Sjogren's disease and environment

What the research found.

• One case-control study compared 140 Italian women with Sjogren's syndrome against 109 female controls. Having a first-degree relative with an autoimmune disease came out at an odds ratio of 7.4, from 2.8 to 20.1. The link grew stronger as the number of affected relatives rose, which is the shape you want to see.

  Priori and colleagues, Clinical and Experimental Rheumatology, 2007

• One cohort across many countries compared how Sjogren's disease presented at diagnosis against national pollution figures. Dry eyes were commoner by 4.61 percentage points in the countries with the highest nitrogen and sulfur oxide exposure, which is the largest difference it found. Carbon monoxide followed at 3.59 points, non-methane compounds at 3.32, and PM2.5 at 3.30.

  Brito-Zeron and colleagues, Clinical and Experimental Rheumatology, 2023

• In that same cohort, dry mouth was commoner by 2.05 percentage points in the worst nitrogen and sulfur oxide countries, at P under 0.0001. Overall disease activity scores were higher in the worst countries too. That held for nitrogen and sulfur oxide, carbon monoxide, and PM2.5, all at P under 0.0001.

  Brito-Zeron and colleagues, Clinical and Experimental Rheumatology, 2023

• One gene study tested air pollutants against several autoimmune diseases. Predicted nitrogen oxides were associated with higher odds of Sjogren's syndrome, at 2.29 from 1.08 to 4.89. Coarse particles, the ones between PM2.5 and PM10, went the other way, at 0.29 from 0.10 to 0.90.

  Zhang and colleagues, Scientific Reports, 2025

A thin page, and it says so

This site holds three sources on the environment in Sjogren's disease, which is the thinnest environment base here after IgG4-related disease. Two of the three are about air pollution. Both of those have real problems in how they were built, and the third is about something that isn't an exposure at all.

The strongest finding across all three is whether somebody in your family has an autoimmune disease. That's worth knowing before any of the rest of it, because nothing else on the page comes near the size of it. It's an odd result for an environment page and it's the honest one.

What follows gives each finding with its problem attached rather than separating the two. Here the problems count for as much as the numbers do, and in two of the three cases they count for more. Reading the method is the whole exercise on a page as thin as this one.

Family history

One case-control study compared 140 Italian women who had Sjogren's syndrome against 109 female controls. It ran across seven teaching hospitals in Italy. It asked about lifestyle, medical history, and family history, and family history was by far the strongest link it found.

Women with a close relative who had an autoimmune disease had an odds ratio of 7.4, with the true value ranging from 2.8 to 20.1. That's a wide interval, and it's wide around a number large enough to survive the width. Nothing else on this page comes anywhere near it.

Two further findings in that study line up behind the first one. The link grew stronger as the number of affected relatives rose, which is the shape you want to see in a result like this. Close relatives of the patients were themselves at higher risk than relatives of the controls.

Now for the problems, and there are several of them. Family history came from a questionnaire people filled in themselves, and somebody with an autoimmune diagnosis is more likely to know what their relatives have. They're also more likely to report it, which inflates the difference between the two groups.

The controls were women at those same hospitals for bone and joint problems, which isn't a population sample. The whole study covers 249 women in all. So hold the size loosely and hold the direction less loosely, because a family history of autoimmune disease is worth telling your team about whatever the true figure turns out to be.

Air pollution, first study

One cohort across many countries asked a different question from the one you would expect. It didn't ask who develops Sjogren's disease, it asked how the disease looks at the moment of diagnosis. Then it compared that picture against national pollution figures.

Dryness was commoner in the countries with the worst pollution, and dry eyes were commoner by 4.61 percentage points where nitrogen and sulfur oxide exposure was highest. Carbon monoxide followed at 3.59 points, non-methane compounds at 3.32, PM2.5 at 3.30, and PM1.0 at 1.60. Those are percentage points rather than multiples, which is the first thing to hold onto.

Dry mouth followed that order too, and by smaller margins throughout the comparison. It was commoner by 2.05 percentage points in the worst nitrogen and sulfur oxide countries, at P under 0.0001. Non-methane compounds gave 1.21 points, at a P value of 0.007.

Overall disease activity was higher too, with mean scores higher in the worst countries for nitrogen and sulfur oxide, carbon monoxide, and PM2.5. All of those came in at P under 0.0001. Moderate or severe whole-body activity was more frequent in those countries as well.

