Association Between Long-term Exposure to Ambient Air Pollution and Lesion Ischemia in Patients with Atherosclerosis

Abstract

Background and aims:

Air pollution has been associated with coronary artery disease. The underlying mechanisms were understudied, especially in relation to coronary stenosis leading to myocardial ischemia. Advances in computed tomography (CT) allow for novel quantification of lesion ischemia. We aim to investigate associations between air pollution exposures and fractional flow reserve on CT (CT-FFR), a measure of coronary artery blood flow.

Methods:

CT-FFR, which defines a ratio of maximal myocardial blood flow compared to its normal value (range: 0–100%), was characterized in 2017 patients with atherosclerosis between 2015 and 2017. Exposures to ozone (O₃), nitrogen dioxide (NO₂), and fine particulate matter (PM_2.5), were estimated using high-resolution exposure models. Linear and logistic regression models were used to assess the association of each air pollutant with CT-FFR and with the prevalence of clinically relevant myocardial ischemia (CT-FFR <75%).

Results:

Participants were on average 60.1 years old. Annual mean O₃, NO₂, PM_2.5 were 61, 47 and 60 μg/m3, respectively. Mean CT-FFR value was 76.9%. In the main analysis, a higher level of O₃ was associated with a lower CT-FFR value (−1.74%, 95% CI: −2.85, −0.63 per 8 μg/m3) and a higher prevalence of myocardial ischemia (odds ratio: 1.32, 95%CI: 1.05–1.65), adjusting for potential confounders such as risk factors and plaque phenotypes, independent of the effects of exposure to NO₂ and PM_2.5. No associations were observed for PM_2.5 or NO₂ with CT-FFR.

Conclusions:

Long-term exposure to O₃ is associated with lower CT-FFR value in atherosclerotic patients, indicating higher risk of lesion ischemia.

Graphical Abstract

INTRODUCTION

Coronary heart disease (CHD) is the leading cause of death worldwide and is also the top-ranked cause of disability-adjusted life years (DALYs) in individuals over 50 years old¹. Air pollution is an emerging factor in determining the risk of adverse coronary events². Evidence from meta-analyses and large-scale cohort studies has shown that long-term exposures to ambient air pollutants [e.g., fine particulate matter (PM_2.5), nitrogen dioxide (NO₂), and ozone (O₃)] are associated with an increased risk of cardiovascular disease morbidity and mortality³. Cardiovascular risks associated with air pollution exposures are especially elevated for susceptible population with pre-existing coronary artery disease (CAD)⁴.

CHD typically evolves after atherosclerotic plaques develop in the walls of coronary arteries, and vulnerable or ruptured plaques lead to thrombosis, which obstructs coronary blood flow and triggers myocardial ischemia, leading to adverse clinical cardiac events and reduced survival⁵. While pathophysiological mechanisms through which air pollution may accelerate progression of atherosclerosis and potentially cause clinical CHD have not been adequately investigated, recent studies have suggested that air pollution may affect plaque formation and morphologic characteristics including calcification and other aspects of plaque composition⁶. Few studies have investigated the effect of air pollution on both the anatomic and functional hemodynamic characteristics of coronary plaques and these together might induce lesion ischemia, especially in patients with CAD. Fractional flow reserve (FFR) measured invasively by coronary angiography is the established reference standard for evaluating the functional severity of coronary stenosis⁷. However, due to its invasive nature, the implementation of FFR is limited among patients with severe CAD. Noninvasive FFR can now be derived from coronary computed tomography angiography (CCTA) using a physiological simulation technique (CT-FFR). It is an emerging validated method for evaluating lesion-specific ischemia with high accuracy⁸. This noninvasive technique may be preferable when evaluating asymptomatic patients with suspected CAD. Recent findings from systematic review and meta-analysis provided evidence that decreased CT-FFR is associated with an increased risk of all-cause death and myocardial infarction (MI)⁹.

Although there is increasing evidence for associations of long-term air pollution exposure with CHD incidence and mortality in China¹⁰, few studies have focused on the underlying mechanistic pathways. Leveraging the availability of an established cohort of patients with a documented presence of coronary atherosclerosis based on CCTA, the present study aimed to fill the gaps in physiological mechanisms between air pollution and CHD by examining the associations of long-term exposure to ambient air pollutants with functional coronary stenosis and lesion ischemia assessed by novel CT-FFR.

