8/19/2019 Air Pollution and Acute Cardio Respiratory Visits.50
1/15
ABSTRACTS
ISEE-1
SCENARIOS OF VECTOR-BORNE DISEASES AND HEAT STRESS
IN EUROPE
Michael van Liehuit, Pim Martens. ICIS
Abstract: Global climate change remains one of the biggest
environmental threats over the coming century. The impacts on human
population health are highly uncertain but are an important focus of the
policy debate regarding mitigation and adaptation. Within the cCASHh-
project complex integrative approach, aims to improve our understanding
of future exposure, and adaptation possibilities and capacities to vector-
borne diseases as well as heat exposure in different areas of the European
Region.
Methodology: An integrated framework, combining quantitative and
qualitative computer modelling and integrated scenario analysis has been
developed to address the future vulnerability of human health in Europe.
The existing models incorporate the dynamic global and local interaction
between the most important drivers.
Results: This work will present the results of the future risk of vector-
borne diseases under a changing environment as well as future risk of
population exposure to heat stress under different adaptation scenarios.Climate change might also affect human health directly. Most climate
scenarios show an increase in the frequency of temperature extremes,
which would entail increases in thermal stress, and consequently on
morbidity and mortality rates. The fact of a greying European population
and a scenario-dependent increase of the degree of urbanisation would
even super add to the potential morbidity and mortality rates related to
heat stress. However, depending on the population’s future health status,
the ability to adapt autonomously and or social and institutional means
available, the residual mortality rates might be lower.
European Commision Grant EVK-2-2000-ENV-00070.
ISEE-2
COOKING FUELS AND INDOOR AIR POLLUTION (IAP)
Mukesh Dherani,* Astrid Fletcher,* Uma R,† Kirk Smith‡. *London
School of Hygiene and Tropical Medicine; †The Energy and Resources
Institute, New Delhi; ‡University of California, Berkeley
Introduction: Individuals using unprocessed biomass fuels for cooking
in developing countries are exposed to very high level of particulate
matter (PM), a major component of biomass smoke. The estimated burden
of disease associated with IAP is high; 1.6 million premature deaths a
year globally according to the WHO (2002). This study focuses on the
household factors and cooking practices related to PM5 concentrations.
Method: A cross sectional survey comprising of 864 households selected
through cluster sampling in a semi-urban area, was conducted. People of
age group 50 years living there were interviewed to assess their current
and past cooking practices. In a sub-sample of 82 households PM5 levels
were measured in the kitchen during cooking, and, living room and outdoors over a 24 hour period. Low volume air samplers (224 &endash;
PCXR7, SKC make, USA) with a cyclone (50% cut off at 5 mm) and
flow rate of 1.9 lit/min were used to measure PM5 levels.
Results: All the households (98%) reported using cattle dung and wood
as their regular fuels, mud stove without a chimney was the most common
type of stove (98%). Hara, a locally made stove that uses dung was
present in 55% of households. Liquefied Petroleum Gas (LPG) stoves
were present in more than 50% of households, however, only 5% used
them as their main stove. Almost all the women either cook currently or
used to cook in the past, while only 3% of men used to cook. The mean
PM5 level in the kitchen was 4800 3200 g/m3 (geometric mean (GM)
3800 g/m3) during cooking. The 24-hr average PM5 levels in living
room and outdoors were 450 280 g/m3 (GM 390 g/m3) and 130
40 g/m3, respectively. Linear regression suggested that the quantity of
each fuel (dung, wood and crop residues) was strongly associated with
PM5 level during cooking and use of Hara was the most important factor
contributing to the high PM5 levels at the cooking places. The daily
average PM5 exposure to men was 385 g/m3 h (sd 300, GM 320 g/
m3-h) and to the women was more than double, 825 g/m3 h (sd 675,
GM 590 g/m3 h). The average life time person-hour stay in the kitchen
for men was 2400 hours and for women was 36700 hours at an average
age of 61 years for both men and women.
Discussion: Kitchen levels of PM5, resulting from the excessive use of
biomass fuel specially cattle dung, were one of the highest reported.
Diseases related to IAP are expected be highly prevalent in this area. Use
of LPG was very low. Further research is needed to find out the
prevalence of related diseases, and the reasons for low LPG usage. The
analysis of the association between IAP and age-related eye diseases
(cataract and macular degeneration) is in progress.
Dr GVS Murthy, Dr Sanjeev Gupta and team of AIIMS, New Delhi, for
conductin of the survey.
ISEE-3
SHIFTING TRENDS IN INDOOR AND OUTDOOR POLLUTION IN
AN INDUSTRIALIZED URBAN AREA
Ivan Benes,* Jiri Skorkovsky,† Joseph Pinto‡. *Public Health Institute
Usti nad Labem; †Regional Hygiene Station Usti nad Labem; ‡U.S.
EPA/NCEA
Introduction: This study investigates trends in indoor and outdoor air
pollutants in an industrial city.
Methods: Measurements of PM2.5, its elemental composition, PAHs,
VOCs, formaldehyde and mutagenicity (Ames test) were obtained in an
industrialized city in a mountainous area in central Europe. Measurements
were made for sampling periods of 12 consecutive days each in 8 indoor
and outdoor sites during winters over the past ten years. The indoor
measurements were obtained in flats, chosen on the basis of differences inventilation, and cooking and smoking history. Four of the flats were
occupied by smokers, half of whom could be classified as heavy smokers.
The indoor-outdoor measurements were compared to observations
collected at a “central” site and they were found to be typically higher
and not as highly correlated as would be the case if the sites were
affected mainly by long range transport of secondary PM.
