Content uploaded by Joel Schwartz
Author content
All content in this area was uploaded by Joel Schwartz on Oct 24, 2015
Content may be subject to copyright.
ORIGINAL ARTICLE
Association between long-term exposure to traffic
particles and blood pressure in the Veterans
Administration Normative Aging Study
Joel Schwartz,
1,2
Stacey E Alexeeff,
2,3
Irina Mordukhovich,
4
Alexandros Gryparis,
3
Pantel Vokonas,
5
Helen Suh,
2
Brent A Coull
3
ABSTRACT
Objectives Particulate air pollution is associated with
cardiovascular events, but the mechanisms are not fully
understood. The main objective was to assess the
relationship between long-term exposure to traffic-
related air pollution and blood pressure (BP).
Methods The authors used longitudinal data from 853
elderly men participating in the Veterans Administration
Normative Aging Study, followed during 1996e2008.
Long-term average exposures to traffic particles were
created from daily predictions of black carbon (BC)
exposure at the geocoded address of each subject, using
a validated spatiotemporal model based on ambient
monitoring at 82 Boston-area locations. The authors
examined the association of these exposures with BP
using a mixed model. The authors included the following
covariates: age, body mass index, smoking, alcohol,
fasting glucose, creatinine clearance, use of
cardiovascular medication, education, census-level
poverty, day of week and season of clinical visit.
Results The authors found significant positive
associations between 1-year average BC exposure and
both systolic and diastolic blood pressure. An IQR
increase in 1-year average BC exposure (0.32 mg/m
3
)
was associated with a 2.64 mm Hg increase in systolic
blood pressure (95% CI 1.47 to 3.80) and a 2.41 mm Hg
increase in diastolic blood pressure (95% CI 1.77 to
3.05).
Conclusions Long-term exposure to traffic particles is
associated with increased BP, which may explain part of
the association with myocardial infarctions and
cardiovascular deaths reported in cohort studies.
INTRODUCTION
Both short-term and long-term exposure to partic-
ulate air pollution has been associated with
cardiovascular morbidity and mortality in
numerous epidemiological studies.
1e6
The effect
sizes of long-term exposure are substantially larger
than those of short-term exposure, suggesting
differences in the mechanisms may at play or
differences in how the mechanisms are impacted by
longer-term exposures. A number of pathways have
been proposed to explain these associations,
including, at the molecular level, increased oxida-
tive stress,
78
systemic inflammation
910
and
thrombotic potential.
11
At the functional level,
potential pathways include changes in autonomic
function, which may result in changes in blood
pressure (BP).
12
Elevated BP is an established risk factor for
coronary heart disease and stroke and an important
intermediate marker of cardiovascular health. The
relationship between air pollution exposure and BP
is still not well understood. Studies of short-term
PM exposure and BP show mixed results, with
some studies showing an inverse association or no
association
13e15
and positive findings in other
studies.
16e20
A key to understanding the mixed
results in the observed health effects of PM is that
PM is a complex mixture and the concentrations of
its individual components vary regionally and
seasonally.
Growing evidence suggests that traffic-related
components of PM pollution contribute signifi-
cantly to particle-related cardiovascular effects. For
example, a recent chamber study examining the
What is known about this subject
Short-term air pollution exposures are associated
with adverse cardiovascular effects, but studying
the effects of longer-term exposures requires more
complex exposure modelling.
What this paper adds
Long-term exposure to traffic-related air pollution is
associated with increases in blood pressure,
a finding that could explain part of the association
of particulate air pollution with cardiovascular
mortality.
Policy implications
This work impacts regulatory decisions about the
level of traffic-related air pollution that affects
cardiovascular health.
<An additional figure is
published online only. To view
this file please visit the journal
online (http://oem.bmj.com/
content/69/6.toc).
1
Department of Epidemiology,
Harvard School of Public Health,
Boston, Massachusetts, USA
2
Department of Environmental
Health, Harvard School of Public
Health, Boston, Massachusetts,
USA
3
Department of Biostatistics,
Harvard School of Public Health,
Boston, Massachusetts, USA
4
Department of Epidemiology,
University of North Carolina,
Chapel Hill, North Carolina, USA
5
Department of Medicine, VA
Normative Aging Study,
Veterans Affairs Boston
Healthcare System, Boston
University School of Medicine,
Boston, Massachusetts, USA
Correspondence to
Stacey E Alexeeff, Department
of Biostatistics, Harvard School
of Public Health, 677 Huntington
Ave, Building 2, 4th floor,
Boston, MA 02215, USA;
salexeeff@fas.harvard.edu
Accepted 4 February 2012
Published Online First
1 March 2012
422 Occup Environ Med 2012;69:422e427. doi:10.1136/oemed-2011-100268
Environment
group.bmj.com on August 21, 2012 - Published by oem.bmj.comDownloaded from
mechanisms of short-term effects of PM
2.5
on BP found that
effects were much stronger for the samples collected from
a high-traffic area.
21
A study of BP and short-term exposure to
a number of air pollutants found the strongest association with
organic carbon and its estimated fossilefuel combustion frac-
tion.
