TY - JOUR
T1 - Long-term exposure to elemental components of fine particulate matter and all-natural and cause-specific mortality in a Danish nationwide administrative cohort study
AU - So, Rina
AU - Chen, Jie
AU - Stafoggia, Massimo
AU - de Hoogh, Kees
AU - Katsouyanni, Klea
AU - Vienneau, Danielle
AU - Samoli, Evangelia
AU - Rodopoulou, Sophia
AU - Loft, Steffen
AU - Lim, Youn Hee
AU - Westendorp, Rudi G.J.
AU - Amini, Heresh
AU - Cole-Hunter, Thomas
AU - Bergmann, Marie
AU - Shahri, Seyed Mahmood Taghavi
AU - Zhang, Jiawei
AU - Maric, Matija
AU - Mortensen, Laust H.
AU - Bauwelinck, Mariska
AU - Klompmaker, Jochem O.
AU - Atkinson, Richard W.
AU - Janssen, Nicole A.H.
AU - Oftedal, Bente
AU - Renzi, Matteo
AU - Forastiere, Francesco
AU - Strak, Maciek
AU - Brunekreef, Bert
AU - Hoek, Gerard
AU - Andersen, Zorana J.
N1 - Publisher Copyright:
© 2023
PY - 2023/5/1
Y1 - 2023/5/1
N2 - Background: Fine particulate matter (PM2.5) is a well-recognized risk factor for premature death. However, evidence on which PM2.5 components are most relevant is unclear. Methods: We evaluated the associations between mortality and long-term exposure to eight PM2.5 elemental components [copper (Cu), iron (Fe), zinc (Zn), sulfur (S), nickel (Ni), vanadium (V), silicon (Si), and potassium (K)]. Studied outcomes included death from diabetes, chronic kidney disease (CKD), dementia, and psychiatric disorders as well as all-natural causes, cardiovascular disease (CVD), respiratory diseases (RD), and lung cancer. We followed all residents in Denmark (aged ≥30 years) from January 1, 2000 to December 31, 2017. We used European-wide land-use regression models at a 100 × 100 m scale to estimate the residential annual mean levels of exposure to PM2.5 components. The models were developed with supervised linear regression (SLR) and random forest (RF). The associations were evaluated by Cox proportional hazard models adjusting for individual- and area-level socioeconomic factors and total PM2.5 mass. Results: Of 3,081,244 individuals, we observed 803,373 death from natural causes during follow-up. We found significant positive associations between all-natural mortality with Si and K from both exposure modeling approaches (hazard ratios; 95% confidence intervals per interquartile range increase): SLR-Si (1.04; 1.03–1.05), RF-Si (1.01; 1.00–1.02), SLR-K (1.03; 1.02–1.04), and RF-K (1.06; 1.05–1.07). Strong associations of K and Si were detected with most causes of mortality except CKD and K, and diabetes and Si (the strongest associations for psychiatric disorders mortality). In addition, Fe was relevant for mortality from RD, lung cancer, CKD, and psychiatric disorders; Zn with mortality from CKD, RD, and lung cancer, and; Ni and V with lung cancer mortality. Conclusions: We present novel results of the relevance of different PM2.5 components for different causes of death, with K and Si seeming to be most consistently associated with mortality in Denmark.
AB - Background: Fine particulate matter (PM2.5) is a well-recognized risk factor for premature death. However, evidence on which PM2.5 components are most relevant is unclear. Methods: We evaluated the associations between mortality and long-term exposure to eight PM2.5 elemental components [copper (Cu), iron (Fe), zinc (Zn), sulfur (S), nickel (Ni), vanadium (V), silicon (Si), and potassium (K)]. Studied outcomes included death from diabetes, chronic kidney disease (CKD), dementia, and psychiatric disorders as well as all-natural causes, cardiovascular disease (CVD), respiratory diseases (RD), and lung cancer. We followed all residents in Denmark (aged ≥30 years) from January 1, 2000 to December 31, 2017. We used European-wide land-use regression models at a 100 × 100 m scale to estimate the residential annual mean levels of exposure to PM2.5 components. The models were developed with supervised linear regression (SLR) and random forest (RF). The associations were evaluated by Cox proportional hazard models adjusting for individual- and area-level socioeconomic factors and total PM2.5 mass. Results: Of 3,081,244 individuals, we observed 803,373 death from natural causes during follow-up. We found significant positive associations between all-natural mortality with Si and K from both exposure modeling approaches (hazard ratios; 95% confidence intervals per interquartile range increase): SLR-Si (1.04; 1.03–1.05), RF-Si (1.01; 1.00–1.02), SLR-K (1.03; 1.02–1.04), and RF-K (1.06; 1.05–1.07). Strong associations of K and Si were detected with most causes of mortality except CKD and K, and diabetes and Si (the strongest associations for psychiatric disorders mortality). In addition, Fe was relevant for mortality from RD, lung cancer, CKD, and psychiatric disorders; Zn with mortality from CKD, RD, and lung cancer, and; Ni and V with lung cancer mortality. Conclusions: We present novel results of the relevance of different PM2.5 components for different causes of death, with K and Si seeming to be most consistently associated with mortality in Denmark.
KW - Air pollution
KW - Cox proportional regression model
KW - Denmark
KW - Long-term exposure
KW - Mortality
KW - PM elemental components
UR - http://www.scopus.com/inward/record.url?scp=85148951110&partnerID=8YFLogxK
U2 - 10.1016/j.envres.2023.115552
DO - 10.1016/j.envres.2023.115552
M3 - Article
C2 - 36822536
AN - SCOPUS:85148951110
SN - 0013-9351
VL - 224
JO - Environmental Research
JF - Environmental Research
M1 - 115552
ER -