OBJECTIVES: The role of outdoor air pollution in the incidence of chronic obstructive pulmonary disease (COPD) remains unclear. We investigated this question using a large, nationally representative cohort based on primary care records linked to hospital admissions. METHODS: A cohort of 812 063 patients aged 40-89 years registered with 205 English general practices in 2002 without a COPD diagnosis was followed from 2003 to 2007. First COPD diagnoses recorded either by a general practitioner (GP) or on admission to hospital were identified. Annual average concentrations in 2002 for particulate matter with an aerodynamic diameter <10 µm (PM10) and <2.5 µm (PM2.5), nitrogen dioxide (NO2), ozone and sulfur dioxide (SO2) at 1 km(2) resolution were estimated from emission-based dispersion models. Hazard ratios (HRs) per interquartile range change were estimated from Cox models adjusting for age, sex, smoking, body mass index and area-level deprivation. RESULTS: 16 034 participants (1.92%) received a COPD diagnosis from their GP and 2910 participants (0.35%) were admitted to hospital for COPD. After adjustment, HRs for GP recorded COPD and PM10, PM2.5 and NO2 were close to unity, positive for SO2 (HR=1.07 (95% CI 1.03 to 1.11) per 2.2 µg/m(3)) and negative for ozone (HR=0.94 (0.89 to 1.00) per 3 µg/m(3)). For admissions HRs for PM2.5 and NO2 remained positive (HRs=1.05 (0.98 to 1.13) and 1.06 (0.98 to 1.15) per 1.9 µg/m(3) and 10.7 µg/m(3), respectively). CONCLUSIONS: This large population-based cohort study found limited, inconclusive evidence for associations between air pollution and COPD incidence. Further work, utilising improved estimates of air pollution over time and enhanced socioeconomic indicators, is required to clarify the association between air pollution and COPD incidence.