Results: During the measurement period there have been shifts from the
use of solid fuels (e.g., lignite, wood, refuse) to cleaner burning natural
gas in home furnaces that are used for heating. This shift has been
responsible, in part for a dramatic decline in PM levels measured at the
“central” site. However, recent increases in the price of natural gas
compared to the more traditional solid fuels have caused shifts back to the
use of solid fuels for home heating. As a result, air quality has
deteriorated in many neighborhoods, especially in lower income areas.
Concentrations of a number of pollutants such as PM2.5, trace elements(S, P, Cl, K, Zn, Br), PAH’s (Anthracene, Chrysene, BenzoaPyrene,
DiBenzoa,hAnthracene), volatile chlorinated and aromatic hydrocarbons
and mutagenicity were generally higher in the flats occupied by smokers.
Open fireplaces installed in many homes and the transport of products of
combustion of solid fuels from nearby homes were also significant sources
of indoor mutagenicity in non smokers’ homes. Mutagenicity levels in the
homes of some non smokers were similar to those found in the homes of
smokers.
Discussion: Because of changes in emissions patterns in the community
over time, measurements at the central site may not accurately reflect
community exposures. There is thus the possibility of error introduced in
epidemiologic studies conducted in areas in which adequate knowledge of
the changing nature of source characteristics is lacking.
Epidemiology • Volume 15, Number 4, July 2004 S15
8/19/2019 Air Pollution and Acute Cardio Respiratory Visits.50
2/15
This research was sponsored by the Czech Ministry of Health, Project NJ/
5907-3.
ISEE-4
ESTIMATING HEALTH EFFECTS FROM EXPOSURES TOOUTDOOR AND INDOOR SOURCES OF AIR POLLUTION
Haluk Ozkaynak. EPA
Abstract: Individuals are exposed to wide variety of pollutants in various
indoor and outdoor microenvironments during the course of a typical day.
Sources of pollution in various indoor and outdoor locations produce
particulate matter (PM) and gaseous pollutants with different physical and
chemical characteristics, which may, in turn, result in different health
responses. Thus, to fully understand the relationships between exposures
to environmental pollutants and various health outcomes, personal
exposures to principal sources of toxic chemicals have to be characterized.
Proper quantification of the magnitude and timing of personal exposures
to these sources will then require identification of key microenvironments,
media, routes and pathways of exposure that contribute most to an
individual’s exposure. Source apportionment techniques can provide auseful method for determining personal exposure to PM and air toxics
from specific sources. Contributions to personal exposures from sources
that have only outdoor, indoor or both indoor and outdoor components
can be developed for generating either personal or population-level
source-specific exposure estimates, for use in epidemiologic analyses. A
number of epidemiological studies have already evaluated the relationship
between health outcomes and sources of ambient particulate matter and
co-pollutants. These studies suggest the importance of examining sources
and constituents of indoor, outdoor, and personal PM and other criteria
pollutants. However, limitations of existing measurement data on indoor,
outdoor and personal measurements to pollutants of concern (e.g., PM and
its constituents or air toxics) and longitudinal time-activity data on
susceptible population groups, hamper the investigation of health impacts
of specific air pollution sources and its components. Future community
health studies need to collect better information on indoor and outdoor
sources of personal exposures to PM, PM species and various toxics
co-pollutants.
Disclaimer: This is an abstract of a proposed presentation and does not
necessarily reflect EPA policy.
ISEE-5
RELATIONSHIP OF MULTIPLE INDOOR AIR POLLUTANTS AND
SICK BUILDING SYNDROME AT AIR-CONDITIONED OFFICE
BUILDINGS IN TAIWAN
Pei-Chih Wu,* Li Yen-Yi,† Chiang Che-Ming,† Su Huey-Jen,*
Lee Chun-Chang*. *Department of Environmental and Occupational
Health, Medical College, National Cheng Kung University, Taiwan, ROC;† Department of Architecture, College of Engineering, National Cheng
Kung University, Taiwan, ROC
Abstract: Exposures to various indoor factors, particularly air pollutants,
have been implicated in several health outcomes; yet most investigations
had been conducted only with measurements of selected hazards. Our
study was designed to examine associations between indoor air quality
and sick building syndrome (SBS) of employees in 8 air-conditioned
of fice buildings through a series of comprehensive environmental
assessments and questionnaire administration.
Airborne microbes, carbon dioxide (CO2), particulate matter (PM10),
formaldehyde, and total volatile organic compounds (TVOC) were
measured in every sampling site inside the study buildings. Frequency of
reporting symptoms and other environmental variables, such as
environmental tobacco smoke exposure and air change rate, were
documented by self-administrated questionnaires. Information on gender,
age of the occupant, degree of job satisfaction, recent experience of stress,
and personal history of diagnosed allergic diseases were also obtained for
adjustment in statistical analysis. Multiple logistic regression analysis was
used to identify associations between dichotomous scales of air pollutants
and health outcomes of interest, after adjusting for variables described
above.
Our data show that concentrations of indoor formaldehyde, TVOC and
airborne microbes are higher than most recommended levels quoted.
Estimates of crude odds ratios show that most air pollutants, when
analyzed by continuous scales, are associated with SBS of one or more
types. Similar relationships are seen with environmental factors of interest
and personal characteristics acquired. PM10 exposures were significantly
associated with mostly irritant outcomes, in addition to headache in
multivariate analysis. Interestingly, indoor TVOC concentrations appeared
to be protective with respect to symptom of sneezing, while exposure to
greater concentrations of indoor HCHO did not present additional adverse
effects with concurrent presence of other indoor pollutants. Most
strikingly, statistical associations were found between selected SBS and
microbial concentrations, dichotomized at 1000 CFU/m3 for its being areference concentrations adopted in many proposed guidelines.
