In depth
Systemic sclerosis and environment
This is the disease with the clearest occupational evidence anywhere on this whole site. Three separate reviews agree on which exposures count, and not one of them agrees with the others on how much. So the direction here is settled and the size of every estimate on the page is not.
What the research found.
One review pooled six exposures at once, which makes it the most useful single table here. Silica gave an odds ratio of 2.81, from 1.86 to 4.23, across 15 case-control studies, and solvents gave 2.00, from 1.32 to 3.02, across 13. Epoxy resins gave 2.97, from 2.31 to 3.83, across 4, and pesticides gave 1.02, from 0.78 to 1.32.
One review pooled 16 studies on silica at work, and the combined relative risk was 3.20, from 1.89 to 5.43. The studies disagreed enormously with each other, at an I-squared of 97.2 percent. Case-control studies gave 2.24, from 1.65 to 3.31, while cohort studies gave 15.49, from 4.54 to 52.87.
McCormic and colleagues, International Archives of Occupational and Environmental Health, 2010
That same silica review split its estimate by sex. In men it was 3.02, from 1.24 to 7.35. In women it was 1.03, from 0.74 to 1.44, which crosses 1 and isn't significant.
McCormic and colleagues, International Archives of Occupational and Environmental Health, 2010
One review pooled case-control studies on solvents at work, and the odds ratio was 2.4, from 1.7 to 3.4. Adjusting for publication bias brought it down to 1.8, from 1.2 to 2.5. In men it was 3.0 and in women 1.8, with the difference between the two at P equals 0.03.
One nationwide survey scored lifetime silica exposure in patients and in a set of matched controls. Scores were higher in the people with systemic sclerosis, at P equals 0.001. The difference came almost entirely from work rather than from home or from hobbies, at P under 0.0001, which narrows the question.
The disease with the clearest occupational evidence
Three separate reviews have pooled the work exposures in systemic sclerosis, and they agree on which ones are linked. Where they disagree, and the disagreement is sharp rather than marginal, is on how large each of those links turns out to be. That combination is unusual enough that it's worth understanding before you reach any of the numbers below.
When reviews agree on direction and disagree on size, the direction is normally the trustworthy part of it. That's the principle this page is built on. It shapes how every figure below gets presented, and it's the reason not one of them appears without the interval attached. An interval says a good deal more than a point estimate does in a literature like this one.
So this page gives the direction plainly, and it treats every point estimate with a good deal of care. That isn't hedging, and it's what these three papers support between them. Anything firmer would be reading a good deal more into them than they hold on their own.
Silica
One review pooled 15 case-control studies, and occupational silica exposure came out at an odds ratio of 2.81. The true value for it ranges from 1.86 to 4.23. That's the most reliable figure anywhere on this page, and it rests on more studies than anything else here.
A second review pooled 16 studies of all designs, and the combined relative risk was 3.20, from 1.89 to 5.43. Those two figures sit close enough together to be reassuring at first glance. The next section explains why a second glance is a good deal less comfortable.
A third study came at it from another angle entirely, scoring lifetime silica exposure in patients and matched controls across a whole country. Scores were higher in people with systemic sclerosis, at P equals 0.001. That's a different method reaching the direction the reviews found, which counts for something.
The useful detail in that third study is where the exposure came from. It came almost entirely from work rather than from home or hobbies, and that difference held at P under 0.0001. So this is an occupational question rather than a household one, which narrows it usefully.
Why the silica estimates disagree
Here is where the honest reading gets harder, and it's the part most summaries leave out. Inside that second review the answer depended entirely on the kind of study being pooled, and two designs gave two very different numbers. Understanding why is more useful than either number on its own.
Case-control studies gave a relative risk of 2.24, with the true value ranging from 1.65 to 3.31. Cohort studies in that same review gave 15.49, from 4.54 to 52.87. Those are one exposure and one disease, measured two different ways, and the two ways disagree by a factor of seven.
That review reports how much its own studies disagreed, and the I-squared was 97.2 percent. That measure ranges from 0 to 100, so 97.2 sits about as high as it ever gets in practice. Heterogeneity like that means the studies are answering slightly different questions.
So a reader could truthfully say silica raises the risk fifteen-fold, and it would mislead badly. What these reviews establish is that the link is real, and what they don't establish is the size of it at all. This page won't pretend otherwise, which is why 15.49 appears only with its interval.
Why men see the higher risk
Systemic sclerosis is much commoner in women, and the occupational risk in these reviews is much higher in men. Both of those statements are true at once. Putting the two of them together is the work of this section and of the one that follows it.
One review split its silica estimate by sex, and in men it came out at 3.02, from 1.24 to 7.35. In women it came out at 1.03, from 0.74 to 1.44, which crosses 1. A second review did likewise for solvents, finding 3.0 in men and 1.8 in women.
The men's and women's solvent intervals ranged from 1.9 to 4.6 and from 1.5 to 2.1. The difference between the two was significant at P equals 0.03. So both of the reviews split in one direction, across two entirely different exposures.
