Air pollution and meteorological factors affecting emergency hospital admissions of 630,059 respiratory patients in Poland

Poland has remained one of the most polluted countries in the EU for years. In 2021, the health costs of exposure to elevated concentrations of fine particulate matter in the air were estimated at 47,300 attributable deaths and 519,000 years of life lost for Polish citizens. Analysis is based on the first and comprehensive nationwide, long-term study for Poland (2012-2019), integrated daily datasets of air pollution and meteorological measurements from institutional WMO and Environmental Protection networks, and detailed registers of emergency hospital admissions due to COPD from the National Institute of Public Health. Surface analysis using TERYT codes enabled the assessment of variability in moving averages of atmospheric and environmental factors and of urgent hospitalization, as well as the identification of additional patient characteristics that affect the individual risk of short-term exposure leading to hospitalization.  

Our distributed lag non-linear models (DLNM) indicated robust short-term associations between ambient air pollution and meteorological conditions and urgent COPD admissions. Main finding: consistent with our previous results, sulfur dioxide (SO2) levels provide the strongest predictive signal for variations in COPD hospitalizations in Poland. Across the study period, higher SO2 levels were consistently associated with increased cumulative hospitalization risk, observed in both sexes, with slightly stronger cumulative estimates in men. While warmer conditions, as reflected in higher daily minimum and maximum temperatures, significantly reduced the risk of urgent hospitalization for COPD. Air temperature associations were observed in both sexes; however, the magnitude varied by indicator. Results for PM10, PM2.5, and NOx were slightly weaker and less consistent, e.g., NOx associations were more apparent in men and less consistent in women. After accounting for lagged effects, this suggests that not all pollutant signals are equally stable (partially seasonally drifted) at the nationwide level. 

Short-term environmental impacts observed on COPD hospitalizations are not acute - lag-response plots in all pollutants were several days, rather than limited to the exposure day. For gaseous pollutants, the lag structure suggested delayed cumulative impacts. For air temperature, risk reductions typically emerged after short delays, reached a peak on lag 14 days, and then attenuated toward later lags. Overall, our findings support the use of distributed lag frameworks in nationwide COPD burden assessments and emphasize the need to implement subgroup-specific prevention strategies.