| Occupational exposure by JEM | Subjects exposed | Severe exacerbation | RR# (95% CI) | p-value for trend |
| DGF¶ | <0.001 | |||
| None | 568 | 34 (6.0) | 1.0 (reference) | |
| Low | 296 | 24 (8.1) | 1.3 (0.8–2.0) | |
| High+ | 102 | 16 (15.7) | 3.1 (1.9–5.1) | |
| Type of DGF | ||||
| Biological dust | <0.001 | |||
| None | 705 | 43 (6.1) | 1.0 (reference) | |
| Low§ | 234 | 26 (11.1) | 1.7 (1.1–2.6) | |
| High+ | 27 | 5 (18.5) | 3.6 (2.2–5.8) | |
| Mineral dust | 0.21 | |||
| None | 781 | 58 (7.4) | 1.0 (reference) | |
| Low | 131 | 10 (7.6) | 1.1 (0.5–2.1) | |
| High§ | 54 | 6 (11.1) | 1.8 (1.02–3.2) | |
| Gas and fumes | 0.08 | |||
| None | 605 | 41 (6.8) | 1.0 (reference) | |
| Low | 300 | 23 (7.7) | 1.1 (0.8–1.6) | |
| High§ | 61 | 10 (16.4) | 2.5 (1.2–5.5) |
Data are presented as n or n (%), unless otherwise stated. JEM: job-exposure matrix; DGF: dust, gas and fumes. #: from regression models with cluster-level variable for country, and covariates for sex, age, and smoking status. Regression models fit separately to determine effect of each type of DGF exposure. ¶: based on highest exposure to biological dust, mineral dust and gas and fumes. §: p≤0.05; +: p≤0.01.