BACKGROUND AND AIMS: Epidemiological studies have shown adverse effects of ambient air pollutants on health with inflammation and oxidative stress playing an important role. We examine the association between blood biomarkers of inflammation and coagulation and physical attributes of particulate matter which are not routinely measured such as particle length or surface area concentration and apparent density of PM. METHODS: Between 3/2007 and 12/2008 187 non-smoking individuals with type 2 diabetes mellitus (T2D) or impaired glucose tolerance (IGT) were examined within the framework of the KORA Study in Augsburg, Germany. In addition, we selected 87 participants with a potential genetic predisposition on detoxifying and inflammatory pathways. This was defined by the null polymorphism for glutathione S-transferase M1 in combination with a certain single nucleotide polymorphism on the C-reactive protein (CRP) gene (rs1205) or the fibrinogen gene (rs1800790). Participants had blood drawn up to seven different times, resulting in 1765 blood samples. Air pollutants were collected at a central measurement station and individual 24-h averages calculated. Associations between air pollutants and high sensitivity CRP, myeloperoxidase (MPO), interleukin (IL)-6 and fibrinogen were analysed using additive mixed models. RESULTS: For the panel with genetic susceptibility, increases were seen for CRP and MPO with most attributes, specifically particle length and active surface concentration. The %change of geometric mean and 95% confidence intervals for the 5-day average exposure for CRP and MPO were 34.6% [21.8;48.8] and 8.3% [3.2;13.6] per interquartile range increase of particle length concentration and 29.8% [15.9;45.3] and 10.4 [4.4;16.7] for active surface area. Results for the panel of T2D and IGT and the other blood biomarkers were less conclusive. CONCLUSIONS: Particle length concentration and active surface concentration showed strong positive associations with blood biomarkers reflecting inflammation. These air pollution metrics might reflect harmful aerosol properties better than particulate mass or number concentration. They might therefore be important for epidemiological studies.