate of new rheumatoid arthritis, at a hazard ratio of 1.74 from 1.06 to 2.86. For ankylosing spondylitis that study found no positive link at all, and that held in its one-pollutant model and in its two-pollutant model alike.

A study across 14 cities in southwest China reported the opposite of that finding. It found fine particles linked to new cases of the disease, possibly with a lag of around two years. Its abstract gives no figure at all for that link, so this page has nothing from it that it can quote.

When a study disagrees with itself

A Taiwanese study shows why this whole area of research is hard to read. It matched 584 patients starting a biologic against a comparison group of 2336 patients who weren't starting one. Carbon monoxide in the previous year came out at an odds ratio of 8.57, on a range ranging from 2.02 to 36.32.

That range spans a factor of eighteen, so the size of the effect is very uncertain indeed. Nitrogen dioxide in that model came out at 0.23, from 0.11 to 0.50, which reads as though nitrogen dioxide protects you from something. That isn't a believable reading, and the authors of the study don't make it either.

They explain it themselves, because air pollutants travel together and a model can't cleanly separate them, so the estimates go unstable. They also note that they had no smoking data available to them at all. There's a second problem sitting underneath that whole study, and it's about the outcome rather than the exposure.

The thing being measured there was starting a biologic rather than getting worse. That depends on what an insurance scheme will approve as much as it depends on how a patient feels. A figure built on that outcome is measuring a health system as much as it's measuring a disease.

What to do with all this

Tell your team that you smoke, and tell them how much you smoke as well. The amount is what the research tracks, so a number is a great deal more use to them than a yes. Ask for real help with quitting. Don't try to work it out alone from a page like this one.

Your heart and your lungs are reason enough on their own here. The joint evidence points in that direction as well, even though no trial has ever proved it. That's two reasons to act rather than one, and the first of them doesn't depend on this page being right.

Don't move house over the pollution research, because the largest study of it found nothing here. The studies that did find something disagree with each other about what it was. Mention workplace dust and fumes anyway, since silica and solvents have never been measured in this disease and your lungs are reason enough to raise it.

Common misconceptions.

Cautions specific to this condition.

• Tell your team you smoke, and tell them how much. The amount is what the figures track, so a number is more use to them than a yes.
• Don't read the former-smoker figures as a promise. They're better than the current-smoker figures, and no trial has shown that quitting brings your disease activity down.
• Don't move house over the air pollution research. The largest study found nothing for this disease, and the ones that found something disagree with each other.
• Ask for help with quitting, because a page can't do that part for you. Your heart and your lungs are reason enough, whatever the joint research eventually shows.
• Take a figure about biologic starts with care. That study measured who got a drug approved, which depends on insurance rules as much as on how you feel.

Questions patients ask.