Available online at www.sciencedirect.

com

Regulatory Toxicology and Pharmacology 51 (2008) 135–140

www.elsevier.com/locate/yrtph

Commentary

Do pollution time-series studies contain uncontrolled or

residual confounding by risk factors for acute health events?
John Bukowski *

WordsWorld Consulting, 658 Evans Lane, Dayton, OH 45459, USA

Received 19 September 2007

Available online 15 December 2007

Abstract

Acute health effects from air pollution are based largely on weak associations identified in time-series studies comparing daily air pol-
lution levels to daily mortality. Much of this mortality is due to cardiovascular disease. Time-series studies have many potential limita-
tions, but are not thought to be confounded by traditional cardiovascular risk factors (e.g., smoking status or hypertension) because
these chronic risk factors are not obviously associated with daily pollution levels. However, acute psychobehavioral variants of these
risk factors (e.g., smoking patterns and episodes of stress on any given day) are plausible confounders for the associations observed
in time-series studies, given that time-series studies attempt to predict acute rather than chronic health outcomes. There is a fairly com-
pelling literature on the strong link between cardiovascular events and daily ‘‘triggers’’ such as stress. Stress-related triggers are plausibly
associated with daily pollution levels through surrogate stressors such as ambient temperature, daily workload, local traffic congestion,
or other correlates of air pollution. For example, variables such as traffic congestion and industrial activity increase both stress-related
health events and air pollution, suggesting the potential for classical confounding. Support for this argument is illustrated through exam-
ples of the well-demonstrated relationship between emotional stress and heart attack/stroke.
Ó 2007 Elsevier Inc. All rights reserved.

Keywords: Time-series; Air pollution; Confounding; Trigger; Cardiovascular

1. Background The recognized or suspected limitations of TS studies

fall into four major categories:
Over the past 10–15 years, numerous articles have dealt
with associations between short-term health effects (especially
sudden death) and relatively low levels of air pollution (Dom-
inici et al., 2002, 2003a,b, 2005; Bell et al., 2004; Jerrett et al.,
2005; Analitis et al., 2006; Ostro et al., 2006). These studies
have generally used the time-series (TS) approach in which
daily fluctuations in pollution are modeled against daily num-
bers of health events within a city or group of cities. The pri-
mary endpoint in these studies is often death, because
accurate mortality statistics are maintained for all large cities,
but other acute endpoints (e.g., emergency department visits,
hospitalization, etc.) have also been explored.
*
Fax: +1 937 439 0221.
E-mail address: wwc-johnb@woh.rr.com
(1) Ecological bias: TS studies are ecological in nature.
As Morgenstern (1998) points out, ‘‘ecological bias
may be severe in practice’’ and ‘‘adjustment for extra-
neous risk factors may not reduce the ecologic bias’’
and may even ‘‘increase bias’’. A recent methodolog-
ical analysis has stratified the ecological associations
with particulate matter (PM) into national and local
components, concluding that there appears to be no
real association between PM and mortality at the
local level (Janes et al., 2007).
(2) Residual confounding by area-wide weather variables:
The weather variables included in TS models are
themselves ecological surrogates for the heat stress
experienced by individuals. For example, ambient
temperature is often measured at a single location