The third study explains why, and the explanation it gives is about jobs rather than about biology. Men had much higher lifetime occupational exposure scores than women in this disease, at P under 0.0001. Non-occupational scores didn't differ by sex at all. That last detail is what settles the question, because it separates the exposure from the person.
What that does and doesn't mean for a woman
Two wrong readings sit either side of this question, and both of them are easy to reach from what's above. Neither one of them follows from the studies themselves. What follows below takes the two of them one at a time, because they need separating.
The first is that a workplace exposure can't be the point, because this is a women's disease. That isn't what the numbers show at all. What they show is that fewer women held the jobs where the dust was in the first place.
The second is that silica is therefore irrelevant to a woman, and that doesn't follow either. The women's estimate was 1.03 and its interval ranged from 0.74 to 1.44, so that range contains no effect and also contains a 44 percent rise. Fewer exposed women means a wider interval and less power to see anything at all.
So the question is unanswered in women rather than answered no, and those two are different things. If you have had occupational dust exposure of any kind, at any point in your working life, it's worth telling your team. That holds whatever your sex and whatever these reviews happened to find.
Solvents, and a figure its authors corrected
One review pooled 13 case-control studies on solvents, and the odds ratio came out at 2.00. The true value for it ranges from 1.32 to 3.02. That's a smaller figure than the silica one, and it rests on a comparable number of studies behind it.
A second review pooled case-control studies on the same question and found 2.4, with the true value from 1.7 to 3.4. Then it did something worth noticing, which is that it adjusted its own estimate for publication bias. The figure fell to 1.8, with the true value from 1.2 to 2.5.
The reason for that adjustment is set out plainly in the paper itself. Smaller studies reported larger effects, at P equals 0.003, which is the usual sign that negative results went unpublished. That pattern turns up right across medicine, and it's very rarely corrected for like this.
So the number to take away from solvents is 1.8 rather than 2.4. A review that corrects its own estimate downward is doing the harder of the two things available to it. It's also doing the more honest of the two, and that's worth saying.
Epoxy resins, and one exposure with nothing behind it
The highest single point estimate in that first review wasn't silica at all. Epoxy resins came out at an odds ratio of 2.97, with the true value from 2.31 to 3.83. That figure rests on 4 case-control studies, which is a thin base by any standard.
The interval on it is fairly tight all the same. Read that as a real signal which needs more work, not as a settled number. Four studies can agree with each other and still all be wrong in one direction, which is why the base counts.
That same review found one exposure with nothing behind it at all. Pesticides came out at 1.02, from 0.78 to 1.32, and that interval falls squarely across 1. So a review that found three real links also found one clear absence among them.
The negative result is worth as much as the positive ones, and it's usually skipped. A review that finds three exposures linked and one not linked is telling them apart rather than finding a link everywhere it looks. That's a reason to trust the other three rather than to discount them.
What these trials didn't test
Every study on this page measured an exposure that people already had. Not one of them assigned an exposure to anybody, and not one removed one and watched what happened afterward. That's a limit on all four together rather than a criticism of any one.
So a link here is a reason to look further rather than proof that the exposure caused the disease. It says nothing at all about what happens if the exposure stops. The study that would answer that particular question has never been run, and it doesn't look likely soon.
Cutting avoidable dust exposure is sensible on its own terms, and it doesn't need any of these figures to hold up. Calling it a treatment for systemic sclerosis goes past everything written above it. The difference between those two sentences is worth holding onto with some care.
What the figures are genuinely good for is a conversation with the people treating you, and that's worth having. Tell them what you did for work, going back as far as your first job. Silica and solvent exposures often ended decades before a diagnosis, and that won't come to light at all unless you say so.
Common misconceptions.
Myth. The evidence on silica is settled.
Reality. The direction is settled and the size isn't. One review pooled 16 studies and found a relative risk of 3.20, from 1.89 to 5.43, while inside that same review case-control studies gave 2.24 and cohort studies gave 15.49, from 4.54 to 52.87. The studies disagreed at an I-squared of 97.2 percent, which is about as high as that measure goes. So the exposure is real and the number attached to it isn't.
Myth. This mostly affects women, so a workplace exposure can't be the point.
Reality. Both reviews found the occupational risk appearing in men rather than in women. One found silica at 3.02 in men and 1.03 in women, where it wasn't significant, and the other found solvents at 3.0 in men and 1.8 in women. That difference came in at P equals 0.03, and a third study explains why it happened. Men had much higher work exposure scores than women, while scores outside work didn't differ by sex at all. That's who was exposed rather than who is vulnerable.
Myth. So silica is irrelevant to a woman with this disease.
Reality. That doesn't follow either, and it's the mirror of the mistake above it. The women's estimate for silica was 1.03, from 0.74 to 1.44, and that interval contains no effect and also contains a 44 percent rise. Fewer exposed women means a wider interval and less power to see anything at all. So the right reading is that the question is unanswered in women rather than answered no.
Myth. All the workplace exposures point one way.
Reality. They don't, and one review says so within a single table. Silica came out at 2.81, from 1.86 to 4.23, and epoxy resins came out at 2.97, from 2.31 to 3.83. Pesticides came out at 1.02, from 0.78 to 1.32, and that interval falls squarely across 1. So a review that found three real links also found one clear absence among them, which is a reason to trust the other three.
