Quantile Regression Analysis of the Distributional Effects of Air Pollution on Blood Pressure, Heart Rate Variability, Blood Lipids, and Biomarkers of Inflammation in Elderly American Men: The Normative Aging Study

Abstract

Background: Previous studies have observed associations between air pollution and heart disease. Susceptibility to air pollution effects has been examined mostly with a test of effect modification, but little evidence is available whether air pollution distorts cardiovascular risk factor distribution.

Objectives: This paper aims to examine distributional and heterogeneous effects of air pollution on known cardiovascular biomarkers.

Methods: A total of 1,112 men from the Normative Aging Study and residents of the greater Boston, Massachusetts, area with mean age of 69 years at baseline were included in this study during the period 1995-2013. We used quantile regression and random slope models to investigate distributional effects and heterogeneity in the traffic-related responses on blood pressure, heart rate variability, repolarization, lipids, and inflammation. We considered 28-day averaged exposure to particle number, PM2.5 black carbon, and PM2.5 mass concentrations (measured at a single monitor near the site of the study visits).

Results: We observed some evidence suggesting distributional effects of traffic-related pollutants on systolic blood pressure, heart rate variability, corrected QT interval, low density lipoprotein (LDL) cholesterol, triglyceride, and intercellular adhesion molecule-1 (ICAM-1). For example, among participants with LDL cholesterol below 80 mg/dL, an interquartile range increase in PM2.5 black carbon exposure was associated with a 7-mg/dL (95% CI: 5, 10) increase in LDL cholesterol, while among subjects with LDL cholesterol levels close to 160 mg/dL, the same exposure was related to a 16-mg/dL (95% CI: 13, 20) increase in LDL cholesterol. We observed similar heterogeneous associations across low versus high percentiles of the LDL distribution for PM2.5 mass and particle number.

Conclusions: These results suggest that air pollution distorts the distribution of cardiovascular risk factors, and that, for several outcomes, effects may be greatest among individuals who are already at high risk.

Citation: Bind MA, Peters A, Koutrakis P, Coull B, Vokonas P, Schwartz J. 2016. Quantile regression analysis of the distributional effects of air pollution on blood pressure, heart rate variability, blood lipids, and biomarkers of inflammation in elderly American men: the Normative Aging Study. Environ Health Perspect 124:1189-1198; http://dx.doi.org/10.1289/ehp.1510044.

Figure 1

Associations between traffic-related air pollutants and quantiles of the distributions of SBP and DBP (adjusted for temperature, relative humidity, sine and cosine terms of the days of the season, age, physician-diagnosed diabetes, body mass index, smoking status, cumulative cigarette pack-years, current use of statin, and current use of antihypertensive medications). The y-axes represent the outcome difference (in the outcome unit) for an IQR increase in exposure. IQR for particle number = 13,845 number per cm³, IQR for PM_2.5 black carbon = 0.43 μg/m³, and IQR for PM_2.5 mass = 4.0 μg/m³. The numbers next to each point estimate indicate the deciles. Error bars represent 95% bootstrap CIs.

Figure 2

Associations between traffic-related air pollutants and quantiles of the distributions of heart rate, SDNN, log(LF:HF ratio), and corrected QT interval (adjusted for temperature, relative humidity, sine and cosine terms of the days of the season, age, physician-diagnosed diabetes, body mass index, smoking status, cumulative cigarette pack-years, current use of statin, and current use of antihypertensive medications). For SDNN, we also controlled for heart rate because standard deviation is likely to be larger as heart rate increases. The y-axes represent the outcome difference (in the outcome unit) for an IQR increase in exposure. IQR for particle number = 13,845 number per cm³, IQR for PM_2.5 black carbon = 0.43 μg/m³, and IQR for PM_2.5 mass = 4.0 μg/m³. The numbers next to each point estimate indicate the deciles. Error bars represent 95% bootstrap CIs. Note: ms, millisecond; s, second.

Figure 3

Associations between traffic-related air pollutants and quantiles of the distributions of HDL cholesterol, LDL cholesterol, and triglycerides (adjusted for temperature, relative humidity, sine and cosine terms of the days of the season, age, physician-diagnosed diabetes, body mass index, smoking status, cumulative cigarette pack-years, and current use of statin). The y-axes represent the outcome difference (in the outcome unit) for an IQR increase in exposure. IQR for particle number = 13,845 number per cm³, IQR for PM_2.5 black carbon = 0.43 μg/m³, and IQR for PM_2.5 mass = 4.0 μg/m³. The numbers next to each point estimate indicate the deciles. Error bars represent 95% bootstrap CIs.

Figure 4

Associations between traffic-related air pollutants and quantiles of the distributions of fibrinogen, C-reactive protein, ICAM-1, and VCAM-1 (adjusted for temperature, relative humidity, sine and cosine terms of the days of the season, age, physician-diagnosed diabetes, body mass index, smoking status, cumulative cigarette pack-years, and current use of statin). The y-axes represent the outcome difference (in the outcome unit) for an IQR increase in exposure. IQR for particle number = 13,845 number per cm³, IQR for PM_2.5 black carbon = 0.43 μg/m³, and IQR for PM_2.5 mass = 4.0 μg/m³. The numbers next to each point estimate indicate the deciles. Error bars represent 95% bootstrap CIs.