Estimated Effects of Air Pollution and Space-Time-Activity on Cardiopulmonary Outcomes in Healthy Adults: A Repeated-measures Study

39 40 Background: Exposure to air pollution is known to affect both short and long-term 41 outcomes of the cardiopulmonary system; however, findings on short-term outcomes 42 have been inconsistent and often from isolated and long-term rather than coexisting and 43 short-term exposures, and among susceptible/unhealthy rather than healthy populations. 44 Aims: We aimed to investigate separately the annual, daily and daily space-time-activity45 weighted effect of ambient air pollution, as well as confounding or modification by other 46 environmental (including noise) or space-time-activity (including total physical activity) 47 exposures, on cardiopulmonary outcomes in healthy adults. 48 Methods: Participants (N=57: 57% female) had indicators of cardiopulmonary outcomes 49 [blood pressure (BP), pulse (HR) and heart rate variability (HRV {SDNN}), and lung 50 function (spirometry {FEV1, FVC, SUM})] measured on four different mornings (at least 51 five days apart) in a clinical setting between 2011 and 2014. Spatiotemporal ESCAPE52 LUR models were used to estimate daily and annual air pollution exposures (including 53 PM10, PMCoarse, but not Ozone {derived from closest station}) at participant residential 54 and occupational addresses. Participants’ time-activity diaries indicated time spent at 55 either address to allow daily space-time-activity-weighted estimates, and capture total 56 daily physical activity (total-PA), in the three days preceding health measurements. 57 Multivariate-adjusted linear mixed-effects models (using either annual or daily estimates) 58 were adjusted for possible environmental confounders or mediators including levels of 59 neighborhood noise and greenness. Causal mediation analysis was also performed 60 separately considering these factors as well as total-PA (as metabolic-equivalents-of-task 61 {METs}). All presented models are controlled by age, height, sex and season. 62 Results: An increase in 5 μg/m3 of daily space-time-activity-weighted PMCoarse exposure 63 was statistically significantly associated with a 4.1% reduction in total heart rate 64 variability (SDNN; p=0.01)), and remained robust after adjusting for suspected 65 confounders [except for occupational-address noise (β=-2.7, p=0.20)]. An increase in 10 66 ppb of annual mean Ozone concentration at the residential address was statistically 67 significantly associated with an increase in diastolic BP of 6.4 mmHg (p<0.01), which 68 lost statistical significance when substituted with daily space-time-activity-weighted 69 estimates. As for pulmonary function, an increase in 10 μg/m3 of annual mean PM10 70 concentration at the residential address was significantly associated with a 0.3% 71 reduction in FVC (p<0.01) and a 0.5% reduction in SUM (p<0.04), for which again 72 significance was lost when substituted for daily space-time-activity-weighted estimates 73 These associations with pulmonary function remained robust after adjusting for 74 suspected confounders, including annual Ozone, as well as total-PA and bioaerosol 75 (pollen and fungal spore) levels (but not residential address greenness {β= -0.22, p= 76 0.09; β= -0.34, p= 0.15, respectively}). Multilevel mediation analysis indicated that the 77 proportion mediated as a direct effect on cardiopulmonary outcomes by suspected 78 confounders (including total-PA, residential-greenness, and occupational-noise level) 79 from primary exposures (including PM10, PMCoarse, and O3) was not significant. 80 Conclusion: Our findings suggest that increased daily space-time-activity-weighted 81 PMCoarse levels significantly affect cardiac autonomic modulation (as reduced HRV). 82 Additionally, increased annual levels at the residential address of Ozone and PM10 83 significantly increase diastolic BP and reduce FVC/SUM, respectively, among healthy 84 adults. These associations typically remained robust when adjusting for suspected 85 confounders. Occupational-address noise and residential-neighborhood greenness 86 levels, however, were seen as mediators of HRV and FVC/SUM, respectively. Total-PA 87 was not seen as a mediator of any of the studied outcomes. 88