Variability in the association between long-term exposure to ambient air pollution and mortality by exposure assessment method and covariate adjustment: A census-based country-wide cohort study

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningfagfællebedømt


  • Fulltext

    Accepteret manuskript, 246 KB, PDF-dokument

  • Mariska Bauwelinck
  • Jie Chen
  • Kees de Hoogh
  • Klea Katsouyanni
  • Sophia Rodopoulou
  • Evangelia Samoli
  • Richard Atkinson
  • Lidia Casas
  • Patrick Deboosere
  • Claire Demoury
  • Nicole Janssen
  • Jochem O. Klompmaker
  • Wouter Lefebvre
  • Amar Jayant Mehta
  • Tim S. Nawrot
  • Bente Oftedal
  • Matteo Renzi
  • Massimo Stafoggia
  • Maciek Strak
  • Hadewijch Vandenheede
  • Charlotte Vanpoucke
  • An Van Nieuwenhuyse
  • Danielle Vienneau
  • Bert Brunekreef
  • Gerard Hoek
Ambient air pollution exposure has been associated with higher mortality risk in numerous studies. We assessed potential variability in the magnitude of this association for non-accidental, cardiovascular disease, respiratory disease, and lung cancer mortality in a country-wide administrative cohort by exposure assessment method and by adjustment for geographic subdivisions.

We used the Belgian 2001 census linked to population and mortality register including nearly 5.5 million adults aged ≥30 (mean follow-up: 9.97 years). Annual mean concentrations for fine particulate matter (PM2.5), nitrogen dioxide (NO2), black carbon (BC) and ozone (O3) were assessed at baseline residential address using two exposure methods; Europe-wide hybrid land use regression (LUR) models [100x100m], and Belgium-wide interpolation-dispersion (RIO-IFDM) models [25x25m]. We used Cox proportional hazards models with age as the underlying time scale and adjusted for various individual and area-level covariates. We further adjusted main models for two different area-levels following the European Nomenclature of Territorial Units for Statistics (NUTS); NUTS-1 (n = 3), or NUTS-3 (n = 43).

We found no consistent differences between both exposure methods. We observed most robust associations with lung cancer mortality. Hazard Ratios (HRs) per 10 μg/m3 increase for NO2 were 1.060 (95%CI 1.042-1.078) [hybrid LUR] and 1.040 (95%CI 1.022-1.058) [RIO-IFDM]. Associations with non-accidental, respiratory disease and cardiovascular disease mortality were generally null in main models but were enhanced after further adjustment for NUTS-1 or NUTS-3. HRs for non-accidental mortality per 5 μg/m3 increase for PM2.5 for the main model using hybrid LUR exposure were 1.023 (95%CI 1.011-1.035). After including random effects HRs were 1.044 (95%CI 1.033-1.057) [NUTS-1] and 1.076 (95%CI 1.060-1.092) [NUTS-3].

Long-term air pollution exposure was associated with higher lung cancer mortality risk but not consistently with the other studied causes. Magnitude of associations varied by adjustment for geographic subdivisions, area-level socio-economic covariates and less by exposure assessment method.
TidsskriftScience of the Total Environment
StatusUdgivet - 2022

ID: 283743140