Long-term exposure to air pollution and 5-year mortality after lung cancer diagnosis: a Danish nurse cohort study

While the association between air pollution and lung cancer is well established, its relationship with post-diagnosis mortality remains limited, especially in low-exposure settings. We investigated the association of long-term exposure to air pollution with 5-year mortality after lung cancer diagnosis in Denmark. In this nested cohort, we included 449 participants from the Danish Nurse Cohort who were diagnosed with lung cancer between their baseline year (1993/1999) and 2020. We estimated mean residential annual concentrations of particulate matter (PM_2.5 and PM₁₀), nitrogen dioxide (NO₂), elemental carbon (EC), and ozone (O₃) using DEHM/UBM/AirGIS. We used time-varying Cox regression models to minimize immortal time bias (ITB), adjusting for age, baseline year, date of diagnosis, time since diagnosis, smoking, socioeconomic factors, and additional lifestyle variables. Of the 449 participants diagnosed with primary lung cancer, 377 died within 5 years after diagnosis. Hazard ratios (HRs) per interquartile range increase in 5-year moving average exposures were 0.99 (95% CI: 0.71-1.39) for PM_2.5, 0.93 (0.71-1.22) for PM₁₀, 0.99 (0.84-1.17) for NO₂, 0.95 (0.82-1.09) for EC, and 1.04 (0.89-1.22) for O₃ in the fully adjusted model. A methodological comparison demonstrated that time-fixed exposure averaging substantially overestimated associations (PM_2.5: HR 3.00 vs. 0.99), confirming the importance of time-varying approaches to avoid ITB. We found no association between long-term air pollution exposure and 5-year mortality in Danish female nurses diagnosed with lung cancer. Our study provides valuable evidence on this scarcely researched topic and highlights the critical methodological consideration of ITB in survival analyses with time-varying environmental exposures.