22
More research is needed to examine the relationship
between traffic-related components of PM and BP, which will
also help us understand the overall relationship between BP and
PM.
Less is known about the relationship between long-term
exposures to air pollution and BP, although mortality studies
have found strong associations with long-term air pollution
exposures.
23 24
In particular, only one recent study has investi-
gated the relationship between long-term average air pollution
exposures and BP. This study in Taiwan found a strong associ-
ation between BP and 1-year averages of PM
2.5
.
25
Since traffic
components of PM have been implicated as a key component in
relation to cardiovascular disease, research is needed to address
long-term exposure to traffic-related air pollution and BP.
We sought to address these research gaps by examining the
relationship between BP and 1-year average exposures to traffic-
related air pollution in a cohort study within the greater Boston
area. An important tool for studying within-city variation in air
pollution is the development of geographic-based exposure
models. Black carbon (BC) is a traffic-related particle and
a common surrogate for traffic particles in general, weighted
towards diesel particles. We have developed and applied a land-
use regression model for traffic particles based on BC in the
greater Boston metropolitan area.
26 27
We hypothesised that estimated 1-year average BC at partici-
pants’addresses would be associated with elevated BP. We examined
this in a longitudinal study in a closed cohort of elderly men in the
greater Boston area with repeated measurements of BP taken
roughly every 4 years.
MATERIALS AND METHODS
Study population
Our study participants were from the Veterans Administration
Normative Aging Study (NAS), a longitudinal study established
by the Veterans Administration in 1963.
28
The NAS is a closed
cohort of male volunteers from the Greater Boston area aged
21e80 years at entry, who enrolled after an initial health
screening determined that they were free of known chronic
medical conditions. Participants were re-evaluated every
3e5 years using detailed on-site physical examinations and
questionnaires. Air pollution data were collected from 1995
onward, so 1-year average BC concentrations were available
starting in 1996. This analysis restricted the study population to
subjects who were still participating in clinic visits after 1
January 1996, and subjects were followed through December
2008. Our analysis included 853 participants with complete
information regarding BC concentrations and all covariates.
These participants presented for a total of 2136 examinations
during the study period. At each study visit, systolic blood
pressure (SBP) and diastolic blood pressure (DBP) were measured
once in each arm while the subject was seated, using a standard
cuff, and the mean of right and left arm values was used in these
analyses.
BC exposure estimation
BC exposures were estimated from a spatiotemporal model that
we developed and validated previously, which has been described
in detail previously.
27
Daily concentrations at the Boston
central-site monitor were used as a predictor to reflect average
concentration levels for a given day, serving as a direct estimate
of the daily time effect. Data from 82 other stationary air
monitors were used to fit the model and estimate the effect of
each covariate in the land-use regression model. Covariates in
the BC prediction model included measures of land use for each
address (cumulative traffic density within 100 m, population
density, distance to nearest major roadway and per cent
urbanisation), geographic information system, location (latitude
and longitude), daily meteorological factors (apparent tempera-
ture, wind speed and height of the planetary boundary layer)
and temporal factors (day of week and day of season).
Separate models were fitfor the warm and cold season.
Interaction terms between the temporal meteorological predic-
tors and land-use variables allowed for spaceetime interactions.
Regression splines allowed main effect terms to non-linearly
predict exposure levels, and thin-plate splines modelled the
residual spatial variability additional spatial variability unac-
counted for by the spatial predictors. A latent variable frame-
work was used to integrate BC and EC exposure data, where BC
and EC measurements were treated as surrogates of some true
unobservable traffic exposure variable, see Gryparis et al
27
for
further details.
Our BC model showed more than a threefold range of varia-
tion in long-term average exposure across the measuring sites,
and the adjusted R
2
for this model was 0.83. A subsequent
validation sample using monthly monitoring data collected at 30
additional locations found an average correlation of 0.59
between predicted values and observed BC levels.
All addresses of participants in the NAS have been geocoded
and we used the model to generate daily predicted BC at the
address of each participant. Daily BC predictions outside of the
observed range of the monitored exposure measurements were
excluded. Long-term average exposures were created by aver-
aging the daily estimates at the participant’s residential address
or addresses for the 365 days before each clinical visit.
Statistical methods
We analysed associations between 1-year average BC exposure
and BP using linear mixed effects models with a random inter-
cept for each subject. We evaluated SBP and DBP as dependent
variables. The models took the general form:
29
Yit ¼
b
0þuiþ
b
1X1it þ þ
b
kXkit þ
b
BCBCit þ
3
it;
where Y
it
is the level of SBP or DBP in subject iat visit t;
covariates for subject iat visit tare denoted by X
1it
to X
kit
.BC
it
is
the 1-year average BC concentration for subject iduring the
365 days before visit t. Here, u
i
represents a subject-specific
intercept, reflecting unexplained heterogeneity in subjects’
overall level of outcome. We assume that the u
i
are generated
from a normal distribution with common variance, yielding the
compound symmetry variance structure. This model requires
estimation of two variance components, which represent
between- and within-subject variation. Models with unbalanced
data (ie, varying numbers of repeated measurements on each
subject) typically yield accurate estimates of within-subject
variation, provided a sufficient number of repeated measure-
ments contribute to the estimate.