What's wrong with that study

The design has a name and it's worth knowing, because exposure was assigned by country and not to any individual. Personal air was never measured at any point. That means everything else about those countries falls inside the pollution number, and countries differ in a great many ways.

Health systems differ, referral routes differ, and how early people get told they have this differs too. That last one is the serious problem, because the thing being measured is how much dryness and disease activity appear at diagnosis. A country where people wait longer for a diagnosis will show more of both, and that has nothing to do with its air.

Two smaller problems come with that one as well. Pollution was measured per unit of national output rather than as a concentration anybody breathed. The differences range from 1.60 to 4.61 percentage points, which is small, and the very low P values come from the size of the cohort instead.

Its own authors present the work as a first suggestion rather than as a settled finding. That is the right description of what they did, and it's an unusually careful one. This page has no reason to improve on their wording.

Air pollution, second study

One genetic study came at that question from another angle, using gene variants that predict how much exposure somebody had. Air was never measured for any individual in that study either. So it's a different kind of weakness rather than the absence of one, and that's worth saying before you read the figures.

Nitrogen oxides were associated with higher odds of Sjogren's syndrome, at an odds ratio of 2.29 ranging from 1.08 to 4.89. That result came in at P equals 0.032. On its own it would read as a modest positive finding, and that's how it usually gets quoted.

Then that analysis found something its authors couldn't explain and didn't try to. Coarse particles, the ones between PM2.5 and PM10, went the other way, at 0.29 ranging from 0.10 to 0.90. That also came in at P equals 0.032, which is the part that makes it hard to set aside as noise.

Read those two results together, because separately each one looks like a finding. One pollutant raised the odds and another lowered them, in one study, in one disease, and the paper offers no explanation for the split. A result that points both ways at once is a reason for caution rather than a reason for confidence in either direction.

What this page doesn't cover

Three things a reader will look for are missing from this page, and the reason for all three is identical. This site holds no source on any of them in Sjogren's disease. That absence may be real or it may be a hole in our own harvest, and saying which would be a guess.

The first is smoking, which leads the evidence on the rheumatoid arthritis and lupus environment pages. Nothing in our set measures it here at all, in either direction. That's the absence most likely to bother somebody who reads across several of these condition pages in a row.

The second is infection, and viral triggers are a standard part of how this disease gets discussed elsewhere. Our set holds nothing at all on them. A page that named one without a source would be inventing a mechanism, which is the thing this site exists not to do.

The third is silica and the other exposures people meet at work. Silica is linked to lupus, systemic sclerosis, and ANCA-associated vasculitis on three other pages in this set. Nothing in our sources measures it in Sjogren's disease, and whether that's a real absence is an open question.

All three absences are recorded for the physician review rather than left silent. A visible hole is more use to you than a confident sentence with nothing behind it, which is the whole principle here. That's the reason this section exists on a page that could easily have left it out.

What these trials didn't test

Not one of these studies changed an exposure and then watched what followed from it. Every one of them measured something people already had, and not one followed anybody forward from an exposure to a diagnosis. That's a limit on all three of them rather than a criticism of any one.

So nothing here says that pollution causes Sjogren's disease. Nothing here says what happens if an exposure changes, because the studies that would answer that question haven't been done yet. Saying otherwise would be going a good deal further than any of these three papers allows.

Cleaner air is worth wanting for reasons that don't depend on any of this. Moving house on the strength of these figures would be a mistake, because they hold a good deal less than they look like they do. That's a judgment about how the studies were built rather than about air quality.

What's genuinely worth doing is telling your team about autoimmune disease in your family. That's the one finding here with real size behind it, at an odds ratio of 7.4 ranging from 2.8 to 20.1. It costs nothing to mention at a visit, and it's the most useful sentence on this page.

Common misconceptions.

Cautions specific to this condition.

• Tell your team about autoimmune disease in your family. It's the strongest link any study on this page found, at an odds ratio of 7.4 from 2.8 to 20.1.
• Read the pollution differences as percentage points, not as multiples. The largest was 4.61 points and the smallest 1.60.
• Treat a study that assigns exposure by country as a starting point. Everything else that differs between those countries falls inside the pollution figure.
• Don't act on a gene study whose two pollutants point opposite ways. One came out at 2.29 and one at 0.29 within a single analysis.
• Nothing here has been tested by changing an exposure. Every study measured what people already had.
• This page gives no exposure limit, no threshold, and no advice about where to live.

Questions patients ask.