METHODS

Study population

Participants in this study were a part of an ongoing multi-center prospective cohort study in China, focused on identifying risk factors for CAD in patients with documented coronary atherosclerosis. Details of this cohort study have been described previously¹¹. Briefly, the cohort recruited Chinese participants who were suspected of low to moderate risk of CAD according to the American College of Cardiology (ACC)/ American Heart Association (AHA) guidelines for atherosclerotic cardiovascular diseases (ASCVD) risk and underwent CCTA imaging between 2015 to 2017. Patients with a prior history of coronary events, such as coronary revascularization, myocardial infarction, and other heart diseases, were excluded. We focused on atherosclerotic patients who had any plaques detected at baseline entry into the cohort (i.e., total plaque volume > 0) identified via cardiac imaging. Participants without plaque may represent other cardiovascular diseases (CVD) other than CAD and therefore may reduce the precision of our results. In this analysis, we included all participants residing in Beijing city and Hebei Province who had validated CT-FFR data collected from six major cardiovascular hospitals.

General cardiovascular risk factors were measured at baseline using self-administered questionnaires (gender, age, education, smoking, residence location), face-to-face interviews (measurements of height and weight, medication use), clinical examinations (hypertension, diabetes mellitus, hyperlipidemia, plaque phenotypes by CCTA), and biochemical laboratory tests according to standard protocols (total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), high-sensitivity C-reactive protein and triglycerides (TG)). Area-level information was derived from residential address (population density, percentage of dependents, Gross domestic product (GDP), education level, rural or urban living). All participants provided written informed consent upon enrolment. The study was approved by the institutional review board (IRB) of the Chinese Academy of Medical Sciences, Fuwai Hospital in Beijing, China, and the University at Buffalo, NY, USA.

Exposure estimation

The main exposures in our study were three criteria air pollutants: PM_2.5, NO₂, and O₃. These three air pollutants are strictly regulated in China due to their important adverse health effects on residents. Air pollutant concentrations were estimated using a hierarchical machine-learning based modeling approach with a high spatial resolution (within the 6^th ring road: 100m; outside: 1km)^{12, 13}. The model incorporated continuous daily measurements from 1419 regulatory monitors nationwide between 2013 to 2019, and along with 292 predictor variables from a wide variety of geographic features obtained from multiple sources (e.g., traffic network, industrial emissions, population density, and land use), satellite-derived predictors and meteorological data, with regression residuals smoothed by universal kriging. The model exhibited good performances which explained 89%, 78%, and 75% variations of PM_2.5, NO₂, and O₃ based on out-of-sample validation. Our primary long-term exposure is the daily air pollutant levels averaged over two years during and before clinical exams at the participants’ homes. We excluded the participants whose addresses were not validated or whose exposure predictions were deemed invalid.

Outcome measurement

FFR is a sub-clinical measure of CAD that assesses coronary blood flow limitation. FFR defines the ratio of maximal myocardial blood flow compared to its normal value (range: 0–100%), with higher values indicating better blood flow¹⁴. An FFR value of < 75% is considered indicative of blood flow obstruction associated with inducible coronary ischemia¹⁴. An FFR value of 85% indicates no hemodynamically relevant obstruction. Values between 75% and 85% are defined as intermediate for diagnostic stratification. FFR is derived from the non-invasive routine diagnostic coronary CT angiography, using a novel image analysis to calculate CT-FFR via computational fluid dynamic modeling¹⁵ (see details in Figure S1). This method contrasts with the established but invasive angiographic reference standard for functional assessment of FFR via the pressure wire technique. Compared to the pressure wire method, non-invasive CT-FFR has been found to have a sensitivity of 91% and specificity of 90% for the identification of lesion-specific ischemia¹⁶. In this study, the primary outcome is the minimum FFR (CT-FFR_min) among the three coronary arteries combined and the secondary outcome is the FFR derived separately for the three coronary arteries, which are the left anterior descending coronary artery (CT-FFR_LAD), left circumflex coronary artery (CT-FFR_LCX), and right coronary artery (CT-FFR_RCA). Moreover, we assessed the presence of lesion-specific coronary ischemia determined by an FFR value less than 75% at individual coronary arteries and counted the total number of vessels with coronary ischemia, respectively.