To examine effect modification by a subgroup, we used an
interaction term to fit separate pollution slopes for each
subgroup and also controlled for group main effects. We tested
for interactions with diabetes status, obesity and medication
use.
Occup Environ Med 2012;69:422e427. doi:10.1136/oemed-2011-100268 423
Environment
group.bmj.com on August 21, 2012 - Published by oem.bmj.comDownloaded from
The following covariates were chosen a priori based on
established relationships with BP and air pollution and included
in the regression models regardless of statistical significance: age,
body mass index, cigarette smoking (never, current or former),
pack years of smoking, alcohol intake (<2 drinks per day vs 2+
drinks per day), fasting glucose, use of antihypertensive medi-
cations (ACE inhibitors,
b
blockers, calcium channel blockers,
angiotensin receptor blockers and diuretics), ablockers, creati-
nine clearance, weekday of clinical visit and season of clinical
visit. In addition, to control for socioeconomic status (SES) at
the individual level and the neighbourhood level, we included
years of education for each subject and neighbourhood poverty
level for each address as measured by per cent below poverty
level of each census block group in the 1999 census. We
considered additional potential confounders (short-term
apparent temperature, median income level and per cent of
population over 25 years of age without a high school diploma
as measured by the 1999 census) and included them if effect
estimates changed by more than 5% to ensure that all measured
relevant confounders were included.
RESULTS
The characteristics of the population are shown in table 1.
Subjects were older people, with a mean age of 70 at the first
study visit. Less than 10% of our participants were current
smokers, but more than half were former smokers. At the first
study visit, participants had a mean body mass index of 28 kg/
m
2
. Average SBP and DBP at this visit were 137 mm Hg and
82 mm Hg, respectively. Fifty-three per cent of participants
were initially antihypertensive medications users and that usage
increased over the follow-up period. One-year average BC
concentrations at participant addresses were 0.51 mg/m
3
, and
the IQR was 0.32 mg/m
3
. Of the 853 subjects, 605 subjects (71%)
were followed for two or more clinical study visits, with 29% of
subjects having only one visit. Subjects who participated in only
one study visit were 5 years older on average than those who
were followed for two or more study visits, but there were no
differences observed in other covariates. By the end of follow-up
in 2008, 403 of the 853 subjects (43%) were deceased, 32 subjects
(4%) dropped out and 23 subjects (3%) became too ill to
participate.
We evaluated the association of SBP and DBP with geocoded
long-term BC concentrations and expressed the results as the
change associated with a 0.32 mg/m
3
increase in BC exposure,
which was the IQR for this study (table 2). We observed
significant associations between 1-year average BC exposure and
both SBP and DBP, adjusting for confounders. In our evaluation
of additional potential confounders, we found that the other
neighbourhood SES measures and the measures of short-term
apparent temperature had a negligible effect on the estimate of
BC exposure and did not meet our criteria of at least a 5%
change (data not shown). Table 2 also reports the association for
the crude model, which includes only age and 1-year average BC
exposure as covariates for comparison. The estimated effects are
similar and estimated to be slightly higher than in the adjusted
model, also with p values <0.001.
Because short-term exposure to air pollution has been asso-
ciated with changes in BP in other studies, we wanted to ensure
that our results were not driven solely by a relationship with
short-term air pollution exposures. In figure 1, we compare the
effects of average BC exposure over a range of exposure times
from 24 h to 1 year. We see an upward trend across each time
window of average BC exposure. This provides strong evidence
that the effects of traffic particles on BP are not limited only to
short-term effects, and long-term effects warrant further scien-
tific investigation.
We examined the correlations between the short-term and
long-term exposures and found that the 1-year average predicted
BC exposure for each address was only moderately correlated
with the predicted exposure 24 h before the study visit (r¼0.53)
and the predicted exposure 1 week before study visit (r¼0.60).