The critical role of exposure to indoor microbes and PM10 is
quantitatively demonstrated in this investigation. Further work should be
designed to clarify the exact contribution of indoor TVOC exposures for
increasing use of modern compounds, like essential oils, and decoration of
synthetic materials are likely to result in rising concentrations of indoor
TVOC in future indoor environments.
ISEE-6
TOBACCO CONTROL LAWS AND SERUM COTININE LEVELS IN
NONSMOKING ADULTS, 1999-2000
Melanie Pickett,*† Susan Schober,* Debra Brody,* Randy Curtin*.*National Center for Health Statistics; † Association of Teachers for
Preventive Medicine
Introduction: State and local bans or restrictions on indoor smoking are
implemented to reduce exposure to environmental tobacco smoke (ETS), a
known human carcinogen. Cotinine, a major metabolite of nicotine, is a
biomarker of exposure to ETS in nonsmokers. Data from the National
Health and Nutrition Examination Survey (NHANES) show that median
cotinine levels in nonsmokers have declined by 70% since the late 1980 ’s.
The objective of this study is to determine the relationship between State
or local ordinances on indoor smoking and cotinine levels in the
nonsmoking adult population, taking into account racial and ethnic
differences.
Methods: NHANES is a cross-sectional survey designed to monitor the
health and nutritional status of the US population. Household interviews,
physical examinations, and cotinine level assessments were performed on4039 adults (83% response rate), of whom 3003 were non smokers. For
each of 26 study locations, laws regulating indoor smoking (in enclosed
public spaces, workplaces, restaurants, and bars) were categorized into one
of three groups indicating minimal, moderate, or extensive regulations.
The proportion of non-smoking adults with detectable levels of serum
cotinine were compared across these three groups.
Results: Approximately 49% of nonsmoking adults had detectable levels
of cotinine. Non-Hispanic blacks had a higher proportion of detectable
levels (69%) than did non-Hispanic whites (47%) and Mexican-Americans
(44%). For Mexican Americans and non-Hispanic whites, persons living
in areas with extensive smoking regulations had a significantly smaller
proportion with detectable cotinine levels (29%) than persons living in
areas with moderate (52%) and minimal (50%) smoking regulations. The
Abstracts Epidemiology • Volume 15, Number 4, July 2004
© 2004 Lippincott Williams & WilkinsS16
8/19/2019 Air Pollution and Acute Cardio Respiratory Visits.50
3/15
8/19/2019 Air Pollution and Acute Cardio Respiratory Visits.50
4/15
Abstract: Studies of short term effects of air pollution have for many
years been focused on respiratory effects. Nevertheless, recent studies
have shown that a major impact on mortality caused by air pollution is
due to cardiovascular deaths.
The aim of this panel study was to investigate the relation between
fluctuations in air pollution levels and two inflammatory markers in blood;fi brinogen and CRP, which are known risk factors for cardiovascular
disease. The study was conducted in a mid sized Swedish city (Växjö)
during the three first moths of year 2003.
Continuous measurements of PM10, PM2.5 and NO2 were performed in
an urban background station placed in a park in the city centre. The study
panel consisted of 40 females, aged 30-65 (mean 46) years of age and all
working at the same hospital. Blood samples were taken at 5 occasions
during days with high and low predicted pollution levels. Each occasion
was chosen based on the trend in PM concentrations and temperature
observed during previous days. The participants were notified one day
before sampling, and all testing was performed by a nurse at the hospital.
Air pollution levels during the study were moderate. In total 192 blood
samples were analysed. The fi brinogene and CRP results were first
inspected by plotting the ratio between each participant’s level at each
occasion and the mean level of all occasions against the measured
pollution levels (two day average). For fi brinogen the results revealed a
slight increase in the ratio when plotted against NO2, but no similar trend
was found for particulates. A single pollution model, with individual
intercept and adjusted for reported infections, was used to further analyse
the relation between NO2 and fi brinogen and CRP, respectively. The
estimated beta-coef ficients were 0.003 and 0.013 but insignificant.
In this study we found no significant relations between short term
exposure to air pollution and increased levels of fi brinogen or CRP in
blood. One limitation in the study so far is the use of data from only one
measuring station. A further analysis will be based on data from
dispersion models which will increase the precision in the exposure
assessment. In addition other measures of particles will be used.
ISEE-10
DOES DEATH DUO TO MYOCARDIAL INFARCTION INCREASE
AFTER 1 HOUR OF A HIGH CONCENTRATION OF SPM?
ANALYSIS OF HOURLY MEASURED AIR POLLUTION DATA
FROM TOKYO
Yoshitaka Murakami, Masaji Ono. National Institute for Environmental
Studies
Introduction: To investigate the short-term effects of a high
concentration of suspended particulate matters (SPM) on the risk of death
from myocardial infarction (MI), we examined hourly measured air
pollution data in Tokyo from 1990-1994.
Methods: Mortality data of myocardial infarction (ICD9:410) were
obtained from the Vital Statistics of Japan. Air pollution and
meteorological data were obtained from the database of the NationalInstitutes for Environmental Studies and from the Japan Meteorological
Agency. In Japan, SPM is used as the criterion for pollutant particulate
matter, with a diameter less than 10 um (intermediate to that of PM10 and
PM2.5). The data of the Tokyo metropolitan area (including 23 wards and
12 cities) from 1990 to 1994 were used for the examination. To determine
the effect of 1 hour of a high concentration of SPM on MI deaths, we
defined: 1) the level of high concentration of SPM; and 2) an effect
period after the SPM high concentration. We defined a threshold value as
200 ug/m3 or 300 ug/m3 from a distribution of SPM from 1990 to 1994.