Statistical analysis

We analyzed the associations between long-term ambient air pollution exposures (PM_2.5, NO₂ and O₃) and CT-FFR (continuous variable) by performing separate, single-pollutant, multiple linear regression models. In addition, binary and ordinal logistic regression models were used to assess the relationships between air pollution exposures and presence (yes vs. no) and severity (presence in the number of vessels from 0 to 3) of coronary ischemia respectively excluding patients with CT-FFR intermediate values (75%-85%).

The multivariable statistical models were developed in stages, by incrementally adding different sets of covariates based on an a priori understanding of other studies of risk factors for CAD and of exposures to air pollution¹⁷. The main model included individual-level variables as follows: age, gender, body mass index (BMI), education, smoking (status, intensity [cigarettes per day], and duration [years]), and Beijing residence. We also included individual comorbid conditions such as diabetes, hypertension, and hyperlipidemia. To address potential behavioral and socio-economic confounding due to geography, we added area-level variables: urban population (≥2500 population per 1 × 1-km² grid), county-level GDP, educational attainment (percentage of the population with college or a higher level of education) and dependency level (percentage of the number of households with children and elderly members) in the main models. To evaluate the independent effects of air pollution through the functional stenosis pathway, we additionally adjusted subclinical measures of plaque phenotypes [(i.e. total plaque burden and presence of high-risk plaques (HRP)] and lumen stenosis in major vessels, evaluated by CCTA (see supplement for a detailed description of the variables). Medication uses (Statin and antihypertensive), and family history of CVD determined by questionnaire were also included as covariates.

Potential effect modification was evaluated by the interaction between air pollution and specific participant characteristics or risk factors including age (<65 vs. ≥65 years), gender (male vs. female), BMI (<24 kg/ m² vs. ≥24kg/ m²), smoking status (never/former vs. current), presence of hypertension (yes vs. no), diabetes mellitus (yes vs. no), CAD (CCTA stenosis >50% or not) or HRP (appearance or not), statin usage (yes vs. no), antihypertensive drugs usage (yes vs. no), Beijing resident (yes or no), urban area living (yes or no). Stratified analyses within subgroups were also conducted.

We performed several sensitivity analyses. Firstly, the main model was fitted with all three air pollutants to assess the independent associations of each pollutant with the primary CT-FFR outcome variable. Secondly, we fitted the main model with biomarkers of clinical conditions (total cholesterol, triglycerides with the natural logarithmic transformation, HDL, LDL, and glucose levels). Thirdly, a generalized additive model (GAM) model with spline function (degrees of freedom = 3) was built and compared with the linear model results according to Akaike Information Criterion (AIC) value. Fourthly, a mixed effect model included zip code as a random intercept. Lastly, we fitted the main model with complete data with missing values imputed by inverse probability weighting (IPW)¹⁸ or Multivariate Imputation by Chained Equations (MICE)¹⁹ methods. All analyses were performed twice and checked by two researchers in SAS 9.4, Stata IC 16.0, and R 4.1.2.

RESULTS

Study population and environmental exposures

There were 2,017 eligible participants with exposure and outcome data. As shown in Table 1, the participants were on average 60.1 years old, 62.5% were male, and 86.9% had an education lower than college. Among the participants with CHD risk factors, 39.8% were current smokers with an average history of 10 years; 56.8% and 19.8% had a diagnosis of hypertension or diabetes mellitus, respectively. Participants were mainly from Beijing city (46.8%) and resided in an urban area (69.7%). The mean (standard deviation, SD) minimum CT-FFR value across the three coronary arteries combined was 76.9% (SD: 12.8%). Coronary ischemia (CT-FFR_min<75%) was identified in 33.2% of the participants, of whom 2.0% had hemodynamically significant obstructions in all three vessels.

Table 1.

