Ambient Air Pollution and Pulse Wave Velocity in Patients with Hypertension Treated with Intensive vs Standard Blood Pressure Control

Ambient air pollution (especially fine particulate matter <2.5 μm, PM_2.5) has been linked to increased cardiovascular risk¹. Increase in blood pressure has been hypothesized to be one mechanism, and has indeed been noted in both humans and experimental models¹. In the systolic blood pressure (BP) intervention trial (SPRINT), we have previously shown that the CV risk reduction due to intensive BP lowering was more robust among individuals living in areas with higher air pollution levels². We sought to investigate the link between ambient particulate matter (PM_2.5) and change in pulse wave velocity (PWV), a marker of arterial stiffness, and the impact of intensive BP lowering in SPRINT participants. We hypothesized that differential responsiveness of arterial stiffness to BP lowering in relation to chronic PM_2.5 exposure, may shed light on mechanisms by which air pollution modulates BP and cardiovascular risk.

A subset of SPRINT participants underwent PWV measurements at baseline, and at 3 years using the Sphygmocor CPV® system (AtCor Medical, Naperville, IL)³ and using unattended blood pressure measurements performed at the randomization visit. Using residential ZIP code of participant’s residence, we linked participants with integrated satellite-based ambient PM_2.5. PM_2.5 estimates were derived from a model that includes satellite-based methods (aerosol optical depth), chemical transport models, and adjustment for ground monitors as we previously described². We analyzed patients by PM_2.5 levels as a continuous variable and using threshold (<12 vs ≥12 μg/m³) corresponding to annual National Ambient Air Quality Standards (NAAQS). Paired t-tests were used to analyze changes over time in PWV. Spearman’s correlations and linear regression analyses were performed to investigate the association between PM_2.5 (both at entry and mean annual levels during the trial), baseline PWV, and change in PWV through year 3 (ΔPWV). Three linear regression models were performed to evaluate the association between PM_2.5 and ΔPWV: unadjusted model, model 1 (adjusted for age, sex, race) and model 2 (model 1 + total cholesterol, baseline systolic and diastolic blood pressure, Clinical/subclinical cardiovascular disease, baseline eGFR, smoking status, and randomization arm, body mass index, heart rate, and baseline pulse wave velocity).

A total of 517 participants were included (259 randomized to standard blood pressure, and 258 to intensive blood pressure). Mean PM_2.5 was 9.5±1.8 μg/m³ and mean baseline PWV was 10.7±2.7 m/s. Over 3 years, PWV increased by a mean of 0.27 m/sec from baseline (P=0.016). There was no correlation between PM_2.5 and PWV at baseline (Spearman’s ρ=0.01, P=0.84). PM_2.5, however, significantly correlated with ΔPWV (Spearman’s ρ =0.26, P<0.001). The associations between PM_2.5 and ΔPWV were similar in the intensive (Spearman’s ρ =0.29, P<0.001) and standard BP arms (Spearman’s ρ=0.25, P<0.001), Figure 1A. No significant interaction was noted between PM_2.5 and trial assignment (intensive vs standard BP control) with respect to ΔPWV (PM_2.5*Assignment, P_interaction=0.74). Participants living in areas with PM_2.5 levels above the NAAQS threshold of 12 μg/m³ (n=61) had higher ΔPWV compared with participants living below NAAQS threshold (n=456), ΔPWV 1.3 m/s vs 0.13 m/s, P=0.001. After multiple adjustments (model 2), entry and mean PM_2.5 remained associated with ΔPWV (per 1 μg/m³ of PM_2.5, entry PM_2.5: [β 0.408, standard error 0.058, P<0.001]; mean PM_2.5: [β 0.315, standard error 0.053, P<0.001]), figure 1B.

Figure 1:

PM_2.5 and Pulse Wave Velocities in SPRINT. (A) Scatter plot of mean PM_2.5 and change in PWV over 3 years. (B) unadjusted and adjusted association between PM_2.5 with change in PWV at 3 years. Model 1: Age, sex, Black race. Model 2: Model 1 + total cholesterol, baseline systolic and diastolic blood pressure, Clinical/subclinical cardiovascular disease, baseline eGFR, smoking status, and randomization arm, body mass index, heart rate, and baseline pulse wave velocity

Arterial stiffness is an established risk factor for cardiovascular disease⁴, and has been proposed as a pathway linking pollution-triggered cardiovascular events. Although exposure to air pollutants have demonstrated changes in acute arterial stiffness, results from epidemiologic studies with longer exposures have been mixed.¹ Our analysis, from a randomized controlled trial that enrolled moderate to high-risk patients with hypertension, demonstrated a significant association between PM_2.5 and PWV after controlling for baseline characteristics. Higher progression in PWV was seen in areas with PM_2.5 levels above NAAQS, but the relationship between PM2.5 and ΔPWV continued to very low exposure levels. Importantly, intensive BP lowering did not modify this association. Prior studies involving exposure to a variety of pollutants including diesel exhaust, have noted change in arterial stiffness, through mechanisms that involve hemodynamic changes and alterations in nitric oxide signaling.^{1, 5} The fact that BP lowering did not alter the association between PM_2.5 and ΔPWV could suggest that the pathways that lead to arterial stiffness may involve non-hemodynamic pathways.

This post-hoc study has limitations and should be interpreted as hypothesis-generating including potential exposure misclassification, patient bias by virtue of enrolling in this ancillary study and small cohort size. Nevertheless, the longitudinal structure of this study, and the randomized design in a carefully performed trial lends unique aspects to these data.

In conclusion, SPRINT participants who live in areas with higher ambient air pollution (PM_2.5) levels experienced larger increase in PWV over 3 years, without a significant impact of intensive BP. Our study findings provide a mechanistic basis for elevated cardiovascular risk at exposure levels well below NAAQS.