An effect period was defined as the subsequent time (3 hours to 120
hours) after SPM concentration exceeded threshold value. Deaths
occurring during the effect/non-effect period were divided by the periods’
person-hours, and risk ratio (effect period ’s risk divided by non-effect
period ’s risk) was calculated. A Poisson regression model was used to
adjust for the effect of temperature, hour of death, and region (Tokyo
metropolitan ward or city). Temperature was treated as a single variable in
the model using the average temperature prior to the period that death
occurred. We determined the effect of temperature on MI deaths using
several models with temperature averaged over different time periods.
Results: The risk ratio with a SPM threshold value of 300 ug/m3 was
larger than the risk ratio with a threshold value of 200 ug/m3. The risk
ratio varied according to the duration of the effect period. When the
threshold value was 300 ug/m3, the risk ratios were 1.14 (95% CI:
0.88 – 1.49), 1.12 (95% CI: 0.91 – 1.38), 1.05 (95% CI: 0.89 – 1.23), 1.06
(95% CI: 0.93 – 1.19) and 1.07 (95% CI: 1.01 – 1.15) when the effect
periods were 3, 6, 12, 24 and 120 hours, respectively. Changing the
duration of the period over which temperature was averaged in the model
had a subtle influence on the risk ratio.
Conclusions: We showed increased risk of MI death in the short period
after a high concentration of SPM, especially 3 – 6 hours after a high
concentration of SPM.
ISEE-11
THE ROLE OF ULTRAFINE PARTICLES AND OTHER TRAFFIC-
RELATED POLLUTANTS ON ISCHEMIC HEART DISEASES:
MAIN RESULTS OF THE HEAPSS PROJECT
Francesco Forastiere,* Pasi Aalto,† Tom Bellander,‡ Niklas Berglind,§
Daniela d ’Ippoliti,* Markku Kulmala,† Timo Lanki, ¶ Fredrik Nyberg,§
Pentti Paatero,† Juha Pekkanen, ¶ Annette Peters, Sally Picciotto,*Massimo Stafoggia,* Jordi Sunyer,** Pekka Tiittanen, ¶
Stephanie Von Klot, Roberto Elosua**. *Dept. Epidemiology, ASL RME;† University of Helsinki; ‡ Dept. of Occupational and Environmental
Health, Stockholm County Council; § Institute of Environmental Medicine,
Karolinska Institutet; ¶ KTL-National Public Health Institute; GSF-
National Research Center for Environment and Health; **IMIM-
Municipal Institute for Medical ResearchBackground: Health Effects of Air Pollution on Susceptible
Subpopulations (HEAPSS) is an EU-funded study of the effects of air
pollution on cardiac health in five European cities characterized by
different air quality and climates. It aimed to determine acute effects of
air pollution on: 1) fatal non-hospitalized coronary events; 2) hospitalized
acute myocardial infarctions (AMI), as well as; 3) re-hospitalizations for a
new AMI or any cardiac cause; and 4) late mortality, in AMI survivors (a
hypothesized susceptible subpopulation).
Methods: Hospital discharge records and AMI registries identified
26,888 AMI patients, 1992-2000. In one city, data on 5,144 out-of-
hospital coronary deaths were collected. Mortality and hospital
readmissions (for acute myocardial infarction and other cardiac causes)
were assessed in follow-ups of 28-day survivors. PM10, CO, NO2, NO,
SO2, and O3 were collected for each city from existing networks of fixed
monitors. Particle Number Concentration (PNC) was measured for oneyear (using condensation particle counters) and retrospectively estimated
for the study period. City specific time series analyses with Poisson
regression were conducted. Results were pooled using fixed or random
effects models.
Results: Estimated PNC (unit: 10,000 particles/cm3) and CO (unit: 0.2
mg/m3) at immediate lag showed the most consistent results. PNC (2.8%
increase, 95% CI 1.1-4.6%) and CO (1.0% CI 0.3-1.8%) were
strongly associated with out-of-hospital coronary events. Hospitalizations
for first AMI were weakly associated with PNC and CO. The associations
were stronger in the three cities using hospital discharge records and no
upper age limit: PNC was associated with 1.3% (CI 0.0-2.6%) and CO
with 0.7% (CI 0.1-1.2%) increase in first MI hospitalizations, and the
strongest associations were observed among persons aged 65 and over.
Abstracts Epidemiology • Volume 15, Number 4, July 2004
© 2004 Lippincott Williams & WilkinsS18
8/19/2019 Air Pollution and Acute Cardio Respiratory Visits.50
5/15
There was variation between cities with respect to effect modification by
age. PNC (2.6%, CI 0.5-4.8) and CO (1.4%, CI 0.1-2.6%) were
associated with cardiac readmissions among myocardial infarction
survivors in the five cities. Air pollution was weakly associated with all-
cause mortality in AMI survivors, with considerable heterogeneity
between cities. In the two centers using AMI register data, an association
was found for both PNC (7.0%, CI 2.0-12%) and CO (6.0%, CI 0-
12) at lag 0-1.
Conclusions: Air pollution from traf fic is associated with out-of-hospital
coronary deaths, hospitalizations for myocardial infarction, and subsequent
cardiac readmissions among AMI survivors. Weak and inconsistent results
were found for total mortality. Differences in air pollution levels, age and
sex structure, and treatment practices among centers might explain the
heterogeneity of the results and require further analysis.
Funded by EU (QLK4-2000-00708).
ISEE-12
ASSOCIATIONS BETWEEN CARDIAC ARRHYTHMIA AND
AMBIENT AIR POLLUTANTS IN AN ELDERLY POPULATION
Stefanie T. Ebelt,* Helen H. Suh,* Brent A. Coull,* Joel Schwartz,*
Peter H. Stone,† Diane R. Gold ‡. *Harvard School of Public Health;
† Cardiovascular Division, Brigham and Women’ s Hospital, Harvard
Medical School; ‡The Channing Laboratory, Brigham and Women’ s
Hospital, Harvard Medical School
Introduction: Cardiovascular mortality and morbidity has been strongly
associated with ambient air pollutant concentrations in numerous
epidemiological studies. A proposed mechanism by which particles exert
their effects involves autonomic nervous system dysfunction, which may
lead to arrhythmia. Here we examine air pollution-induced cardiac
arrhythmia in an elderly population.