0273-2300/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.yrtph.2007.12.003
136 J. Bukowski / Regulatory Toxicology and Pharmacology 51 (2008) 135–140
(e.g., an airport) and does not reflect the heat stress to
which all municipal residents are exposed. Area-wide
variables provide no information on individual-level
factors, such as access to air conditioning and resi-
dence near ‘‘heat islands’’. Heat-related variables
have been shown to have a strong and nonlinear
effect on health (Armstrong, 2006), which may not
be adequately accounted for in linear TS models.
(3) Statistical errors: TS studies use highly complex sta-
tistical models to tease out weak associations. Rela-
tively minor errors in statistical application or
model choice/specification can lead to substantial
errors in model coefficients (Moolgavkar, 2005).
For example, minor problems were identified several
years ago with the software application and modeling
specifications common to early TS studies investigat-
ing the mortality associated with PM air pollution.
This so-called ‘‘GAM/S-Plus’’ problem resulted in
an artificial doubling of risk (Dominici et al., 2005).
Subsequent analyses have corrected this particular
problem, but do not eliminate the potential for other
uncontrolled sources of bias in such complex and
model-intensive approaches (Moolgavkar, 2005).
(4) Weak associations: In past epidemiologic studies,
weak associations have generally been defined as
relative risks (RR) below 1.5–3.0 (Wynder, 1987;
Shapiro, 1994, 2000; Szklo, 2001). It has been
argued that these weak associations might provide
insight into important risks, but that they are not
sufficiently reliable to form the basis for expo-
sure-response functions used for regulatory assess-
ment (Hertz-Picciotto, 1995; Shapiro, 1994, 2000).
Time-series studies generally report RR below
1.02, although values vary by city, region, time
of year, etc. The RR for the 95-city National Mor-
bidity, Mortality, and Air Pollution Study (NMM-
APS), which is the most comprehensive and
authoritative TS study to date, reported RR for
total mortality in the range of 1.002 per 10 micro-
grams of fine particulate matter (PM2.5) (Dominici
et al., 2005). Spurious elevations of this magnitude
would not be unusual, especially given the limita-
tions inherent in the TS design.
There may be additional areas of uncontrolled/residual
confounding that have been overlooked in TS studies.
The current article explores the hypothesis that TS studies
may not adequately control for confounding risk factors
(including smoking, stress, and other variables) that can
trigger acute cardiovascular health events.
2. Chronic vs. acute risk factors
Many of the daily health events (e.g., death, hospital
admissions, etc.) of importance in TS studies are from
cardiovascular diseases (CVD) such as heart disease and
stroke, which make up approximately one-third of total
US mortality and a larger percentage of external/nontrau-
matic deaths (Miniño et al., 2006). Traditional risk factors
for CVD include variables such as smoking status, high
blood pressure, diets high in fat and salt, etc. (Wilson
and Cullerton, 1998; Vinereanu, 2006). These are chronic
risk factors that do not change within individuals from
day-to-day and are not obviously associated with short-
term pollution exposures. Such risk factors are not con-
sidered to be potential confounders in TS studies because
‘‘smoking and dietary habits [and similar variables] are
unlikely to covary with air pollution on timescales briefer
than season.’’ (Schwartz et al., 2003).
Risk factors such as smoking status predict chronic
health outcomes (e.g., cancer or heart disease), not the
acute health events (e.g., sudden death or hospitalization)
of concern in TS studies. Put another way, chronic risk fac-
tors predict someone’s long-term probability of dying from
a heart attack, but not when that fatal heart attack will
occur. However, there are acute variants (i.e., ‘‘triggers’’)
for these risk factors that need to be considered when pre-
dicting near-term (e.g., daily) risk. For example, an active
cigarette smoker is at overall increased risk for a heart
attack, but the risk on any given day is driven in part by
the number and timing of cigarettes smoked on that partic-
ular day. Research has shown that coronary blood clots are
larger in heart attacks occurring near a smoking episode
(Giri et al., 2001) and that ischemia is much more likely
to occur during smoking (Gabbay et al., 1996), which
means that the length and number of smoking episodes
on any given day is important. It has also been shown that
smoking episodes increase cortisol/catecholamine levels,
heart rate, blood pressure, coronary vasoconstriction, oxy-
gen demand, and other hemodynamic processes that can
trigger heart attack and stroke (Strike and Steptoe, 2005;
Robinson and Cinciripini, 2006).
On a similar vein, high blood pressure, aggressive per-
sonality, stressful occupation, and similar factors increase
overall risk of heart attack or stroke. However, the risk
on any given day would be more strongly related to the
number and magnitude of stressful events occurring on
or near that day. For example, it has been clearly and con-
sistently shown that episodes of emotional stress (e.g.,
anger, stress, or anxiety) can trigger myocardial infarction
(MI) and stroke (Mittleman et al., 1995; Möller et al., 1999;
Koton et al., 2004; Strike and Steptoe, 2005; Strike et al.,
2006a).
3. Confounding from acute triggers
In time-series studies, chronic risk factors are correctly
ignored as irrelevant to the acute situation. However,
time-series studies also ignore the acute triggers that have
direct bearing on acute health events. Various factors have
been implicated as possible acute triggers of CVD events.
These include physical exertion, sexual activity, sleep dis-
turbance, tobacco smoking, alcohol consumption, high-
fat meals, and emotional distress. In clinical epidemiology,