Characteristics of the study participants (N = 2017)

Characteristics	Statistic n (%) or mean (standard deviation, SD)	Ischemic patientsa	Non-ischemic patientsa	P-valued
Demographics
Male, n (%)	1261 (62.5)	424 (63.6)	837 (62)	0.494
Age (C), mean (SD)	60.1 (9.81)	60.2 (9.66)	60.0 (9.78)	0.434
BMI (kg/m2), mean (SD)	25.5 (3.41)	25.4 (3.27)	25.5 (3.44)	0.624
Education, n (%)
Did not attend college	1753 (86.9)	579 (86.8)	1174 (87.0)	0.524
College	236 (11.7)	76 (11.4)	160 (11.9)
Post-graduate	28 (1.4)	12 (1.8)	16 (1.2)
Risk factors
Current smoker, n (%)	803 (39.8)	275 (41.2)	528 (39.1)	0.361
Smoking intensity (cigarettes per day), n (%)
≤10	182 (9.0)	56 (8.4)	126 (9.3)	0.451
10 – 20	451 (22.4)	158 (23.7)	293 (21.7)
≥20	171 (8.5)	63 (9.4)	108 (8.0)
Smoking duration (years), mean (SD)	10.54 (16.02)	11.19 (16.20)	10.22 (15.93)	0.201
Hypertension, n (%)	1146 (56.8)	382 (57.3)	764 (56.6)	0.084
Hyperlipidemia, n (n%)	952 (47.2)	340 (51.0)	612 (45.3)	0.017 *
Diabetes, n (%)	399 (19.8)	119 (17.8)	280 (20.7)	0.124
Antihypertensive drugs, n (%)	817 (40.5)	290 (43.5)	527 (39.0)	0.056
Statin, n (%)	297 (14.7)	105 (15.7)	192 (14.2)	0.242
Total plaque burden, mean (SD)	25.95 (13.22)	26.68 (13.68)	25.59 (12.98)	0.082
High-risk plaque presence, n (%)	420 (20.8)	152 (22.8)	268 (19.9)	0.126
Lumen diameter stenosis, mean (SD)	44.83 (31.19)	44.33 (30.72)	45.08 (31.42)	0.612
Biomarkers
Total cholesterol (mmol/L), mean (SD)b	4.71 (1.34)	4.77 (1.23)	4.68 (1.31)	0.139
Triglycerides (mmol/L), mean (SD)b	1.78 (1.28)	1.79 (1.34)	1.77 (1.35)	0.754
HDL (mmol/L), mean (SD)b	1.25 (0.89)	1.29 (1.40)	1.21 (0.38)	0.050
LDL (mmol/L), mean (SD)b	2.97 (2.33)	3.10 (3.65)	2.91 (1.03)	0.076
Glucose (mmol/L), mean (SD)b	5.87 (1.93)	5.97 (2.22)	5.91 (1.94)	0.534
Area-level variables c
Beijing resident, n (%)	944 (46.8)	308 (46.2)	636 (47.1)	0.692
Urban resident, n (%)	1406 (69.7)	456 (68.4)	950 (70.4)	0.357
% Dependency, mean (SD)	21.1 (4.2)	21.2 (4.3)	21.1 (4.1)	0.612
GDP, mean (SD)	1.10×106 (2.10 ×106)	0.98×106 (1.93× 106)	1.15×106 (2.17×106)	0.086
% College and higher, mean (SD)	22.9 (16.6)	22.5 (16.22)	23.1 (16.79)	0.444
Outcomes
CT-FFRmin (%), mean (SD)	76.9 (12.8)	64.6 (12.8)	83.0 (7.2)	<0.001
Coronary ischemia
Overall, n (%)	667 (33.2%)	NA	NA
1 vessel, n (%)	450 (22.5%)	NA	NA
2 vessels, n (%)	176 (8.7%)	NA	NA
3 vessels, n (%)	41 (2.0%)

^aIschemic participants: participants with any vessel’s CT-FFR below 75%; non-ischemic participants: participants with all vessels’ CT-FFR above 85%;

^bThere is 5% missing data in the biomarker data among the participants;

^cArea-level variables as proxies for behavioral or socioeconomic confounders.

^dWe apply chi-square test for categorical variables with count and T-test for continuous variables with mean and standard deviation.