Methods: 32 non-smoking older adults (mean age 70.8 years, 3 male,
29 female) participated in a repeated measures study during the summer
and fall of 2000. The cardiovascular health of each participant wasassessed weekly for 24 weeks following a standardized protocol that
included continuous EKG measurements through 5-minute intervals of
supine baseline rest, standing, exercise outdoors, supine second rest and
deep breathing. For each part of the protocol, EKG data were analyzed for
counts of supraventricular ectopy (SVE) and ventricular ectopy (VE).
Ambient hourly PM2.5, gases and meteorological parameters were
collected at a central site within one mile from participant residences and
clinic rooms. Logistic mixed effects regressions predicting the presence of
arrhythmia were performed for the entire protocol and for each interval,
controlling for subject, apparent temperature and time trend, for pollutant
lags up to 24-hours and moving averages up to 120-hours prior to the
health assessment. Results were scaled to interquartile range (IQR)
differences in pollution.
Results: Over the 24-week study period, a total of 645 participant visits
were conducted. Mean counts for SVE (2787) and VE (2892) over the protocol varied strongly by subject, with subject-specific means between
0-363 for SVE and 0-350 for VE. The odds of having an arrhythmia were
most strongly associated with PM2.5
, O3
and NO2
, and were weaker for
CO, NO and SO2. For SVE, both short- and long-term exposures appeared
important. For example, while standing, significant O3 odds ratios (OR) of
up to 3.80 (95% CI: 1.89-7.61, IQR 38.5 ppb) were observed in the 14-
21 hours prior to the health assessment, as well as for the mean 48-hours
prior (OR 3.50, 95% CI: 1.62-7.55, IQR 21.7 ppb). For VE, long-term
exposures appeared the most important, with an OR of 1.72 (95% CI:
1.02-2.92, IQR 11.2 ug/m3) for 96-hour PM2.5 and an OR of 5.68 (95%
CI: 1.78-18.12, IQR 21.4 ppb) for 96-hour O3.
Discussion: Results suggest that there is an increased risk of cardiac
arrhythmia with elevated ambient PM2.5
and O3
concentrations. Future
analyses will examine pollution associations with the counts of
arrhythmias within each visit, and will examine confounding and synergy
among pollutants.
This work is supported by funding from the NIH/NIEHS (ES-09825), U.S.
EPA (R826780-01-0, R827353-01-0) and the U.S. DOE
(EP-P4464/C2166).
ISEE-13
CARDIOVASCULAR DISEASE HOSPITALIZATIONS AMONG
MEDICARE ELDERLY AND PARTICULATE MATTER AIR
POLLUTION IN 6 U.S. CITIES
Lynn Disney, Case Western University, OH
Introduction: The Harvard Six Cities study was seminal in establishing
the link between particulate air pollution and premature mortality. It has
been reviewed and reanalyzed numerous times finding similar results. In
this study, the same cities were studied examining hospitalization rates
among the Medicare elderly ( 65).
Methods: Zip codes corresponding to the 6 cities (Watertown, MA;
Topeka, KS; Portage, WI; Harriman, TN; Steubenville, OH and St. Louis,
MO) were identified from the U.S. Postal Service. Medicare eligible
beneficiaries who resided in the six Harvard cities were selected from the
1989 denominator files. The cohort was restricted to white residents over
age 65 who were not enrolled in an HMO in 1989. Hospitalization data
(Part A) were merged by HCFA claim number. CVD (ICD-9 400.0-440.0)
admissions were then identified among these beneficiaries for 9 years after
cohort year (1989-1997). The first identified hospitalization with ICD-9 in
the specified range was coded as a CVD admission. CVD admission
compared to no CVD admission was analyzed using logistic regression
analysis. Age at baseline in 1989, sex and high school graduation rates
were all measured confounders. High school graduation rates wereobtained from zip code level census data.
Results: 149,634 white residents over age 65 were identified. The
demographic distribution of the study population is displayed below.
Table 1. Demographics of Study Population
Number %Female %Not HS graduate
MA 5,015 65.3 16.2
TN 3,811 58.5 32.0
OH 5,792 61.1 27.9
WI 2,036 58.1 21.6
KS 15,421 63.0 14.7
MO 117,559 64.0 23.4
In all, 90,360 hospitalizations were identified for CVD from 1989-1997.The logistic regression results are shown graphically in Figure 1.
Discussion: The observed results showing the association between
cardiovascular diseases and fine particulate matter are very similar to
those observed by Dockery et al. The area with the highest levels of
particulate matter also had the highest probability of admission for CVD.
Similarly, residence in Portage, WI and Topeka, KS are at the lowest risk
of CVD admission. This study supports that of others that fine particulate
matter may contribute to the prevalence of CVD. The main disadvantage
of this study is the lack of direct measurement of confounding variables.
Key to this analysis is lack of information on places of residence and
personal risk factors such as cigarette smoking, obesity and occupational
exposures. These results show that claims data can be used to support
population based studies of environmental risks.
Epidemiology • Volume 15, Number 4, July 2004 Abstracts
© 2004 Lippincott Williams & Wilkins S19
8/19/2019 Air Pollution and Acute Cardio Respiratory Visits.50
6/15
ISEE-14
EFFECTS OF NITROGEN DIOXIDE ON HEART RATE
VARIABILITY
Chang-Chuan Chan,* Kai-Jen Chuang,* Ta-Chen Su,† Lian-Yu Lin†.
