Association between long-term exposure to low ambient PM2.5 and cardiovascular hospital admissions: A UK Biobank study

Abstract

Introduction: A causal link between air pollution exposure and cardiovascular events has been suggested. However fewer studies have investigated the shape of the associations at low levels of air pollution and identified the most important temporal window of exposure. Here we assessed long-term associations between particulate matter < 2.5 µm (PM_2.5) at low concentrations and multiple cardiovascular endpoints using the UK Biobank cohort.

Methods: Using data on adults (aged > 40) from the UK Biobank cohort, we investigated the associations between 1-year, 3-year and 5-year time-varying averages of PM_2.5 and incidence of major adverse cardiovascular events (MACE), myocardial infarction (MI), heart failure, atrial fibrillation and flutter and cardiac arrest. We also investigated outcome subtypes for MI and stroke. Events were defined as hospital inpatient admissions. We fitted Cox proportional hazard regression models applying extensive control for confounding at both individual and area level. Finally, we assessed the shape of the exposure-response functions to assess effects at low levels of exposure.

Results: We analysed data from 377,736 study participants after exclusion of prevalent subjects. The average follow-up (2006-2021) was 12.9 years. We detected 19,353 cases of MACE, 6,562 of acute MI, 6,278 of heart failure, 1,258 for atrial fibrillation and flutter, and 16,327 for cardiac arrest. Using a 5-year exposure window, we detected positive associations (for 5 μg/m³ increase in PM_2.5) for 5-point MACE of [1.12 (95 %CI: 1.00-1.26)], heart failure [1.22 (1.00-1.50)] and cardiac arrest [1.16 (1.03-1.31)]. We did not find any association with acute MI, while non-ST-elevation MI was associated with the 1-year exposure window [1.52 (1.12-2.07)]. The assessment of the shape of the exposure-response relationships suggested that risk is approximately linear for most of the outcomes.

Conclusions: We found positive associations between long-term exposure to PM_2.5 and multiple cardiovascular outcomes for different exposure windows. The cardiovascular risk tends to rise even at exposure concentrations below 12-15 μg/m³, indicating high risk below UK national and international thresholds.

Keywords: Air pollution; Cardiovascular disease; Hospital admissions; Long-term.

Fig. 1

Flow diagram representing the selection of the sample of the UK Biobank.

Fig. 2

Concentration-response functions of the associations between lag04 (5-year time-dependent) of PM_2.5 with cardiovascular events in the UK Biobank cohort. Models were fully adjusted (Model 2). The associations representing hazard ratios (HRs, with 95% confidence intervals) were estimated using penalized splines with degrees of freedom selected using the Akaike information criterion (AIC, solid line) with 95% confidence intervals (surrounding dashed lines).