LAUSR

Background CAO is often associated with smoking but also seen in non-smokers. As risks of CAO attributable to different exposures are dependent on the prevalence of exposure, PAR is by nature variable but has not been systematically estimated across diverse sites. Aims We estimated the proportion of the total population with CAO attributable to exposures (PAR) for smoking, passive smoking, poor education, low body mass index (BMI), dusty work and tuberculosis (TB) for participants aged over 40 living in 40 centres participating in an international study on lung health. Methods We defined CAO as a ratio of Forced Expiratory Volume in 1 s to Forced Vital Capacity less than the lower limit of normal (LLN) for white Americans. As PAR depends on both relative risk and prevalence of exposure we used a Bayesian method to improve the precision of local PAR estimates, borrowing information across centres on the relative risk of interest (mutually adjusted for all other risk factors). Results In 31 centres there was less ‘unexplained’ CAO than expected from a definition of CAO using the LLN. This defines 5% of the ‘normal’ reference population (never smokers without symptoms or diagnosis) as having obstruction. The highest levels of unexplained CAO beyond the 5% expected from this definition were in Salzburg (+2.5%) and Maastricht (+2.0%). The most important causes were (figure 1): Conclusions Most CAO can be explained. The most important risk factors after tobacco are associated with social conditions. The mechanistic explanation for these remains unclear.Abstract M4 Table 1 Risk (Definition) Mean PAR (%) (Range) Tobacco (Smoking and passive smoking) 3.52 (0 to 10.8) Social indicators (Poor education and low BMI) 1.9 (0.3 to 6.2) Dusty work 0.6 (0 to 1.6) TB 0.2 (0 to 2.6)