 J. Bukowski / Regulatory Toxicology and Pharmacology 51 (2008) 135–140
these factors have generally been measured via retrospec-
tive recall or prospective diaries, usually focusing on the
1–2 h or 1–2 days before symptom onset (Strike and Step-
toe, 2005).
People are known to increase risky behaviors (e.g.,
anger, smoking, over-eating, lack of sleep, etc.) because
of stress (Cummings et al., 1985; Niaura et al., 2002;
McKinzie et al., 2006; Falkin et al., 2007; Yamamoto
et al., 2007). It would be expected that these behaviors
would similarly increase from the irritability and stress
associated with high ambient temperature, heavy traffic,
traffic/work noise, job strain, or other factors associated
with periods of increased pollution. For example, work
stress has been associated with increases in eating, smok-
ing, and other acute triggers (Siegrist and Rodel, 2006;
Radi et al., 2007; Kouvonen et al., 2007). The strenuous
physical activity associated with heightened industrial
activity could similarly be associated with both pollution
and acute risk. Such associations between risk and corre-
lates of exposure are suggestive of classical confounding.
3.1. Case example: emotional stress and heart attack
Considerable research has linked negative emotional
states such as anger, frustration, and anxiety with either
acute coronary syndrome (ACS) (Myers and Dewar,
1975; Mittleman et al., 1995; Gabbay et al., 1996; Möller
et al., 1999, 2005; Servoss et al., 2002; Gullette et al.,
1997; Strike and Steptoe, 2005; Strike et al., 2006a) or
the pathophysiological processes attending ACS. Mental/
emotional stress decreases vagal tone (Bacon et al., 2004)
and increases vasoconstriction, blood pressure, release of
catecholamines and cortisol, and platelet aggregation (Ver-
rier et al., 1987; Verrier and Mittleman, 1996; Jern et al.,
1989; Muller et al., 1989; Boltwood et al., 1993; Toffler,
1997; Strike et al., 2006b). These forces are thought to
increase cardiac arrhythmia and plaque disruption, subse-
quently leading to MI and/or unstable angina (Muller
et al., 1989; Toffler, 1997; Lampert et al., 2002; Servoss
et al., 2002; Burg et al., 2004). In most cases, the RR from
stress are substantial, ranging from approximately 2.0–15.0
(Table 1).
Many of the studies linking stress and heart disease have
used case-crossover designs that control between-subject
differences and most subsequent confounding (Mittleman
et al., 1995; Gullette et al., 1997; Möller et al., 1999,
2005; Lampert et al., 2002; Burg et al., 2004; Strike and
Steptoe, 2005). Recent research has coupled prospective
daily diaries with ambulatory electronic monitoring of sub-
clinical cardiac events (e.g., arrhythmia or transient ische-
mia), in order to generate real-time data that avoid
biased recall and provide a clear temporal relationship
between episodes of stress and documented cardiovascular
events (Gabbay et al., 1996; Gullette et al., 1997; Lampert
et al., 2002; Bacon et al., 2004; Burg et al., 2004).
Anger and frustration are common in daily life. Results
from the above studies suggest approximately 1–17% expo-
137
sure prevalence among cases for clinically important time
windows (Table 1). Furthermore, the substantial excess
risks associated with these stressors (Table 1) dwarf the tiny
risks identified in TS pollution studies. For example, the
1.002 RR per 10 micrograms PM2.5 that was reported for
NMMAPS (Dominici et al., 2005) implies less than 1%
excess mortality over the typical range of PM2.5 concentra-
tions. By comparison, stress-related triggers increase acute
MI risk by approximately 100–1400% (i.e., 2- to 15-fold)
(Table 1).
3.2. Confounding by stress-related triggers
Confounding by acute triggers would be expected if such
triggers were associated with both acute cardiovascular
events and pollution (or variables correlated with pollu-
tion, such as temperature, traffic density, or increased
industrial activity). Two acute triggers that fit this pattern
well are traffic density and workplace stress.
3.3. Traffic stress and CVD
Driving in or living near heavy traffic increases the stress
that can cause both CVD and acute cardiovascular events
(Netterstrom and Juel, 1988; Robinson, 1989; Babisch
et al., 1993; Gee and Takeuchi, 2004). A cohort study of
over 2000 Danish bus drivers identified ‘‘driving in heavy
traffic’’ and ‘‘increased work pace’’ as significant risks for
MI, with RR ranging from approximately 3.0–6.0 (Netter-
strom and Juel, 1988). Vehicular travel during vacation has
also been linked to MI, with hospitalized cases being signif-
icantly more likely to travel by car (RR 3.2, 95% CI 1.1–
8.8) compared to hospital controls (Kop et al., 2003). Such
findings are consistent with telephone interviews (adminis-
tered while subjects were driving) that suggest that stress
increases with traffic congestion (Wickens and Wiesenthal,
2005). Research has also found that stress-induced blood
parameters consistent with the initiation of ACS increase
during car travel (Robinson, 1989). Living in the proximity
of heavy traffic or living with chronic noise exposure has
been similarly associated with stress and heart disease
(Ising et al., 1999; Gee and Takeuchi, 2004; Babisch
et al., 2005). These associations show the impact of traffic
stress and are consistent with the large amount of evidence
that negative emotional states are acutely associated with
cardiovascular pathophysiology.
3.4. Industrial activity and CVD
Sudden increases in industrial activity cause short-term
health measures to decline because of job-related stress
and the ‘‘growing pains’’ associated with adapting to new
technologies (Netterstrom and Juel, 1988; Brenner, 2005;
Tapia Granados, 2005). More importantly, systematic
reviews have identified a firm association between work-
related stress and CVD (Strike and Steptoe, 2005). The
incidence of MI has been shown to peak on Monday, pri-
138 J. Bukowski / Regulatory Toxicology and Pharmacology 51 (2008) 135–140