Bi-annually averaged concentrations of PM_2.5, NO₂ and O₃ were generally higher (mean: 60.1 μg/m³ for PM_2.5, 47.3 μg/m³ for NO₂, 61.1 μg/m³ for O₃) across the participants’ homes than the national safety level (Table 2). Moreover, the exposure levels varied considerably across the study area, especially for PM_2.5, which ranged from 19.8 μg/m³ to 101.6 μg/m³.

Table 2.

Summary statistics of air pollutant exposures for the study population (N = 2017)

Exposure	Min	P25	P50	P75	Max	Correlation coefficient
						PM2.5	NO2	O3
PM2.5 (μg/m3)	19.8	46.4	62.1	72.9	101.6	1
NO2 (μg/m3)	15.6	36.7	51.2	58.8	67.7	0.78	1
O3 (μg/m3)	31.9	57.0	58.3	65.3	82.7	−0.29	−0.57	1

Air pollution and change in CT-FFR

Primary results are shown in Figure 1 with detailed staged model results in Table S1. A representative example of a patient with coronary ischemia (CT-FFR_LAD) who was exposed to high concentrations of O₃ is shown in Figure S2. Higher exposure to ambient O₃ was associated with lower CT-FFR_min value (−1.74% per 8 μg/m³ O₃, 95% CI: −2.85%, −0.63%), indicating significantly lower blood flow across the three coronary arteries combined, after adjusting for the full set of covariates. Statistically significant inverse associations were also evident with CT-FFR for each vessel. Associations were marginally different with additional adjustments for blood biomarkers (Table S1). Similarly, including all the exposure variables together in the same model did not have much impact on the associations for O₃. As shown in Figure S3, there is suggestive evidence of a nonlinear relationship between O₃ and CT-FFR_min (AIC = 2.4 comparing the nonlinear and linear model). The concentration-response curve shows a steeper decline in CT-FFR_min when the O₃ level is higher than 65 μg/m³. No associations were found for NO₂ or PM_2.5 with CT-FFR_min in the main models. A sensitivity analysis that fitted the main model using two imputed datasets (20% of individuals with missing covariates were imputed) showed similar associations for O₃ (IPW: −1.11% per 8 μg/m³ O₃, 95% CI: (−2.08%, −0.14%), MICE: −1.20% per 8 μg/m³ O₃, 95%CI: −2.21%, −0.19%)) and no association for NO₂ or PM_2.5 (Figure S4).

Figure 1.

Association between air pollutants exposure and degree of CT-FFR values in main model. CT-FFR was estimated as the minimum FFR (CT-FFR_min) among the three coronary arteries combined and those derived separately for the three coronary arteries, including the left anterior descending coronary artery (CT-FFR_LAD), left circumflex coronary artery (CT-FFR_LCX), and right coronary artery (CT-FFR_RCA).

Figure 2 shows the logistic regression analyses with binary outcomes. An interquartile (IQR) increase in O₃ exposure (8 μg/m³) was associated with significantly higher odds of having myocardial ischemia defined by CT-FFR < 75% (odds ratio: 1.32, 95%CI: 1.05, 1.65). No associations were found between the other exposure variables and the presence of myocardial ischemia (PM_2.5: 0.93, 95%CI: 0.72, 1.21; NO₂: 0.81, 95%CI: 0.56, 1.18) (Table S2). Similarly, no associations were observed between any air pollutants and the severity of myocardial ischemia across the three coronary arteries using ordinal logistic regression.

Figure 2.

Association between air pollutant exposures and presence of coronary ischemia in main model. Coronary ischemia was defined as CT-FFR<0.75 at individual vessels (CT-FFR_LAD, CT-FFR_LCX, CT-FFR_RCA) and those combined (CT-FFR_min).

Effect modification

In the effect modification analyses (Table 3), we found stronger associations between O₃ and CT-FFR_min among patients without diabetes mellitus (interaction p-value = 0.009) and those who live outside urban areas (interaction p-value = 0.022). There were no clear differences in the pollutant effect estimates across the subgroups of other personal characteristics.

Table 3.