*Institute of Occupational Medicine and Industrial Hygiene, College of
Public Health, National Taiwan University; † Department of Internal
Medicine (Cardiology Section), National Taiwan University Hospital
Introduction: Although many studies have shown decrease in heart rate
variability (HRV) was associated with particulate air pollutants, relatively
few studies have investigated its association with gaseous air pollutants.
Objective: This study was to examine the relationship between nitrogen
dioxide (NO2) and HRV in the elderly population.
Methods: We used a panel study to investigate short-term effects of air
pollution on HRV. Our study subjects are 83 patients (mean age 61) fromthe cardiology section, Department of Internal medicine, National Taiwan
University Hospital in Taipei. Continuous ambulatory electrocardiographic
(ECG) monitoring was performed on each study subject, which measured
time and frequency domains of HRV, such as standard deviation of
normal-to-normal (NN) intervals (SDNN) and the square root of the mean
of the squared differences between adjacent NN intervals (r-MSSD), high
frequency (HF, 0.15-0.40 Hz) low frequency (LF, 0.04-0.15 Hz). NO2 and
other co-pollutants, such as sulfur dioxide (SO2), carbon monoxide (CO),
ozone, PM10 (particulate matter less than 10 um in diameter) and PM2.5
(particulate matter less than 2.5 um in diameter) measured at each
subject’s close-by monitoring stations were used to represent personal
exposure. We used linear mixed-effects models in single pollutant and
multiple pollutant models to estimate the relationship between air
pollution and log10-transformed HRV. Potential confounders, such as
subject’s age, sex, body mass index (BMI), disease status of coronaryheart diseases, smoking status, and ambient temperature were also
adjusted in all models.
Results: Single pollutant models showed that NO2 is the most consistent
air pollutants responsible for both time and frequency domains of HRV
decreases among all air pollutants. Such effects occurred at 1 to 4-hour
moving averages. CO is also responsible for decreases in SDNN, r-MSSD,
LF, and HF at 1-hour moving CO average. By contrast, we found no
associations between HRV and PM2.5, PM10, SO2, or ozone. Multiple
pollutant models with NO2 and CO further showed that NO2 was
responsible for SDNN decreases at 1 to 4-hour moving NO2 average after
controlling confounders. The models showed a 1ppb increase in NO2
exposure with 1-4 hour moving averages was associated 0.08-0.17%
decreases in SDNN.
Discussion: Our findings suggested that exposures to traf fic-related air
pollutants in urban environment, such as NO2 or CO, are associated with
decreased HRV in susceptible population.
ISEE-15CAN THE PM2.5-INDUCED HEART RATE VARIABILITY
CHANGES BE PREVENTED? RESULTS FROM A RANDOMIZED
STUDY OF FISH AND SOY OIL SUPPLEMENTATION
Fernando Holguin,* Mara Tellez-Rojo,† Mariana Lazo,†
Abigail Manzano,† Marlene Cortez,† Mauricio Hernandez,†
David Mannino,‡ Stephen Redd,‡ Pierre Julien,§ Marie-Claire Belanger,§
Isabelle Romieu†. *Emory University Department of Medicine and
Centers for Disease Control; † Instituto Nacional de Salud Publica;
‡Centers for Disease Control; § Laval University, Lipid Research Center,
Quebec
Introduction: Supplementation with omega-3 fatty acids is associated
with increased heart rate variability. However, it is unknown whether
omega-3 fatty acids can prevent or lessen the reductions in heart rate
variability associated with environmental exposure to PM2.5.Methods: Nursing home residents in Mexico City were followed for 5
months with alternate 6-min heart rate variability (HRV) measurements
and simultaneous monitoring of daily indoor and outdoor 24-h average
levels of PM2.5. After 1 month of follow up, participants were
randomized to receive 2 gr/day of fish oil (85% Eicosopentanoic acid and
Docosahexaenoic acid) in divided doses vs 2 gr/day of soy oil (6.78%
alpha-Linolenic acid). We used a mixed effects model to estimate the
effect of the same-day PM2.5 on HRV parameters, and adjusted the
model for age, gender and HR and an interaction term for fish and soy
oil.
Results: A total of 50 subjects were enrolled in the study. The average
age was 81 years (65-95), and 35% were male. During the study, the
levels of PM2.5 were: 19.18 g/m3 (SD 10.46) (5.09 – 106.6). A total of
1610 HRV measurements were done during follow up. The concentration
of omega-3 (in red cell membranes) increased from 6.4 to 13.5% in thefish oil group (p0.01) and from 7 to 8.8% (p0.04) in the soy oil
group. Pre-randomization, exposure to same-day PM2.5 was associated
with a significant reduction in the high frequency (HF) HRV component
coef 0.034 (p0.01) and SDNN (standard deviation of normal to
normal heart beats) coef 0.013 (p0.01). Fish oil supplementation
prevented the reduction associated with same exposure to PM2.5 in the
HF-HRV (coef 0.0021; p0.5) and SDNN (coef 0.0005; p0.6)
parameters. Soy oil supplementation did not prevent reductions in HF-
HRV (coef 0.008; p0.01); however, it prevented the reduction in
SDNN (coef 0.0025; p0.11).
Conclusions: This is the first study to show that the cardiac autonomic
response to ambient PM2.5 in the elderly can be prevented by
supplementation with omega-3 fatty acids, and to a lesser extent with
alpha-Linolenic acid (which is partially metabolized to omega-3 fatty
acids) in soy oil.
