# Burden of disease attributable to PM2.5 at low exposure levels: impact of methodological choices

## TL;DR

This study shows how different methods for calculating health risks from low levels of air pollution can lead to very different results, emphasizing the need for transparency in reporting.

## Contribution

The study quantifies how methodological choices affect disease burden estimates from low PM2.5 exposure, using Norway as a case study.

## Key findings

- Methodological choices in concentration-response curves caused up to 104% variation in attributable deaths.
- Exposure and health data choices led to 40% differences in death estimates.
- Years of life lost accounted for 74% of PM2.5 attributable DALYs.

## Abstract

Exposure to fine particles (PM2.5) has been associated with adverse health outcomes, even at low exposure levels (< 10 µg/m3). Burden of disease assessments can quantify these associations; however, their sensitivity to methodological choices limits comparability between studies.

This study aimed to quantify the impact of methodological choices on disease burden attributable to low levels of ambient PM2.5, using Norway as a case study. Key methodological choices included (i) population exposure data, (ii) concentration-response curves, and (iii) population health data. Data from national and international sources were applied, including the global burden of disease (GBD) study. Attributable mortality and disability-adjusted life years (DALY) were estimated using burden of disease methodology. Additionally, the impact of choices related to concentration-response curves was assessed for higher exposure levels, using a scenario where exposure distributions were shifted to mean exposures up to 30 µg/m3.

Methodological choices related to the concentration-response curves had the largest impacts on the estimated attributable deaths, ranging from − 91% to 104% change relative to the reference estimate (1,448 deaths, 95% CI 502–1497). These choices had a smaller impact on higher exposure levels, varying from − 46% to 53%. The choice of exposure and population health data led to 40% differences in attributable death estimates. DALYs attributable to PM2.5 were predominantly driven by years of life lost (YLL: 74%). The choice of relative risk (RR) for the concentration response curve caused around 30% variation in DALY estimates relative to the reference (11,730 DALYs; 5,980 − 16,790), with larger differences for ischemic heart disease (-44 to 79%).

Attributable burden estimates for PM2.5 are highly sensitive to key methodological choices, particularly at low exposure levels. Consequently, transparent reporting of the methodological choices and data sources in PM2.5 health risk assessments are required to improve comparability and facilitate interpretations of the burden estimates.

The online version contains supplementary material available at 10.1186/s12940-025-01250-y.