When adjusting for 24 h BC exposure, we observed a slight
increase in the estimated effect of 1-yrea BC exposure on SBP,
3.22 mm Hg (95% CI 1.93 to 4.52), and on DBP, 2.63 mm Hg
(95% CI 1.92 to 3.34). Adjusting for 1-week average BC exposure
did not noticeably change the estimated effect of 1-year BC
exposure on SBP nor on DBP (data not shown). Overall, the
trend in figure 1 and our examination of the short-term
Table 1 Descriptive statistics at first study centre visit (n¼853) and
over all visits (N¼2136) given as mean (SD) or number (%)
Study variables
First visit
(n[853)
All visits
(N[2136)
Continuous variables Mean (SD) Mean (SD)
Age (years) 70.1 (7.5) 72.6 (7.4)
Body mass index (kg/m
2
) 27.9 (3.9) 28.0 (4.2)
Lifetime smoking (pack-years) 31.8 (29.4) 29.8 (27.1)
Systolic blood pressure 136 (17.9) 131 (18.4)
Diastolic blood pressure 82 (9.1) 77 (10.9)
Fasting plasma glucose (mg/dl) 109.6 (32.8) 108.3 (28.1)
Creatinine clearance (mg/dl) 1.04 (0.23) 1.09 (0.29)
Years of education (individual) 14.4 (2.7) 14.6 (2.8)
Per cent below poverty level
(census tract, 1999)
5.89 (5.19) 5.72 (4.96)
Black carbon 1-year average (mg/m
3
) 0.61 (0.29) 0.51 (0.26)
Study variables
First visit
(n[853)
All visits
(N[2136)
Categorical variables N (%) N (%)
Smoking status
Never-smoker 235 (28%) 606 (28%)
Former smoker 572 (67%) 1444 (68%)
Current smoker 46 (5%) 86 (4%)
Alcohol intake (2 + drinks per day) 178 (21%) 408 (19%)
Obese 222 (26%) 559 (26%)
Diabetes 94 (11%) 280 (13%)
Antihypertensive medication
Any 407 (48%) 1215 (57%)
Calcium channel blockers 153 (18%) 355 (17%)
ACE inhibitors 141 (17%) 515 (24%)
Angiotensin receptor agonists 14 (2%) 88 (4%)
b
blockers 204 (24%) 720 (34%)
Diuretics 116 (14%) 424 (20%)
ablockers 72 (8%) 259 (12%)
Table 2 Estimated change in blood pressure associated with a IQR
increase (0.32 mg/m
3
) in 1-year average black carbon (BC) levels for
2136 clinical visits (n¼853 subjects)
Crude modelyAdjusted modelz
Effect (95% CI) Effect (95% CI)
Systolic blood pressure (mm Hg) 4.00 (2.90 to 5.10)*** 2.64 (1.47 to 3.80)***
Diastolic blood pressure (mm Hg) 3.19 (2.58 to 3.81)*** 2.41 (1.77 to 3.05)***
***p<0.001.
yCrude regression models included only age and 1-year average BC level.
zAdjusted regression models controlled for age, cigarette smoking, pack- years of smoking,
season of clinical visit, weekday of clinical visit, body mass index, fasting glucose level, use
of antihypertensive medications (ACE inhibitors, angiotensin receptor agonists, Calcium
channel blockers,
b
blockers and diuretics), use of ablockers, years of education, per cent
below poverty level in the census tract, creatinine clearance and daily alcohol intake.
424 Occup Environ Med 2012;69:422e427. doi:10.1136/oemed-2011-100268
Environment
group.bmj.com on August 21, 2012 - Published by oem.bmj.comDownloaded from
correlations show that 1-year averages BC exposures have an
effect that is not explained by shorter-term exposures.
We also assessed the linearity of the effect of long-term BC on
BP by fitting a penalised spline for the 1-year BC exposures. For
SBP, the effect was estimated to be linear with 1 df, and for DBP,
the effect was estimated to have 4 df, although the overall trend
is still roughly linear with no evidence of a threshold (supple-
mentary figure 1).
Of the 853 subjects in the analysis, 30 subjects (3.5%) indi-
cated having a seasonal residence and 63 subjects (7.4%) indi-
cated a change in their primary address during the year prior
a clinic visit. Of the subjects who moved, the mean number of
days lived at the new address was 144 days. As a sensitivity
analysis, excluding the subjects who had a seasonal residence or
who had moved during the exposure period did not noticeably
change the estimated effect of 1-year BC exposure on SBP,
2.68 mm Hg (95% CI 1.47 to 3.88), nor on DBP, 2.43 mm Hg
(95% CI 1.77 to 3.09).
We examined interactions with diabetes, obesity and use of
antihypertensive medications and considered an interaction
significant if the p value was <0.05. These results are in table 3.
Most of these interactions were not significant; the only signifi-
cant interaction was for DBP with use of antihypertensive
medications, where those taking antihypertensive medications
were estimated to have a greater effect of BC exposure on BP. The
interaction between antihypertensive medications and SBP was
suggestive (p value¼0.085). In addition, as a sensitivity analysis,
we examined this interaction among the subgroup of participants
who already had hypertension prior to the first visit.
The results were similar but not as strong, with a p ¼0.05 for
DBP, while the interaction with SBP was not significant or
suggestive, which may reflect reduced power from looking at
only a subset of participants. Overall, these results are only
suggestive and exploratory since our ability to examine the
effect of medication usage in this observational epidemiological
study is extremely limited and cannot be considered causal.
DISCUSSION
We found that 1-year average BC concentration estimated at
each participant’s home was positively associated with SBP and
DBP in a cohort of elderly men. This association was quite
strong for an epidemiological study of its kind, with p values all
<0.001 for adjusted and crude models. Our interactions with
obesity and diabetes were not significant. The interactions with
antihypertensive medication use were suggestive, with a statis-
tically significant association for DBP only.