Table 1
Studies showing a significant association between mental stress and acute heart disease
Reference Design (n) Pea (%) Stress exposure RR (95% CI)b
Willich et al. Case-control (1194) 4 Emotional upset within 24 h of myocardial 2.7 (1.1–6.6)
(1993) infarction (MI)
Mittleman et al. Case-crossover (1623) 2 Anger within 2 h of MI 2.3 (1.7–3.2)
(1995)
Gabbay et al. Similar to case-crossover that used daily diaries 14 Anger preceding ischemia 2 (p < 0.057)d
(1996) and ambulatory EKGc (63)
Gullette et al. Case-crossover using daily diaries and ambulatory 4 Tension within 1 h of ischemia 2.2 (1.1–4.5)
(1997) EKG (58)
4 Frustration within 1 h of ischemia 2.2 (1.1–4.3)
Möller et al. Case-crossover
(1999) (690 with symptomse) 1 Anger within 1 h of MI 9.0 (4.4–18.2)
(399 w/o symptoms) 2 Anger within 1 h of MI 15.7 (7.6–32.4)
Burg et al. (2004) Case-crossover using diaries and ICDf (42) 15 Anger within 15 min preceding ICD shock 5.3 (1.4–22.8)
Möller et al. Case-crossover (1381) 1 High-pressure work deadline within 24 h 6.0 (1.8–20.4)
(2005) of MI
1 Competitive work pressure within 2–7 6.0 (1.3–26.8)
days of MI
Strike et al. Case-crossover (295) 17 Anger within 2 h of acute coronary
(2006a) syndrome (ACS)
Compared to 24 h earlier 2.1 (1.7–2.2)
Compared to 6 mos. earlier 7.3 (5.2–10.2)
a
Prevalence of stress exposure among cases.
b
Relative risk (95% confidence interval).
c
Study similar to case-crossover, but not formally presented as such. Ambulatory patients filled out detailed diaries while fitted with ambulatory EKG
devices.
d
Authors presented findings graphically without formal RR (p-value within parentheses).
e
Separate analyses for those with or without telltale premonitory symptoms.
f
Ambulatory patients filled out detailed diaries while fitted with implanted cardioverter-defibrillators (ICD).

