To investigate the impact of short-term exposure to air pollutants and meteorological variation on ED visits for primary spontaneous pneumothorax (PSP).

We retrospectively identified PSP cases that presented at the ED of our tertiary center between January 2015 and September 2016. We classified the days into three types: no PSP day (0 case/day), sporadic days (1–2 cases/day), and cluster days (PSP, ≥3 cases/day). Association between the daily incidence of PSP with air pollutants and meteorological data were determined using Poisson generalized-linear-model to calculate incidence rate ratio (IRRs) and the use of time-series (lag-1 [the cumulative air pollution level on the previous day of PSP], lag-2 [two days ago], and lag-3 [three days ago]).

Using multivariate logistic regression analysis, O₃ (p = 0.010), NO₂ (p = 0.047), particulate matters (PM)₁₀ (p = 0.021), and PM_2.5 (p = 0.008) were significant factors of PSP occurrence. When the concentration of O₃, NO₂, PM₁₀, and PM_2.5 were increased, PSP IRRs increased approximately 15, 16, 3, and 5-fold, respectively. With the time-series analyses, atmospheric pressure in lag-3 was significantly lower and in lag-2, was significantly higher in PSP days compared with no PSP days. Among air pollutant concentrations, O₃ in lag-1 (p = 0.017) and lag-2 (p = 0.038), NO₂ in lag-1 (p = 0.015) and lag-2 (p = 0.009), PM₁₀ in lag-1 (p = 0.012), and PM_2.5 in lag-1 (p = 0.021) and lag-2 (p = 0.032) were significantly different between no PSP and PSP days.

Increased concentrations of air pollutants and abrupt change in atmospheric pressure were significantly associated with increased IRR of PSP.