Comparison to the literature
The cardiovascular effects of long-term exposure to particulate
air pollution are largely unknown. A recent study in Taiwan of
1-year averages of several air pollutants and cardiovascular
outcomes found a strong association between BP and 1-year
averages of PM
2.5
.
25
The 1-year average PM
2.5
exposures in that
study were quite high, with mean 35.3 mg/m
3
(SD 15.9 mg/m
3
).
In the USA, the Environmental Protection Agency criteria level
for annual average exposure to PM
2.5
is 15 mg/m
3
, which is
substantially lower than many of the levels measured in the
Taiwan study. A rough rescaling of the effect in the Chuang
study is a 3.1 mm Hg change in DBP per 2 mg/m
3
increase in
PM
2.5
. Since BC represents a variable fraction of PM
2.5
,we
cannot directly compare effect sizes, but effects appear to be of
similar magnitudes relative to small changes in PM
2.5
. Taken
together, our results show no evidence of a thresholding effect in
the relationship between BP and BC, the traffic component of
PM
2.5
, at these lower exposure levels observed in our study. In
addition, our results suggest that traffic components of PM
2.5
may explain a large part of the association between BP and
PM
2.5
.
The literature on short-term to medium-term PM exposure
and BP is mixed.
13e20
This may be due to differences in the
composition of the particles across study sites. In particular, if
the association is really with traffic particles, differing associa-
tions between PM
2.5
and BC across locations and over time
within locations could obscure the true relation. For example,
among recent studies examining the short-term and medium-
term effects of PM
2.5
on BP, investigators found some evidence
that PM
2.5
in higher traffic areas had stronger effects on BP
compared with PM
2.5
in lower traffic areas.
16 21
More research
is needed to examine the relationship between BP and
Figure 1 Estimated change in blood
pressure associated with an IQR
increase in black carbon (BC) exposure.
DBP, Diastolic blood pressure; SBP,
systolic blood pressure.
Table 3 Modification of the effects of 1-year average black carbon
exposure on blood pressure by obesity, diabetes and antihypertensive
medication use
Outcome
Effect (95% CI) Effect (95% CI)
No diabetes (N[1856) Diabetes (N[280)
SBP mm Hg 2.63 (1.42 to 3.84) 2.83 (0.29 to 5.95)
DBP mm Hg 2.34 (1.68 to 3.00) 2.38 (0.69 to 4.08)
Outcome
Effect (95% CI) Effect (95% CI)
Not obese (N[1577) Obese (N[559)
SBP mm Hg 2.76 (1.43 to 4.09) 2.35 (0.39 to 4.30)
DBP mm Hg 2.59 (1.86 to 3.32) 1.96 (0.89 to 3.03)
Outcome
Effect (95% CI) Effect (95% CI)
Not taking
antihypertensive
medicines (N[921)
Taking antihypertensive
medicines (N[1215)
SBP mm Hg 1.78 (0.18 to 3.39) 3.32 (1.88 to 4.76)
DBP mm Hg 1.73 (0.85 to 2.60) 2.96 (2.17 to 3.75)
N is the number of observations in each analysis.
DBP, Diastolic blood pressure; SBP, systolic blood pressure.
Occup Environ Med 2012;69:422e427. doi:10.1136/oemed-2011-100268 425
Environment
group.bmj.com on August 21, 2012 - Published by oem.bmj.comDownloaded from
traffic-related components of PM to understand its role in the
overall relationship between BP and PM.
Exposure strengths and weaknesses
An important issue in the relationship between air pollution and
cardiovascular disease is how short-term and long-term effects
fit together. Our results show that 1-year average BC exposure is
associated with increases in BP and that association has a larger
effect size than short-term associations and cannot be explained
by short-term effects. In a previous study of long-term air
pollution exposure, lag time and mortality, we found the
strongest association was with the most recent year ’s exposure
in the Harvard Six City cohort.
30
Another recent mortality and
air pollution study of the Nurses Health cohort showed little
additional explanatory power for PM
2.5
exposures in the past
2 years.
23
Further research is needed to understand how expo-
sures longer than 1 year affect BP and whether there is a point of
attenuation in these effects.
Laden and coworkers
31
examined the association between
particles from different sources and mortality and reported that
the strongest effects (per micrograms per cubic metre of expo-
sure) were for traffic particles. Moreover, these particles seemed
to be particularly associated with cardiovascular deaths. This is
consistent with our finding that traffic particles (in our case BC)
are associated with increased BP. Traffic-related particles have
also been associated with elevated homocysteine concentra-
tions,
32
increased intercellular cell adhesion molecules (ICAM)
and vascular cell adhesion molecules (VCAM),
33
decreased
flow-mediated dilatation of the brachial artery,
34
reduced
parasympathetic tone
35
and with acute effects on BP.
22 36
Another related study, by Hoffmann and coworkers,
37
reported
an association between distance to roadway measures and
ankleebrachial index as well as an increased risk for peripheral
arterial disease among those living within 50 m of a major road
compared with living >200 m away in a cohort in Germany. The
study also examined long-term PM
2.5
exposure (average esti-
mated 1-year PM
2.5
concentration at the home address) and
found that it was not related to ankleebrachial index or
peripheral arterial disease in that cohort nor was it correlated
with distance to roadway in that cohort. The findings of the
Hoffmann study support the hypothesis that traffic particles are
a key component in the toxicity of PM
2.5
and may be the more
relevant exposure with respect to cardiovascular health.