Myth. The solvent figure is straightforward.
Reality. Its own authors adjusted that figure downward, which is worth knowing before you quote it. The pooled odds ratio was 2.4, from 1.7 to 3.4, and after adjusting for publication bias it fell to 1.8, from 1.2 to 2.5. They found smaller studies reporting larger effects, at P equals 0.003, which is the usual sign that negative work went unpublished. The lower figure is the one to take away.
Cautions specific to this condition.
- Tell your team what you did for work, going back to your first job. Silica and solvents both showed dose and duration effects across these reviews, and the exposure often ended decades before the diagnosis.
- Read a cohort estimate of 15.49 with its interval of 4.54 to 52.87. An interval spanning a tenfold range says the direction is real and the size is unknown.
- Don't take a null result in women as an answer. The interval there ranges from 0.74 to 1.44, which is too wide to rule anything out.
- Ask about epoxy resins as well as silica. One review put them at 2.97, from 2.31 to 3.83, which is the highest point estimate in the whole table.
- No study here assigned an exposure to anybody. Every one measured an exposure people already had, which is weaker evidence than a trial.
- Nothing here has been tested as a treatment. No study asked what happens to the disease when an exposure stops.
Discuss any change with the rheumatologist who manages your care. Nothing here replaces that conversation.
Questions patients ask.
Which exposure has the strongest evidence here?
Silica dust at work, by a clear distance from anything else here. One review pooled 15 case-control studies and found an odds ratio of 2.81, from 1.86 to 4.23, while a second pooled 16 studies of all designs and found a relative risk of 3.20, from 1.89 to 5.43. A third scored lifetime exposure and found it higher in patients than in controls, at P equals 0.001.
Why do the silica numbers vary so much?
Because the study design changes the answer, and it changes it enormously. Inside one review, case-control studies gave 2.24, from 1.65 to 3.31, while cohort studies in that same review gave 15.49, from 4.54 to 52.87. The studies disagreed at an I-squared of 97.2 percent, which is near the top of that scale. So the link is solid and the size of it isn't.
Does this apply to women?
The studies can't say, and that isn't the same as saying that the answer is no. One review found silica at 3.02 in men and 1.03 in women, where the interval ranged from 0.74 to 1.44. That interval contains no effect and it also contains a 44 percent rise. Far fewer women were exposed at work, which makes that estimate imprecise rather than reassuring.
Why is the risk higher in men when the disease is commoner in women?
Because the exposure is higher in men, and one nationwide survey shows it directly. It scored lifetime silica exposure in patients and in controls, and men had much higher scores than women in this disease, at P under 0.0001. That difference came through work, because scores outside work didn't differ by sex at all. So it's about which jobs people held rather than about who the disease affects.
What about solvents?
They come second, and the figure attached to them needs care. One review pooled 13 case-control studies and found an odds ratio of 2.00, from 1.32 to 3.02, while another pooled study found 2.4, from 1.7 to 3.4. That second figure fell to 1.8 after adjusting for publication bias, because smaller studies reported larger effects at P equals 0.003.
Is there anything with no link at all?
Pesticides, on the one review that looked at them directly. The odds ratio was 1.02, from 0.78 to 1.32, and that interval falls squarely across 1. That's a useful result rather than a disappointing one, because a review finding three exposures linked and one not linked is doing its job and making the three easier to trust.
Can changing my exposure now help my disease?
No study here tested that, because every one of them measured exposures people already had. So not one of them can say what happens when an exposure stops. Cutting avoidable dust exposure is sensible on its own terms, and calling it a treatment goes past what any of this shows, which is a real limit rather than a technicality.
References.
- Rubio-Rivas M; Moreno R; Corbella X. Occupational and environmental scleroderma. Systematic review and meta-analysis. Clin Rheumatol. 2017;36:569-582. 10.1007/s10067-016-3533-1Systematic review and meta-analysis of all studies published on systemic sclerosis and occupational or environmental exposure
- McCormic ZD; Khuder SS; Aryal BK et al. Occupational silica exposure as a risk factor for scleroderma: a meta-analysis. Int Arch Occup Environ Health. 2010;83:763-9. 10.1007/s00420-009-0505-7Meta-analysis of 16 studies on occupational silica exposure and systemic sclerosis
- Kettaneh A; Al Moufti O; Tiev KP et al. Occupational exposure to solvents and gender-related risk of systemic sclerosis: a metaanalysis of case-control studies. J Rheumatol. 2007;34:97-103. PMID 17117485Meta-analysis of 11 case-control studies comprising 1
- Cavalin C; Lescoat A; Sigaux J et al. Crystalline silica exposure in patients with rheumatoid arthritis and systemic sclerosis: a nationwide cross-sectional survey. Rheumatology (Oxford). 2023;62:2707-2715. 10.1093/rheumatology/keac675Nationwide cross-sectional survey using a purpose-built and validated lifetime exposure questionnaire covering occupational and non-occupational crystalline silica exposure
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.
Across the site