Objectives In this study we aimed to estimate the associations between respiratory health at inclusion and drop-out from a 4-year longitudinal study of lung function among cement production workers in Europe and Turkey.

Method Non-administration workers (n = 3203) aged 17–54 yrs in 22 cement production plants in eight countries were included. Geometric mean (GM) exposure to the thoracic fraction of the workplace aerosol was allocated individually by job type and plant from a database of 6111 measurements from the follow-up study. Drop out was analysed in logistic regression with year of hire, exposure level, airway symptoms, chronic obstructive pulmonary disease (COPD) and use of airway protection at inclusion as explanatory variables. COPD was defined as FEV₁/FVC < 0.7.

Results COPD at inclusion was associated with drop out, but only among those hired 6–10 years before inclusion (OR 2.4 [CI 95% 1.1–5.5]). GM exposure exceeding a level of 1.61 mg/m³ was associated with drop out (OR= 1.9; CI 95% 1.3–2.7) compared to exposure below 0.42 mg/m³. The use of respiratory protection at inclusion was inversely associated with drop out (OR 0.6 [CI 95% 0.4–0.9]). Age, sex, allergy, asthma and previous occupational exposure to dust and gases did not confound the associations.

Conclusions Workers in the highest exposed group had increased risk, while those who reported using a respiratory mask had reduced risk of dropping out. Participants classified with COPD at inclusion were to some extent depleted from the study, but only among those hired 6–10 years before their inclusion.