Mechanisms and interactions
BC may elevate BP by increasing oxidative stress, inducing
endothelial dysfunction and promoting inflammatory activity.
These pathways may interact. For example, oxidative stress may
promote inflammatory activity, and systemic oxidative stress and
inflammation may induce endothelial dysfunction by reducing
levels of nitric oxide, a vasodilator important in maintaining
vascular tone. Other mechanisms by which BC may elevate BP
include activation of the sympathetic nervous system, alterations
in blood coagulability and direct vasoconstriction.
It is important to note that research regarding mechanisms
mediating BC’s hypertensive effect is quite limited. Plasma
markers of systemic oxidative stress and inflammation have
been associated with BC exposure in epidemiologic studies.
However, much of our discussion regarding potential mecha-
nisms of BC cardiotoxicity is necessarily based on particulate
matter research in general.
Previous studies have shown that diabetes status,
9 34
and
sometimes obesity,
938
modify the effects of particles on
cardiovascular outcomes. Our evaluation of whether the asso-
ciations we found were modified by these factors was incon-
clusive. There was no evidence to support a differential effect for
these subpopulations in this study. This could be because our
population is so older people that there is too much comorbidity
to examine these subgroups individually. Alternatively, the
mechanisms at play in the short-term effects of air pollution
may differ from the mechanisms involved in long-term effects.
Our finding that antihypertensive medications may interact
with this effect of BC is suggestive but considered exploratory.
In our sensitivity analysis restricting to those with hypertension
at baseline, the interaction was not as strong, which may be due
to lack of power from examining a smaller subgroup. With an
observational study with participants often taking multiple
medications, we cannot effectively evaluate the differential
effects of any of these medications. In addition, the inclusion of
antihypertensive medication use as a covariate is problematic
because hypertension and subsequent medication use is
a potential consequence of exposure and an important predictor
of BP, so there is a potential for bias by including or not
including this covariate.
Another important consideration of the findings is the clinical
relevance of an increase of 2 mm Hg in DBP or SBP. While levels
of 2 mm Hg may seem relatively small, some regions have
higher levels of combustion particles and may see greater effects.
Although the IQR of 1-year average BC was only 0.32 mg/m
3
in
our cohort, the overall distribution was skewed and ranged from
0.02 to 1.90 mg/m
3
. An extreme example would be the levels
observed in Taiwan in the Chuang 2011 study, where the range
of annual PM
2.5
exposures was 8.8e82.3 mg/m
3
. A meta-analysis
of BP to cause-specific mortality found that a 10 mm Hg
decrease in DBP was associated with a notable decreased risk of
ischaemic heart disease of 0.62 (95% CI 0.60 to 0.64) for those
ages 70e79. Thus, the clinical relevance of a 2 mm Hg increase
in DBP per 0.32 mg/m
3
increase in 1-year average BC depends on
the regional exposure levels and the relationship between BP and
cardiovascular disease.
Limitations and generalisability
Our estimates of long-term exposure are model based, rather
than based on measurements. Our model is relatively rich
and was validated on a large number of sampling sites, but
prediction error is always a concern as measurement error in
spatial models can sometimes lead to bias and/or CIs that are
too narrow.
39
However, we note that this is a limitation in any
air pollution study using stationary monitors or modelled
exposures because the true exposure of interest is not measured
exactly.
Although we did control for SES at the individual level and at
the neighbourhood level, it is possible that the number of years
of education does not provide strong enough control of indi-
vidual-level SES; hence, there may still be some residual
confounding by SES. In addition, there is potential for
confounding by other exposures such as environmental tobacco
smoke, as we only measured personal smoking and not whether
there was another smoker in the home. Another possible
unmeasured confounder is road noise, which also comes from
heavy traffic and has been associated with cardiovascular
outcomes. Road noise is particularly difficult to address since
traffic pollution and road noise come from the same source.
Overall, we have included most key confounders in our model so
we do not expect that unmeasured confounding would have
a strong influence on our results.
Our study population was homogeneous, consisting entirely
of elderly men, 97% of whom were Caucasian. Thus, these
426 Occup Environ Med 2012;69:422e427. doi:10.1136/oemed-2011-100268
Environment
group.bmj.com on August 21, 2012 - Published by oem.bmj.comDownloaded from
results cannot be generalised to other populations without
further research on how effects vary by age, gender and race.
Additionally, the study population is a self-selected group of
people who continued to participate in an ongoing study for
many years and may not be representative of all elderly men in
the USA. In particular, there may be survivor bias if the subjects
who continue to participate are healthier than other older
people, which would bias effect estimates towards the null, so
the true effect in the general population may be stronger.
