Adult-onset Still's and environment

What the research found.

• One study matched 72 patients against 216 controls on age, sex, education, and marital status. Ever smokers were 11 of the 72 patients, which is 15.3 percent, and among controls they were 25 of 216, or 11.6 percent. After adjustment there was no significant increase in risk.

  Rahmanpour and colleagues, Internal Medicine Journal, 2024

• One study followed 185 patients with this disease, of whom 45 smoked. Macrophage activation syndrome occurred in 28.9 percent of the smokers against 6.4 percent of the non-smokers. In the adjusted model smoking came out at a hazard ratio of 6.21, with the range ranging from 2.46 to 15.70.

  Ruscitti and colleagues, Clinical Rheumatology, 2022

• Among those 185 patients, smokers had more lung involvement, at an adjusted hazard ratio of 3.90 from 1.36 to 11.23. Smoking also raised the risk of death, at 4.25 from 1.33 to 13.55. That death figure is unadjusted, so read it with more care than the others.

  Ruscitti and colleagues, Clinical Rheumatology, 2022

• One nationwide study in Poland counted 1050 hospitalized patients over ten years. The average yearly rate was 0.32 per 100000 people, with the range ranging from 0.30 to 0.34 across that decade. At the end of 2018 the point prevalence was 2.7 per 100000 people.

  Bogdan and colleagues, Annals of Agricultural and Environmental Medicine, 2021

Two questions, two answers

There are two different questions about smoking and this disease, and they have different answers. Running them together is the easiest mistake to make here, and it gets made constantly in what people write about this. So this page keeps them apart before it gives you either answer.

The first question is whether smoking causes adult-onset Still's disease, and the answer from the one study that asked appears to be no. The second is what happens to smokers who already have it. The answer there is that they do considerably worse, which is the harder half of the page.

The risk question

One study matched 72 patients against 216 controls, matching them on age, sex, education, and marital status. Ever smokers were 11 of the 72 patients, which is 15.3 percent. Among the controls they were 25 of 216, which is 11.6 percent, and the difference between those two is small.

After adjusting for those four things, there was no significant increase in risk. The authors conclude that smoking probably isn't a risk factor for this disease, which is about as plain as a conclusion gets. They also found no difference in how the patients did afterward, which is the part that gets contested below.

That last point is worth holding onto, because the next section disagrees with it. Rahmanpour found no difference between ever smokers and never smokers in remission, in time to remission, or in flare rate. The second study on this page found something quite different about outcomes.

The outcome question

A different study looked at 185 patients who already had the disease, and 45 of those people smoked. The headline finding is about macrophage activation syndrome, which is the most dangerous complication this disease has and needs urgent hospital treatment. It's the reason this page exists in its current shape.

It occurred in 28.9 percent of the smokers and in 6.4 percent of the non-smokers. In the adjusted model, smoking came out at a hazard ratio of 6.21, with a range from 2.46 to 15.70. That range doesn't touch 1, so the link held up after adjustment.

It's an observational study, so it shows a link rather than proving that smoking caused the complication. That distinction is a real one and it holds. It also doesn't shrink the distance between the two groups, which is more than fourfold in the most dangerous complication this disease has.

Lungs, and death

The Ruscitti cohort also found smokers had more lung tissue involvement, at an adjusted hazard ratio of 3.90 from 1.36 to 11.23. It found a higher risk of death in smokers as well. That figure was a hazard ratio of 4.25, with a range from 1.33 to 13.55.

The death figure needs a warning attached to it, because it's unadjusted. That means nothing else about those patients was accounted for, and smokers may differ from non-smokers in ways that affect survival on their own. The paper reports no adjusted figure for death at all.

So the lung figure and the death figure are at different levels of confidence. One survived adjustment and the other was never put through it, which is why this page labels the second one every time it appears. That labeling is deliberate rather than merely cautious, and it happens three times on the page.

Where the two studies disagree

On causing the disease the two aren't in conflict, because only one of them asked the question. On what happens afterward the two of them disagree plainly. The first found no difference between ever smokers and never smokers in remission on treatment, remission off treatment, time to remission, or flare rate.

The second found large differences in complications and in survival, which is a long way from the first result. This page doesn't pick between them, because both rest on small numbers and neither is big enough to overrule the other. That's an unsatisfying place to leave a reader, and it's where the evidence leaves everybody.

Why everything here is small

The reason is that the disease is rare, and one study puts a number on how rare it is. A nationwide Polish count found 1050 hospitalized patients over ten years. The average yearly rate was 0.32 per 100000 people, with a range from 0.30 to 0.34, and point prevalence at the end of 2018 was 2.7 per 100000.

At those numbers, no country has enough patients for the kind of study that settles a question. Rheumatoid arthritis registries hold tens of thousands of people. This disease has case series counted in dozens, which is the whole explanation for everything above.

That Polish count also included only patients who were admitted to hospital, so the true figure is higher than 0.32. It still tells you the order of things, which is what this page uses it for. A large trial in a disease this uncommon is never going to be run.

What to do with all this

Tell your team that you smoke, because the complication figures here are large enough that it belongs in your notes. It belongs in the notes rather than in your own head, where it can't help anybody. That's the single most useful sentence anywhere on this page.

Ask about macrophage activation syndrome specifically if you smoke, because it's the outcome with the biggest difference between groups. It's also the one that needs catching early rather than late. Knowing its name is most of what a patient can do about it in advance.

Don't read the risk study as permission to keep smoking. It found no sign that smoking starts this disease, which is a different claim from smoking being harmless once you have it. Don't expect a settled answer either, because two small studies disagree about outcomes, no bigger one exists, and pretending otherwise would be worse than saying so.

Common misconceptions.

Cautions specific to this condition.

• Tell your team if you smoke. The complication figures in this disease are large enough that it belongs in your notes.
• Don't read the risk study as permission. It found no sign that smoking causes this disease, which says nothing about what smoking does once you have it.
• Take the death figure with extra care. It's unadjusted, meaning nothing else about those patients was accounted for.
• Ask about macrophage activation syndrome if you smoke. That's the complication with the biggest difference between smokers and non-smokers here.
• Don't expect a settled answer. These are two small studies that disagree with each other about outcomes, and no larger one exists.

Questions patients ask.