marily among working populations, suggesting a short-
term impact from stresses associated with the workplace
(Willich et al., 1993). It is uncertain whether this is due
to increasing mental stress or to increased physical exertion
after 2 days of rest. Möller et al. (2005) investigated mental
stress specifically within the Stockholm Heart Epidemiol-
ogy Program. They enrolled more than 1300 patients with
acute, nonfatal first-infarcts into a case-crossover study of
potential triggers. A traditional case-control analysis of
this population showed a significant doubling of risk
among men exposed to conflict at work within the past
12 months. The case-crossover analysis showed that the
RR increased to 6.0 (95% confidence interval (CI) 1.8–
20.3) within 24 h of a high-pressure work-related deadline.
A similar risk (RR 6.0, 95% CI 1.3–26.8) was reported
within 48 h of an event associated with heavy workplace
competition (Möller et al., 2005).
Time-series models are usually adjusted for day of the
week, but this would not remove all bias related to work
stress. More CVD events occur on Monday than on other
days (Willich et al., 1993), but there is no perfect correla-
tion between work stress and day of the week. Also, as
can be seen in the table, most pertinent work stress occurs
within 1–2 h before disease onset. Day of the week would
appear to be too imprecise a metric to capture this. Fur-
thermore, as already noted, the ecological nature of the
TS design suggests only imperfect control for individual-
level confounding, so that residual bias is to be expected.
3.5. Associations with air pollution
The above risk factors have plausible associations with
air pollution, suggesting a real possibility of confounding.
Traffic activity increases the level of air pollution, with
locations near heavy traffic having significantly higher par-
ticulate levels than those near lighter traffic (Brauer et al.,
2003; Levy et al., 2003; Rundell et al., 2006; Soliman
et al., 2006; van Roosbroeck et al., 2006). In fact, air-qual-
ity models based on traffic patterns account for up to 50–
73% of the variability in average annual levels of fine PM
(Brauer et al., 2003; Soliman et al., 2006). Increased eco-
nomic or industrial activity should similarly increase pollu-
tion levels, given that air pollution levels are much higher
in areas with active factories (Ghio, 2004; Sanderson and
Farant, 2004).
4. Conclusions
TS studies suggest very weak associations between air
pollution and acute mortality. Therefore, we need to inves-
tigate all sources of potential bias, because very little
uncontrolled or residual bias would be needed to produce
such weak excess risks. Stress has been consistently shown
to be an acute trigger of cardiovascular events, and is also
plausibly associated with air pollution. Stress can increase
smoking, raise blood pressure, disrupt sleep, and produce
other behavioral or physiological effects that increase the
 J. Bukowski / Regulatory Toxicology and Pharmacology 51 (2008) 135–140
acute risk of heart attack or stroke. These triggers exert
their influence over a period of hours to days before a
CVD event, which is consistent with the time lags used in
TS studies. The RR associated with stress-related triggers
dwarf the very weak associations reported in pollution
TS studies.
Janes et al. (2007) report little association between PM
and mortality at the local level, and suggest that summary
associations at the national level may be confounded.
Using stress-related risk as an example, I have shown that
stresses from things such as temperature, traffic density,
and occupational workload are plausibly associated with
both acute health and air pollution exposure, so that clas-
sical confounding is a real possibility. My arguments have
been directed primarily at mortality studies, but would
apply at least as well to studies evaluating hospitalization
or emergency department visits. These latter endpoints
are earlier in the cause and effect chain and would therefore
be more directly influenced by stress. In fact, the literature
on stress and CHD deals more with heart attack and stroke
than with death.
I have necessarily presented only a limited number of