ISEE-16
ARRHYTHMOGENESIS IN HUMAN RESPONSE TO
PARTICULATE AIR POLLUTANTS
Jiu-Chiuan Chen,* Peter H. Stone,† Joel Schwartz,* Jee-Young Kim,*
Robert F. Herrick,* David C. Christiani*. *Department of Environmental
Health, Harvard School of Public Health; † Cardiovascular Division,
Brigham and Women’ s Hospital
Introduction: Epidemiologic studies have shown that exposure to
particulate matter (PM) increased arrhythmia-related mortality and
morbidity. Patients with arrhythmia are also found to be more sensitive to
FIGURE 1. Odds Ratios from Logistic Regression Model v. Fine
Particulate
Abstracts Epidemiology • Volume 15, Number 4, July 2004
© 2004 Lippincott Williams & WilkinsS20
8/19/2019 Air Pollution and Acute Cardio Respiratory Visits.50
7/15
the PM-mediated acute cardiac effects. Although these epidemiologic
observations are supported by results of many animal studies, direct
human data so far have only provided evidence supporting the
contribution of arrhythmia to population susceptibility.
Methods: We investigated whether particulate pollutants result in
arrhythmogenesis in a short-term prospective study of 10 male
boilermakers (age 34.38.1 years) with no overt cardiac diseases. They
were exposed to high levels of metal particulates in their workplace.
Episodes of extrasystoles within each 5-minute epoch were determined by
validated algorithms across 24-hour Holter recordings. The PM2.5 (PM
with aerodynamic diameter 2.5 um) levels were monitored concurrently
using personal air samplers. Information on the time profile of daily
activities was retrieved from detailed diary records. We used
autoregressive logistic regression to model the probability of extrasystoles,
and employed autoregressive Poisson regression to estimate the average
counts of premature complexes. Supraventricular and ventricular
extrasystoles were analyzed separately.
Results: The cross-shift average PM2.5 level was 1.78 1.25 mg/m3.
Of all the recorded 5-min epochs, 7.6% (95% CI: 6.3, 9.1) had
supraventricular extrasystoles and 23% (95% CI: 19, 28) had ventricular
extrasystoles. Within each positive recording, the average number of extrasystoles was 1.6 (ranging from 1.3-1.9) per 5-min for
supraventricular beats and 5 (ranging from 4.0 to 6.5) per 5-min for
ventricular beats. We found a statistically significant association between
exposures to PM2.5 and the increased extrasystole frequency. For each
1-mg/m3 increase in the preceding 4-hour average concentration of
PM2.5, the associated OR was 1.82 (95% confidence interval CI: 1.52,
2.17) for having supraventricular extrasystoles, and the frequency of
supraventricular premature beats increased by 72% (95% CI: 52, 94). For
each 1-mg/m3 increase in the preceding 6-hour average concentration of
PM2.5, the associated OR was 1.58 (95% CI: 1.18, 2.12) for having
ventricular extrasystoles, and the frequency of ventricular premature beats
increased by 21% (95% CI: 10, 32). After adjusting for circadian rhythm
(morning, afternoon, evening, night), sleeping, smoking, alcohol use,
coffee drinking, 5-min average heart rate, and personal coronary risk
profiles, we observed a consistently positive association between PM2.5and supraventricular extrasystoles. The adjusted analyses revealed a
slightly stronger association between PM2.5 and the probability and
frequency of ventricular premature complexes.
Discussion: These results provide significant evidence indicating that
high levels of metal particulates may induce extrasystoles in healthy
working men without overt heart diseases.
This abstract is funded by the NIH grant (ES 09860). Dr. J.C. Chen was
supported by the Liberty Mutual-Harvard Program for Occupational
Safety and Health.
ISEE-17CARDIOVASCULAR EFFECTS OF AIR POLLUTION IN ADULTS
IN CUBATÃO, SÃO PAULO, BRAZIL
Luiz Alberto Amador Pereira,* Gleice Margarete de Sousa Conceição,†
Alfesio L.F.Braga‡. *Catholic University of Santos; Brazil; † Department
of Internal Medicine, Federal University of Sã o Paulo School of
Medicine; ‡ Program of Community Health, Catholic University of Santos;
and Environmental Pediatrics Program University of Santo Amaro School
of Medicine
Abstract: Cubatão is a southeast seaboard Brazilian city that has a very
important industrial center with a steel mill, several petrochemical and
chemical industries, surrounded on three sides by mountains that interfere
in air pollution dispersion. Because of that, the region was called in the
80’s of the XX Century the “Valley of Death”. Since then, some air
pollution control policies were adopted in order to decrease air pollution
level.
Objective: A time-series study was conducted to estimate air pollution
effects on cardiovascular (CVD) hospital admissions using records from
Brazilian public health system for people older than 30 years of age, in
the city of Cubatão, from 1997 to 2000, and air pollution and weather
data provided by São Paulo State Environmental Agency.
Methods: We performed analysis for the whole group and for two
sub-groups: people from 30 to 64 years and older than 64 years of age.
We adopted generalized linear Poisson regression single-pollutant models
to assess the effects of PM10, CO, NO2, SO2, and O3 on CVD
controlling for seasonality (long and short-term trends) and weather
variables using a semi-parametric smoother (loess). We adopted more
stringent convergence parameters than the S-PLUS software default. The
eight days cumulative effects of air pollutants were assessed using third-
degree polynomial distributed lag models.
Results: PM10 was positively associated with cardiovascular diseases for
two age groups: 30 to 64 years of age and older than 64 years of age. An
interquartile range increase on PM10 concentration (40g/m3) was
associated with increases of 16% (95% CI &endash; 1;30) and 21% (95%
CI&endash;3;38) in CVD admissions for the younger and the older agegroups, respectively. The associations of the other studied pollutants with
CVD were also positive but not statistically significant.
