Estimated effects of air pollution and space-time-activity on cardiopulmonary outcomes in healthy adults: A repeated measures study
Tom Cole-Hunter,Audrey de Nazelle,David Donaire-Gonzalez,Nadine Kubesch,Glòria Carrasco-Turigas,Florian Matt,Maria Foraster,Tania Martínez,Albert Ambros,Marta Cirach,David Martinez,Jordina Belmonte,Mark Nieuwenhuijsen
DOI: https://doi.org/10.1016/j.envint.2017.11.024
Abstract:Background: Exposure to air pollution is known to affect both short and long-term outcomes of the cardiopulmonary system; however, findings on short-term outcomes have been inconsistent and often from isolated and long-term rather than coexisting and short-term exposures, and among susceptible/unhealthy rather than healthy populations. Aims: We aimed to investigate separately the annual, daily and daily space-time-activity-weighted effect of ambient air pollution, as well as confounding or modification by other environmental (including noise) or space-time-activity (including total daily physical activity) exposures, on cardiopulmonary outcomes in healthy adults. Methods: Participants (N=57: 54% female) had indicators of cardiopulmonary outcomes [blood pressure (BP), pulse (HR) and heart rate variability (HRV {SDNN}), and lung function (spirometry {FEV1, FVC, SUM})] measured on four different mornings (at least five days apart) in a clinical setting between 2011 and 2014. Spatiotemporal ESCAPE-LUR models were used to estimate daily and annual air pollution exposures (including PM10, PMCoarse, but not Ozone {derived from closest station}) at participant residential and occupational addresses. Participants' time-activity diaries indicated time spent at either address to allow daily space-time-activity-weighted estimates, and capture total daily physical activity (total-PA {as metabolic-equivalents-of-task, METs}), in the three days preceding health measurements. Multivariate-adjusted linear mixed-effects models (using either annual or daily estimates) were adjusted for possible environmental confounders or mediators including levels of ambient noise and greenness. Causal mediation analysis was also performed separately considering these factors as well as total-PA. All presented models are controlled by age, height, sex and season. Results: An increase in 5μg/m3 of daily space-time-activity-weighted PMCoarse exposure was statistically significantly associated with a 4.1% reduction in total heart rate variability (SDNN; p=0.01), and remained robust after adjusting for suspected confounders [except for occupational-address noise (β=-2.7, p=0.20)]. An increase in 10ppb of annual mean Ozone concentration at the residential address was statistically significantly associated with an increase in diastolic BP of 6.4mmHg (p<0.01), which lost statistical significance when substituted with daily space-time-activity-weighted estimates. As for pulmonary function, an increase in 10μg/m3 of annual mean PM10 concentration at the residential address was significantly associated with a 0.3% reduction in FVC (p<0.01) and a 0.5% reduction in SUM (p<0.04), for which again significance was lost when substituted for daily space-time-activity-weighted estimates These associations with pulmonary function remained robust after adjusting for suspected confounders, including annual Ozone, as well as total-PA and bioaerosol (pollen and fungal spore) levels (but not residential-neighborhood greenness {β=-0.22, p=0.09; β=-0.34, p=0.15, respectively}). Multilevel mediation analysis indicated that the proportion mediated as a direct effect on cardiopulmonary outcomes by suspected confounders (including total-PA, residential-neighborhood greenness, and occupational-address noise level) from primary exposures (including PM10, PMCoarse, and O3) was not statistically significant. Conclusion: Our findings suggest that increased daily space-time-activity-weighted PMCoarse exposure levels significantly adversely affect cardiac autonomic modulation (as reduced total HRV) among healthy adults. Additionally, increased annual levels at the residential address of Ozone and PM10 significantly increase diastolic blood pressure and reduce lung function, respectively, among healthy adults. These associations typically remained robust when adjusting for suspected confounders. Occupational-address noise and residential-neighborhood greenness levels, however, were seen as mediators of cardiovascular and pulmonary outcomes, respectively. Total daily physical activity was not seen as a mediator of any of the studied outcomes, which supports the promotion of active mobility within cities.