examples of potential confounding from acute triggers.
However, there are likely to be other unrecognized
variables that could plausibly confound the complex asso-
ciation between air pollution and acute health. This poten-
tial for bias needs to be thoroughly investigated (given the
weak, ecological associations reported), and caution is
warranted when applying causal associations to the results
of TS studies.
Acknowledgment
This paper was funded in part by the American Petro-
leum Institute.
References
Analitis, A., Katsouyanni, K., Dimakopoulou, K., Samoli, E., Nikolou-
lopoulos, AK., Petasakis, Y., et al., 2006. Short-term effects of
ambient particles on cardiovascular and respiratory mortality. Epide-
miology 17, 230–233.
Armstrong, B., 2006. Models from the relationship between ambient
temperature and daily mortality. Epidemiology 17, 624–631.
Babisch, W., Ising, H., Gallacher, J.E., Sharp, D.S., Baker, I.A., 1993.
Traffic noise and cardiovascular risk: the Speedwell study, first phase
Outdoor noise levels and risk factors. Arch. Environ. Health 48,
401–405.
Babisch, W., Beule, B., Schust, M., Kersten, N., Ising, H., 2005. Traffic
noise and risk of myocardial infarction. Epidemiology 16, 33–40.
Bacon, S.L., Watkins, L.L., Babyak, M., Sherwood, A., Hayano, J.,
Hinderliter, A.L., et al., 2004. Effects of daily stress on autonomic
cardiac control in patients with coronary artery disease. Am. J.
Cardiol. 93, 1292–1294.
Bell, M.I., McDermott, A., Zeger, S.L., Samet, J.M., Dominici, F., 2004.
Ozone and short-term mortality in 95 US urban communities, 1987–
2000. JAMA 292, 2372–2378.
Boltwood, M.D., Taylor, C.B., Burke, M.B., Grogin, H., Giacomini, J.,
1993. Anger report predicts coronary artery vasomotor response to
mental stress in atherosclerotic segments. Am. J. Cardiol. 72,
1361–1365.
139
Brauer, M., Hoek, G., van Vliet, P., Meliefste, K., Fischer, P., Gehring,
U., et al., 2003. Estimating long-term average particulate air pollution
concentrations: application of traffic indicators and geographic infor-
mation systems. Epidemiology 14, 228–239.
Brenner, M.H., 2005. Commentary: economic growth is the basis of
mortality rate decline the 20th century—experience of the United
States 1901–2000. Int. J. Epidemiol. 34, 1214–1221.
Burg, M.M., Lampert, R., Joska, T., Batsford, W., Jain, D., 2004.
Psychological traits and emotion-triggering of ICD shock-terminated
arrhythmias. Psychosom. Med. 66, 898–902.
Cummings, K.M., Jaen, C.R., Giovino, G., 1985. Circumstances sur-
rounding relapse in a group of recent exsmokers. Prev. Med. 14, 195–
202.
Dominici, F., McDermott, A., Zeger, S.L., Samet, J.M., 2002. On the use
of generalized additive models in time-series studies of air pollution
and health. Am. J. Epidemiol. 156, 193–203.
Dominici, F., McDermott, A., Zeger, S.L., Samet, J.M., 2003a. Airborne
particulate matter and mortality: time-scale effects in four US cities.
Am. J. Epidemiol. 157, 1055–1065.
Dominici, F., McDermott, A., Zeger, S.L., Samet, J.M., 2003b. National
maps of the effects of particulate matter on mortality: exploring
geographical variation. Environ. Health Perspect. 111, 39–43.
Dominici, F., McDermott, A., Daniels, M., Zeger, S.L., Samet, J.M.,
2005. Revised analyses of the National Morbidity, Mortality, and Air
Pollution study: mortality among residents of 90 cities. J. Toxicol.
Environ. Health A 68, 1071–1092.
Falkin, G.P., Fryer, C.S., Mahadeo, M., 2007. Smoking cessation and
stress among teenagers. Qual. Health Res. 17, 812–823.
Gabbay, F.H., Krantz, D.S., Kop, W.J., Hedges, S.M., Klein, J.,
Gottdiener, J.S., et al., 1996. Triggers of myocardial ischemia during
daily life in patients with coronary artery disease: physical and mental
activities, anger and smoking. J. Am. Coll. Cardiol. 27,
585–592.
Gee, G.C., Takeuchi, D.T., 2004. Traffic stress, vehicular burden and well
being: a multilevel analysis. Soc. Sci. Med. 59, 405–414.
Ghio, A.J., 2004. Biological effects of Utah Valley ambient air particles in
humans: a review. J. Aerosol. Med. 17, 157–164.
Giri, S., Mitchel, J., Kiernan, F.J., Hirst, J.A., Clive, J., Fram, D.B.,
et al., 2001. Cigarette smoking increases the magnitude of intracor-
onary thrombus observed during acute myocardial infarction
[abstract]. Circulation 103, 1362-e.