Conclusions
Long-term exposure to BC is associated with increases in BP in
this older population, a finding that could explain part of the
association of particulate air pollution with cardiovascular
mortality. More research is needed to address the relation
between traffic-related air pollution exposures and BP among
diverse study populations, including women, other races and
younger populations. Further research is also needed to study
the role diabetes, obesity and antihypertensive medication use in
modifying the effect and to clarify other mechanisms underlying
the association between BC and BP.
Funding This work was supported by the National Institute of Environmental Health
Sciences grants ES015172-01 and ES00002 and the US Environmental Protection
Agency grants RD-832416-01 and RD-834798-01. The Normative Ageing Study is
supported by the Cooperative Studies Program/Epidemiology Research and
Information Center of the US Department of Veterans Affairs and is a component of
the Massachusetts Veterans Epidemiology Research and Information Center
(MAVERIC), Boston, MA.
Competing interests None.
Ethics approval Human Subjects Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
REFERENCES
1. Laden F, Schwartz J, Speizer FE, et al. Reduction in fine particulate air pollution and
mortality: extended follow-up of the Harvard six cities study. Am J Respir Crit Care
Med 2006;173:667e72.
2. Tonne C, Melly S, Mittleman M, et al. A case-control analysis of exposure
to traffic and acute myocardial infarction. Environ Health Perspect
2007;115:53e7.
3. Zanobetti A, Schwartz J. Particulate air pollution, progression, and survival after
myocardial infarction. Environ Health Perspect 2007;115:769e75.
4. Wellenius GA, Schwartz J, Mittleman MA. Particulate air pollution and hospital
admissions for congestive heart failure in seven United States cities. Am J Cardiol
2006;97:404e8.
5. Zeka A, Zanobetti A, Schwartz J. Short term effects of particulate matter on cause
specific mortality: effects of lags and modification by city characteristics. Occup
Environ Med 2005;62:718e25.
6. Katsouyanni K, Touloumi G, Spix C, et al. Short-term effects of ambient sulphur
dioxide and particulate matter on mortality in 12 European cities: results from time
series data from the APHEA project. Air pollution and health: a European approach.
BMJ 1997;314:1658e63.
7. Ying Z, Kampfrath T, Thurston G, et al. Ambient particulates alter vascular function
through induction of reactive oxygen and nitrogen species. Toxicol Sci
2009;111:80e8.
8. Sørensen M, Daneshvar B, Hansen M, et al. Personal PM2.5 exposure and markers
of oxidative stress in blood. Environ Health Perspect 2003;111:161e6.
9. Dubowsky SD, Suh H, Schwartz J, et al. Diabetes, obesity, and hypertension may
enhance associations between air pollution and markers of systemic inflammation.
Environ Health Perspect 2006;114:992e8.
10. O’Neill MS, Veves A, Sarnat JA, et al. Air pollution and inflammation in type 2
diabetes: a mechanism for susceptibility. Occup Environ Med 2007;64:373e9.
11. Nemmar A, Hoet PH, Dinsdale D, et al. Diesel exhaust particles in lung acutely
enhance experimental peripheral thrombosis. Circulation 2003;107:1202e8.
12. Urch B, Silverman F, Corey P, et al. Acute blood pressure responses in healthy
adults during controlled air pollution exposures. Environ Health Perspect
2005;113:1052e5.
13. Harrabi I, Rondeau V, Dartigues JF, et al. Effects of particulate air pollution on
systolic blood pressure: a population-based approach. Environ Res 2006;101:89e93.
14. Ebelt ST, Wilson WE, Brauer M. Exposure to ambient and nonambient components of
particulate matter: a comparison of health effects. Epidemiology 2005;16:396e405.
15. Ibald-Mulli A, Timonen KL, Peters A, et al. Effects of particulate air pollution on
blood pressure and heart rate in subjects with cardiovascular disease: a multicenter
approach. Environ Health Perspect 2004;112:369e77.
16. Auchincloss AH, Diez Roux AV, Dvonch JT, et al. Associations between recent
exposure to ambient fine particulate matter and blood pressure in the multi-ethnic
study of atherosclerosis (MESA). Environ Health Perspect 2008;116:486e91.
17. Ibald-Mulli A, Stieber J, Wichmann HE, et al. Effects of air pollution on blood
pressure: a population-based approach. Am J Publ Health 2001;91:571e7.
18. Choi JH, Xu QS, Park SY, et al. Seasonal variation of effect of air pollution on blood
pressure. J Epidemiol Community Health 2007;61:314e18.
19. Liu L, Ruddy T, Dalipaj M, et al. Effects of indoor, outdoor, and personal exposure to
particulate air pollution on cardiovascular physiology and systemic mediators in
seniors. J Occup Environ Med 2009;51:1088e98.
20. Dvonch JT, Kannan S, Schulz AJ, et al. Acute effects of ambient particulate matter
on blood pressure: differential effects across urban communities. Hypertension
2009;53:853e9.
21. Brook RD, Urch B, Dvonch JT, et al. Insights into the mechanisms and mediators of
the effects of air pollution exposure on blood pressure and vascular function in
healthy humans. Hypertension 2009;54:659e67.