Effect modifications in associations between O₃ exposure and CT-FFR_min from selected demographics and risk factors

	N	Effect Estimate (%) per 8 μg/m3	P value
			0.12
Male	1011	−1.08 (−2.50, 0.33)
Female	575	−2.94 (−4.79, −1.12)
			0.43
≤65 years	1054	−1.72 (−3.10, −0.35)
>65 years	532	−1.90 (−3.77, 0.03)
			0.07
BMI <24	494	−1.18 (−3.21, 0.84)
BMI≥24	1092	−2.19 (−3.52, −0.86)
			0.08
Non/former smokers	943	−1.39 (−2.80, 0.01)
Smokers	643	−2.43 (−4.25, −0.60)
			0.42
Statin (No)	1349	−1.86 (−3.03, −0.69)
Statin (Yes)	237	−1.48 (−4.97, 2.00)
			0.31
Antihypertensive (No)	918	−2.47 (−3.91, −1.02)
Antihypertensive (Yes)	668	−0.96 (−2.71, 0.79)
			0.01
Diabetes (No)	1259	−2.39 (−3.62, −1.17)
Diabetes (Yes)	327	1.67 (−0.97, 4.31)
			0.06
Hypertension (No)	667	−1.48 (−3.13, 0.17)
Hypertension (Yes)	919	−2.09 (−3.59, −0.59)
			0.49
Lumen stenosis (No)	804	−2.18 (−3.78, −0.57)
Lumen stenosis (Yes)	782	−0.94 (−2.48, 0.59)
			0.37
High-risk plaque (No)	1273	−2.18 (−3.42, −0.94)
High-risk plaque (Yes)	313	0.21 (−2.30, 2.73)
			0.02
Urban area (No)	443	−2.82 (−4.27, −1.38)
Urban area (Yes)	1143	0.20 (−1.86, 1.45)
			0.11
Beijing resident (No)	819	−1.71 (−2.90, −0.52)
Beijing resident (Yes)	767	−5.16 (−9.80, −0.51)

DISCUSSION

In the study of adults with coronary atherosclerosis, we found that higher levels of long-term ambient O₃ concentrations were associated with reduced CT-FFR across all major vessels, after controlling for relevant demographic and clinical covariates, suggesting that this pollutant results in greater blood flow obstruction. The associations of other pollutants (NO₂ and PM_2.5) concentrations with CT-FFR were marginal and non-significant (Fig 3). Furthermore, higher levels of O₃ were associated with a greater prevalence of myocardial ischemia (CT-FFR < 75%). Given that FFR is a predictor of functional blood flow obstruction, arterial stenosis, and clinical CHD events¹⁵, this study adds new support that functional coronary artery stenosis and the resulting myocardial ischemia may be a potential pathway by which O₃ air pollution exposure could impose adverse cardiac risks.

Although the evidence linking long-term air pollution exposure to coronary physiology is relatively new, plaque formation and plaque instability are likely culprits. Findings are generally consistent in associations between exposure to PM_2.5 and plaque calcification scores (e.g., coronary artery calcification (CAC) and thoracic aortic calcification (TAC)) on CT, a standard subclinical marker for coronary atherosclerosis, among both general population and clinical patients worldwide^{6, 11, 20}. More recently, a few studies, including our cohort, have provided more insights into plaque phenotypes and ruptures, suggesting that higher PM_2.5 exposure is associated with increased total plaque burdens (both calcified and non-calcified) and high-risk plaque formation assessed by CCTA in patients suspected of CAD^21–23. Conversely, our previous study suggested that O₃ exposure, not PM_2.5, was associated with greater coronary stenosis of lumen diameters by CCTA in the same cohort²³. The finding is further supported by this study suggesting that long-term exposure to O₃, not PM_2.5 or NO₂, may affect the hemodynamic significance of a lesion as a consequence of coronary stenosis. This suggests that the pathogenesis of CAD might be different for exposure to O₃ compared to PM_2.5. Atherosclerotic plaque characteristics do not fully explain the functional change in stenosis, as the same degree of structural stenoses by plaques on visual estimation may yield different FFR values²⁴. Therefore, other pathways (e.g., endothelial dysfunction and vascular permeability²⁵) may play important roles of which air pollution, especially O₃, may reduce coronary flow and lead to significant myocardial ischemia that can be identified non-invasively using CT-FFR. This is suggested by our study, as the association between O₃ and CT-FFR remains robust even after controlling for atherosclerotic plaque phenotypes.

