Rationale: A link is established between air pollution and respiratory diseases. Very few studies evaluated this link with primary spontaneous pneumothorax (PSP). Contrasted results, low statistical power and methodological limits of these studies brought us to evaluate in a more thorough way this link.

Objectives: (1) to estimate the relation between PSP and air pollutants namely nitrogen dioxide (NO₂₎ , ozone (O₃₎ and particulate matter with a diameter ≤ 10 μm (PM₁₀ ); (2) to investigate a time lag effect between these pollutants and occurrence of PSP.

Methods: This study has a case-crossover design. Subjects aged ≥18 years admitted from 1st June 2009 to 31st May 2013, in 14 Emergency Departments centers on the French territory. Were excluded: patients with traumatic, secondary, recurrent or history of previous pneumothorax. NO₂ , O₃ and PM₁₀ data were collected hourly in monitoring stations. Three exposure assessments were retained: quantitative values, fast increase concentration of air pollutants and peak of pollution. These assessments were calculated for the entire exposure period and for each of the four days of all case and control periods.

Results: 948 subjects included. Whatever the pollutant considered, no differences were observed between case and control periods, regardless of whether the quantitative values of air pollutants exposure (p > 0.09), fast increase concentration (p > 0.46) and peak of pollution (p > 0.20).

Conclusions: We failed to show a relation between PSP and short-term air pollution exposure to low levels of NO₂ and PM₁₀ . An association between O₃ exposure and PSP cannot be ruled out. An impact at higher exposure level, and/or a potentiating effect of different meteorological factors remain to be demonstrated.