Table 3 Occupational exposures and the risk of COPD
ExposureUnadjusted OR (95% CI)Adjusted OR (95% CI)PAF (95% CI)
Entire FLOW cohort
VGDF exposure*2.18 (1.69 to 2.83)2.11 (1.59 to 2.82)31% (22% to 39%)
JEM exposure probability*
    Low (referent)1.0 (referent)1.0 (referent)Referent
    Intermediate1.64 (1.00 to 2.67)1.27 (0.74 to 2.19)2% (−2% to 6%)
    High2.66 (1.80 to 3.94)2.27 (1.46 to 3.52)13% (8% to 18%)
Cigarette smoking
    Never (referent)1.01.0Referent
    Current smoker31 (17 to 58)31 (17 to 58)32% (30% to 33%)
    Past smoker4.67 (3.50 to 6.23)4.52 (3.35 to 6.09)42% (37% to 46%)
GOLD stage II or higher
VGDF exposure*2.33 (1.77 to 3.06)2.13 (1.55 to 2.93)31% (21% to 41%)
JEM exposure probability*
    Low (referent)1.0 (referent)1.0 (referent)Referent
    Intermediate1.77 (1.05 to 2.97)1.58 (0.88 to 2.84)3% (−1% to 8%)
    High2.79 (1.86 to 4.19)2.33 (1.45 to 3.72)14% (8% to 20%)
Cigarette smoking
    Never (referent)1.01.0Referent
    Current smoker45 (24 to 86)46 (24 to 88)35% (32% to 37%)
    Past smoker7.05 (5.02 to 9.91)6.27 (4.42 to 8.90)46% (41% to 50%)
  • COPD, chronic obstructive pulmonary disease; GOLD, Global Obstructive Lung Disease initiative; JEM, job exposure matrix for exposure likelihood on longest held job; PAF, population attributable fraction based on multivariate analysis; VGDF, vapours, gas, dust or fume on longest held job by self-report.

  • *Unadjusted results are from logistic regression models with only that set of variables; adjusted results are from two separate models for exposure, one with VGDF and one with JEM, controlling for age, sex, race and smoking. Adjusted smoking risk estimates are derived from the VGDF model; results for smoking were not substantively different in the JEM model. Results are shown separately for the entire cohort and for the subset of subjects with GOLD stage II or greater, together with referents.