Gullette, E.C.D., Blumenthal, J.A., Babyak, M., Jiang, W., Waugh, R.A.,
Frid, D.J., et al., 1997. Effects of mental stress on myocardial ischemia
during daily life. JAMA 277, 1521–1526.
Hertz-Picciotto, I., 1995. Epidemiology and quantitative risk assessment: a
bridge from Science to Policy. Am. J. Public Health 85, 484–491.
Ising, H., Babisch, W., Kruppa, B., 1999. Noise-induced endocrine effects
and cardiovascular risk. Noise Health 4, 37–48.
Janes, H., Dominici, F., Zeger, S.L., 2007. Trends in air pollution and
mortality. An approach to the assessment of unmeasured confounding.
Epidemiology 18, 416–423.
Jern, C., Eriksson, E., Tengborn, L., Risberg, B., Wadenvik, H., Jern, S.,
1989. Changes in plasma coagulation and fibrinolysis in response to
mental stress. Thromb. Haemost. 62, 767–771.
Jerrett, M., Burnett, R.T., Ma, R., Pope, C.A., Krewski, D., Newbold,
K.B., et al., 2005. Spatial analysis of air pollution and mortality in Los
Angeles. Epidemiology 16, 727–736.
Kop, W.J., Vingerhoets, A., Kruithof, G.J., Gottdeiner, J.S., 2003. Risk
factors for myocardial infarction during vacation travel. Pschosomatic.
Med. 65, 396–401.
Koton, S., Tanne, D., Bornstein, N.M., Green, M.S., 2004. Triggering risk
factors for ischemic stroke. Neurology 63, 2006–2010.
Kouvonen, A., Kivimaki, M., Vaananen, A., et al., 2007. Job strain and
adverse health behaviors: the Finnish Public Sector Study. J. Occup.
Environ. Med. 49, 68–74.
Lampert, R., Jaska, T., Burg, M.M., Batsford, W.P., McPherson, C.A.,
Jain, D., 2002. Emotional and physical precipitants of ventricular
arrhythmia. Circulation 106, 1800–1805.
140 J. Bukowski / Regulatory Toxicology and Pharmacology 51 (2008) 135–140
Levy, J.I., Bennett, D.H., Melly, S.J., Spengler, J.D., 2003. Influence of
traffic patterns on particulate matter and polycyclic aromatic hydro-
carbon concentration in Roxbury, Massachusetts. J. Expo. Anal.
Environ. Epidemiol. 13, 364–371.
McKinzie, C., Altamura, V., Burgoon, E., Bishop, C., 2006. Exploring the
effect of stress on mood, self-esteem, and daily habits with psychology
graduate students. Psychol. Rep. 99, 439–448.
Miniño, A.M., Heron, M.P., Smith, B.L. 2006. Deaths: Preliminary Data
for 2004. In: National Vital Statistics Reports, vol. 54, No 19. National
Center for Health Statistics, US Centers for Disease Control and
Prevention.
Mittleman, M.A., MacClure, M., Sherwood, B., Mulry, R.P., Tofler,
G.H., Jacobs, S.C., et al., 1995. Triggering acute myocardial infarc-
tion onset by episodes of anger Determinants of Myocardial Infarction
Onset Study Investigators. Circulation 92, 1720–1725.
Möller, J., Hallqvist, J., Diderichsen, F., Theorell, T., Rueterwall, C.,
Ahlbom, A., 1999. Do episodes of anger trigger myocardial infarction?
A case-crossover analysis in the Stockholm Heart Epidemiology
Program (SHEEP). Psychosom. Med. 61, 842–849.
Möller, J., Theorell, T., de Faire, U., Ahlbom, A., Hallqvist, J., 2005.
Work related stressful life events and the risk of myocardial infarction.
Case-control and case-crossover analyses within the Stockholm heart
epidemiology programme (SHEEP). J. Epidemiol. Community Health
59, 23–30.
Moolgavkar, S.H., 2005. A review and critique of the EPA’s rationale for
a fine particulate standard. Regul. Toxicol. Pharmacol. 42, 123–144.
Morgenstern, H., 1998. Ecological studies. In: Rothman, K.J., Greenland,
S. (Eds.), Modern Epidemiology, second ed. Lippincott-Raven,
Philadelphia, pp. 459–480.
Muller, J.E., Tofler, G.H., Stone, P.H., 1989. Circadian variation and triggers
of onset of acute cardiovascular disease. Circulation 79, 733–743.
Myers, A., Dewar, H.A., 1975. Circumstances attending 100 sudden
deaths from coronary artery disease with coroner’s necropsies. Br.
Heart J. 37, 1133–1143.
Netterstrom, B., Juel, K., 1988. Impact of work-related and psychosocial
factors on the development of ischemic heart disease among urban bus
drivers in Denmark. Scand. J. Work Environ. Health 14, 231–238.
Niaura, R., Shadel, W.G., Britt, D.M., Abrams, D.B., 2002. Response to
social stress, urge to smoke, and smoking cessation. Addict. Behav. 27,
241–250.
Ostro, B., Broadwin, R., Green, S., Feng, W.F., Lipsett, M., 2006.
Particulate air pollution and mortality in nine California counties:
