Introduction Occupational exposure to wood dust is associated with increased risk of adenocarcinoma and probably also squamous cell sinonasal cancer. There is, however, still limited knowledge about the exposure-response relation and possible threshold levels.

Objectives To analyze the exposure-response relation between quantitative measures of wood dust exposure and risk of sinonasal cancer.

Methods A cohort study was conducted of the entire Danish working population 1979–2015 (n=5,421,248) with annual information on occupation since 1977. An inception population of workers 20 years or younger in 1977 was also established (n=3,012,247). Annual wood dust levels were assigned to each participant with a quantitative job exposure matrix (JEM) modelled from 12,653 personal measurements. A total of 2,576 incident cases of sinonasal cancer were identified in the National Patient Register during follow up. Incidence rate ratios (IRR) were analyzed in discrete time hazard models adjusted for gender, age, calendar year, education and JEM estimates of smoking probability.

Results During 36 years of follow up we identified 309 wood dust exposed cases. The adjusted IRR (95%CI) was 1.83 (1.56–2.15) for the highest cumulative exposure tertile (>11.2 mg/m³-years), 1.66 (1.41–1.97) for the longest duration of exposure (>5 years), and 1.64 (1.38–1.95) for the highest mean exposure tertile (>2.06 mg/m³) compared with no exposure. Trend tests were statistically significant only for mean exposure (P>0.001). No increased risk of sinonasal cancer was observed in the inception cohort that, however, only included 28 wood dust exposed cases.

Conclusion We found increased risk of sinonasal cancer associated with high-level wood dust exposure, but with no consistent trends. Future analyses of this material should separate adenocarcinomas from other histological subtypes.