22. Delfino RJ, Tjoa T, Gillen DL, et al. Traffic-related air pollution and blood pressure in
elderly subjects with coronary artery disease. Epidemiology 2010;21:396e404.
23. Puett RC, Hart JE, Yanosky JD, et al. Chronic fine and coarse particulate exposure,
mortality, and coronary heart disease in the nurses’ health study. Environ Health
Perspect 2009;117:1697e701.
24. Gan WQ, Koehoorn M, Davies HW, et al. Long-term exposure to traffic-related air
pollution and the risk of coronary heart disease hospitalization and mortality. Environ
Health Perspect 2011;119:501e7.
25. Chuang KJ, Yan YH, Chiu SY, et al. Long-term air pollution exposure and risk factors
for cardiovascular diseases among the elderly in Taiwan. Occup Environ Med
2011;68:64e8.
26. Maynard D, Coull BA, Gryparis A, et al. Mortality risk associated with short-term
exposure to traffic particles and sulfates. Environ Health Perspect 2007;115:751e5.
27. Gryparis A, Coull B, Schwartz J, et al. Semiparametric latent variable regression
models for spatiotemporal modelling of mobile source particles in the greater Boston
area. J R Stat Soc Ser C 2007;56:183e209.
28. Bell B, Rose C, Damon A. The normative aging study: an interdisciplinary and
longitudinal study of health and aging. Aging Hum Dev 1972;3:4e17.
29. Fitzmaurice GM, Laird NM, Ware JH. Applied Longitudinal Analysis. New York:
John Wiley and Sons. 2004.
30. Schwartz J, Coull B, Laden F, et al. The effect of dose and timing of dose on the
association between airborne particles and survival. Environ Health Perspect
2008;116:64e9.
31. Laden F, Neas LM, Dockery DW, et al. Association of fine particulate matter from
different sources with daily mortality in six U.S. cities. Environ Health Perspect
2000;108:941e7.
32. Park SK, O’Neill MS, Vokonas PS, et al. Traffic-related particles are associated with
elevated homocysteine: the VA normative aging study. Am J Respir Crit Care Med
2008;178:283e9.
33. Alexeeff SE, Coull BA, Gryparis A, et al. Medium-term exposure to traffic-related air
pollution and markers of inflammation and endothelial function. Environ Health
Perspect 2011;119:481e6.
34. O’Neill MS, Veves A, Zanobetti A, et al. Diabetes enhances vulnerability to
particulate air pollution-associated impairment in vascular reactivity and endothelial
function. Circulation 2005;111:2913e20.
35. Schneider A, Hampel R, Ibald-Mulli A, et al. Changes in deceleration capacity of
heart rate and heart rate variability induced by ambient air pollution in individuals with
coronary artery disease. Part Fibre Toxicol 2010;7:29.
36. Mordukhovich I, Wilker E, Suh H, et al. Black carbon exposure, oxidative stress
genes, and blood pressure in a repeated-measures study. Environ Health Perspect
2009;117:1767e72.
37. Hoffmann B, Moebus S, Kroger K, et al. Residential exposure to urban air pollution,
ankle-brachial index, and peripheral arterial disease. Epidemiolology 2009;20:280e8.
38. Madrigano J, Baccarelli A, Wright R, et al. Air pollution, obesity, genes, and cellular
adhesion molecules. Occup Environ Med 2010;67:312e17.
39. Gryparis A, Paciorek CJ, Zeka A, et al. Measurement error caused by spatial
misalignment in environmental epidemiology. Biostatistics 2009;10:258e74.
PAGE fraction trail=6
Occup Environ Med 2012;69:422e427. doi:10.1136/oemed-2011-100268 427
Environment
group.bmj.com on August 21, 2012 - Published by oem.bmj.comDownloaded from
doi: 10.1136/oemed-2011-100268
1, 2012 2012 69: 422-427 originally published online MarchOccup Environ Med
Joel Schwartz, Stacey E Alexeeff, Irina Mordukhovich, et al.
Study
Veterans Administration Normative Aging
traffic particles and blood pressure in the
Association between long-term exposure to
http://oem.bmj.com/content/69/6/422.full.html
Updated information and services can be found at:
These include:
Data Supplement http://oem.bmj.com/content/suppl/2012/02/29/oemed-2011-100268.DC1.html
"Supplementary Data"
References http://oem.bmj.com/content/69/6/422.full.html#ref-list-1
This article cites 38 articles, 14 of which can be accessed free at:
service
Email alerting the box at the top right corner of the online article.
Receive free email alerts when new articles cite this article. Sign up in
Collections
Topic
(538 articles)Other exposures (131 articles)Air pollution, air quality
Articles on similar topics can be found in the following collections
Notes
http://group.bmj.com/group/rights-licensing/permissions
To request permissions go to:
http://journals.bmj.com/cgi/reprintform
To order reprints go to:
http://group.bmj.com/subscribe/
To subscribe to BMJ go to:
group.bmj.com on August 21, 2012 - Published by oem.bmj.comDownloaded from