results from CALFINE. Environ. Health Perspect. 114, 29–33.
Radi, S., Ostry, A., Lamontagne, A.D., 2007. Job stress and other
working conditions: relationships with smoking behaviors in a
representative sample of working Australians. Am. J. Ind. Med. 50,
584–596.
Robinson, A.A., 1989. Cerebrovascular disease, ischaemic heart disease,
and the stress of vehicle travel. Med. Hypotheses 30, 101–104.
Robinson, J.D., Cinciripini, P.M., 2006. The effects of stress and smoking
on catecholaminergic and cardiovascular response. Behavior. Med. 32,
13–18.
Rundell, K.W., Caviston, R., Hollenbach, A.M., Murphy, K., 2006.
Vehicular air pollution, playgrounds, and youth athletic fields. Inhal.
Toxicol. 18, 541–547.
Sanderson, E.G., Farant, J.P., 2004. Indoor and outdoor polycyclic
aromatic hydrocarbons in residences surrounding a Soderberg alumi-
num smelter in Canada. Environ. Sci. Technol. 38, 5350–5356.
Schwartz, J., Zenobetti, A., Bateson, T. 2003. Morbidity and mortality
among elderly residents of cities with daily PM measurement. In:
Revised analyses of time-series studies of air pollution and health.
Special Report. Health Effects Institute, Boston, MA.
Servoss, S.J., Januzzi, J.L., Muller, J.E., 2002. Triggers of acute coronary
syndromes. Prog. Cardiovasc. Dis. 44, 369–380.
Shapiro, S., 1994. Meta-analysis/Shmeta-analysis. Am. J. Epidemiol. 140,
771–778.
Shapiro, S., 2000. Bias in the evaluation of low-magnitude associations: an
empirical perspective. Am. J. Epidemiol. 151, 939–945.
Siegrist, J., Rodel, A., 2006. Work stress and health risk behavior. Scand.
J. Work Environ. Health 32, 473–481.
Soliman, A.S., Jacko, R.B., Palmer, G.M., 2006. Development of an
empirical model to estimate real-world fine particulate emission
factors: the traffic air quality model. J. Air Waste Manag. Assoc. 56,
1540–1549.
Strike, P.C., Steptoe, A., 2005. Behavioral and emotional triggers of acute
coronary syndromes: a systematic review and critique. Psychosom.
Med. 67, 179–186.
Strike, P.C., Perkins-Porras, L., Whitehead, D.L., McEwan, J., Steptoe,
A., 2006a. Triggering of acute coronary syndromes by physical
exertion and anger: clinical and sociodemographic characteristics.
Heart 92, 1035–1040.
Strike, P.C., Magid, K., Whitehead, D.L., Brydon, L., Bhattacharyya,
M.R., Steptoe, A., 2006b. Pathophysiological processes underlying
emotional triggering of acute cardiac events. PNAS 103, 4322–4327.
Szklo, M., 2001. The evaluation of epidemiologic evidence for policy-
making. Am. J. Epidemiol. 154 (Suppl.), S13–S17.
Tapia Granados, J.A., 2005. Increasing mortality during the expansions of
the US economy, 1900–1996. Int. J. Epidemiol. 34, 1194–1202.
Toffler, G.H., 1997. Triggering and the pathophysiology of acute coronary
syndromes. Am. Heart J. 134, S55–S61.
van Roosbroeck, S., Wichmann, J., Janssen, N.A., Hoek, G., van Wijnen,
J.H., Lebret, E., et al., 2006. Long-term personal exposure to traffic-
related air pollution among school children, a validation study. Sci.
Total Environ. 368, 565–573.
Verrier, R.L., Hagestad, E.L., Lown, B., 1987. Delayed myocardial
ischemia induced by anger. Circulation 75, 249–254.
Verrier, R.L., Mittleman, M.A., 1996. Life-threatening cardiovascular
consequences of anger in patients with coronary heart disease. Cardiol.
Clin. 14, 289–307.
Vinereanu, D., 2006. Risk factors for atherosclerotic disease: present and
future. Herz 31 (Suppl. 3), 5–24.
Wickens, C.M., Wiesenthal, D.L., 2005. State driver stress as a function of
occupational stress, traffic congestion, and trait stress susceptibility. J.
Appl. Behavior. Res. 10, 83–97.
Willich, S.N., Lewis, M., Lowel, H., Arntz, H.-R., Schubert, F., Schroder,
R., 1993. Physical exertion as a trigger of acute myocardial infarction.
NEJM 329, 1684–1690.
Wilson, P.W.F., Cullerton, B.F., 1998. Epidemiology of cardiovascular
disease in the United States. Am. J. Kidney Dis. 32 (Suppl. 3), S56–
S65.
Wynder, E.L., 1987. Workshop on guidelines to the epidemiology of weak
associations. Prev. Med. 16, 139–141.
Yamamoto, K., Irie, M., Sakamoto, Y., Ohmori, S., Yoshinari, M., 2007.
The relationship between IMPS-measured stress score and biomedical
parameters regarding health status among public school workers. J.
Physiol. Anthropol. 26, 149–158.