We did not find associations between exposure to PM_2.5 or NO₂ and CT-FFR or the presence of coronary ischemia. In China, PM_2.5 and NO₂ concentrations have steadily decreased since 2013 due to pollution control measures, while O₃ levels have increased during our study period²⁶. Therefore, the chronic process of coronary ischemia may have been initiated by the high level and toxicity (due to industrial and coal-burning sources) of historical PM_2.5 exposure before the regulation took place while O₃ exposure in recent years may play a major role in advancing the functional stenosis. It will be important for future studies to explore critical exposure windows for the effects of different air pollutants on CT-FFR. The apparent faster decline in CT-FFR_min when the O₃ level is higher than 65 μg/m³ suggests that populations with CAD may need to avoid high O₃ exposures.

The mechanisms underlying the associations between O₃ exposure and blood flow limitation by CT-FFR are not well elucidated, although the impact of O₃ on the respiratory system is well-studied. One potential pathophysiological pathway is the generation of oxidative reaction products from the reaction of O₃ and resultant oxidant gases with lipids or cellular membranes in the lung, which are subsequently released into the circulatory system and initiate or propagate a systemic inflammatory response²⁷. Persistent activation of this pathway is associated with the release of endothelium-derived vasoconstrictor factors such as endothelin-1 which may potentially contribute to the reduced flow and the extension of myocardial ischemia²⁸.

An FFR < 75% is consistent with a clinically significant ischemic lesion, and these patients are usually recommended for a revascularization procedure²⁹. The associations observed in this study suggest that the risk of greater degree of lesion ischemia may be associated with higher O₃ concentrations in China and may underlie the observed association between O₃ exposure and coronary events³⁰.

STRENGTHS AND LIMITATIONS

This is the first study to assess the deleterious influence of ambient air pollutant exposures on coronary functional stenosis and lesion ischemia. Strengths of the study include a large sample size, reproducible and non-invasive FFR measurements by CCTA, which allow the inclusion of low-to-moderate risk patients with atherosclerosis, and accurate predictions for population exposure with high spatiotemporal resolution.

Our study has several limitations. First, we use outdoor air pollution as the exposure variable. Exposure to indoor air pollution and infiltration as well as individual activity patterns are therefore not taken into account. This may result in the misclassification of individual exposure and attenuate the study associations. Second, we focused on patients with atherosclerosis; therefore, our study may only be relevant to an at-risk population regarding the effects of pollution exposure on CAD. Third, this is a cross-sectional study involving the baseline examination of an ongoing cohort study. Lastly, while CT-FFR is a validated physiologic simulation technique, most underlying data were acquired from North American and European studies¹⁵. Therefore, uncertainty may be amplified by generalizing the same model algorithm to estimate CT-FFR in Chinese patients.

CONCLUSION

Exposure to higher levels of ambient O₃ was independently associated with increased blood flow resistance and myocardial ischemia characterized by FFR values derived from CCTA. These novel findings suggest important roles for functional stenosis and the resulting lesion ischemia to explain the increased risk of ambient O₃ exposure on coronary events.

Abbreviation list:

CT

Computed Tomography

FFR

Fractional Flow Reserve

CHD

Coronary Heart Disease

CAD

Coronary Artery Disease

O3

Ozone

PM2.5

Fine Particulate Matter

NO2

Nitrogen Dioxide

DALY

Disability-adjusted Life Years

CCTA

Coronary Computed Tomography Angiography

MI

Myocardial Infarction

ACC

American College of Cardiology

AHA

American Heart Association

ASCVD

Atherosclerotic Cardiovascular Diseases

HDL

high-density lipoprotein

LDL

low-density lipoprotein

BMI

Body Mass Index

HRP

High Risk Plaque

GAM

Generalized Additive Model

IPW

Inverse Probability Weighting

MICE

Multivariate Imputation by Chained Equations

IQR

Interquartile

CAC

Coronary Artery Calcification

TAC

Thoracic Aortic Calcification

DATA AVAILABILITY STATEMENT

Data available on reasonable request.