Conclusion: Despite of the polices adopted to reduce air pollution in the
City of Cubatão in the last two decades, this study showed that air
pollution has a remarkable impact on cardiovascular morbidity and
reinforces the necessity of additional air pollutant emission-controlling
polices in that area.
This study was funded by: UNISANTOS, CEACL-UNIFESP, LIM05-
FMUSP, and UNISA.
ISEE-18
THE EFFECT OF PARTICULATE AIR POLLUTION ON THE RISK
OF MYOCARDIAL INFARCTION IN ELDERLY: A MULTI-CITY
CASE-CROSSOVER ANALYSIS
Antonella Zanobetti, Gregory A. Wellenius, Ariana Zeka, Joel Schwartz.
Harvard School of Public Health
Abstract: Few studies have examined the acute effects of air pollution on
the occurrence of hospital admissions for myocardial infarctions (MI),
pneumonia, and chronic obstructive pulmonary disease (COPD), and these
have mostly been single cities studies. We conducted a multi-city case-
crossover study of the acute effect of PM10 on the increased risk of being
admitted to the hospital among the elderly in twenty-four US cities.
We examined the association between daily particulate matter (PM10) and
Medicare data (age GE 65) on hospital admissions from the emergencyroom with a primary diagnosis of MI (ICD-9: 410), COPD (ICD-9: 490-
496, except 493) and pneumonia (ICD-9: 480-487), in those cities for the
years 1985-1998. In the first stage of the analysis, we applied a case-
crossover design in each city with referent exposure days chosen using the
time-stratified approach such that exposures on the case day were
compared to exposures occurring every third day of the same month as
the case day. We controlled for apparent temperature and day-of-week in
the model. In the second stage of the analysis we combined the city-
specific estimates using a random effect approach.
We found that a 10 g/m3 increase in PM10 was associated with a 0.72%
increase (95% CI: 0.3, 1.1) in risk of admissions for MI on the same day.
When we analyzed COPD we found a 0.6% increase (95% CI: 0.1, 1.2)
for 10 g/m3 increase in PM10, while we found a 0.69% increase (95%
Epidemiology • Volume 15, Number 4, July 2004 Abstracts
© 2004 Lippincott Williams & Wilkins S21
8/19/2019 Air Pollution and Acute Cardio Respiratory Visits.50
8/15
CI: 0.4, 1.0) in risk of admissions for pneumonia and PM10 on the same
day.
We conclude that our analysis showed an increased risk of hospitalization
for MI, COPD and pneumonia associated with PM10. More studies are
needed to understand the biological mechanism and to examine effect
modification.Founded by: EPA R82735301 and NIEHS ES00002
ISEE-19
AIR POLLUTION AND MARKERS OF CARDIOVASCULAR RISK
Antonella Zanobetti,* Joel Schwartz,* Paul M. Ridker †. *Harvard School
of Public Health; † Brigham and Women’ s Hospital
Abstract: Air pollution has been associated with increased risk of death
and hospital admissions all over the world. Many of those associations are
from cardiovascular disease, however the mechanism by which airborne
particles increases the risk of cardiovascular deaths is still being explored.
We looked at the effects of medium term and long term (annual mean) of
particulate air pollution (PM2.5), ozone (O3), carbon monoxide (CO), and
nitrogen dioxide (NO2), and changes in blood markers of cardiovascular
risk, including HDL cholesterol, triglycerides and C-reactive protein, in
the PRINCE study, a large, national randomized trial of the use of statins
to lower such risk factors. A total of 5778 participants were matched to
exposure from monitors in their county of resident.
We controlled for subject, trend, season, age, gender, race, body mass
index, education, smoking history, hormone replacement therapy, exercise,
and whether receiving a statin or a placebo.
We found a medium term (mean of the two previous months) effect
between C-reactive protein and PM2.5, CO, NO2, with an increase of
7.7% (95% CI: 0.96-14.96) for 10 mg/m3 of PM2.5, an increase of 3.8%
(95% CI: 0.3-7.3) for 5 ppm of NO2, an increase of 2.1% (95% CI: 0.1-
4.1) for 0.5 ppm of CO. When we examined interactions with statin use
we found a higher increase in subjects on placebo.
We found a significant association between triglycerides and the mean of
the three previous months of ozone with an increase of 0.95% (95% CI:
0.1-1.8) for 5 ppm of O3. For the same time period for PM2.5, we found
that the association was almost entirely in the subjects not receiving
statins. In those subjects we found for the mean of the three previous
months an increase of 8.3% (95% CI: 1.2-15.8).
For HDL cholesterol we fund an association with a longer averaging
period (annual mean) of pollution, with a – 3% increase (95% CI:-5.7- -
0.3) for 10 mg/m3 of PM2.5 and a – 1% increase (95% CI:-1.7- -0.3) for 5
ppm of O3.
These results present new the evidence of possible mechanism for the
association between air pollution and cardiac events, and the cancellation
of the effect by statins points to mechanistic pathways effected by statins.
Founded by: EPA R82735301 and NIEHS ES00002.
ISEE-20
EFFECTS OF PARTICULATE AIR POLLUTION ON MEAN
DECELERATION MAGNITUDE IN PATIENTS WITH CORONARY
HEART DISEASE: A NOVEL APPROACH TO ASSESS HEART
RATE FLUCTUATIONS
Angela Ibald-Mulli,* Regina Rückerl,† Raphael Schneider,‡ Axel Bauer,§
Georg Schmidt,‡ Josef Cyrys,* H.-Erich Wichmann,* Wojciech Zareba, ¶
Annette Peters†. *GSF-Insitute of Epidemiology and IBE Department of
Epidemiology, Ludwig-Maximilians-University Munich; † GSF-Insitute of
Epidemiology; ‡Working Group of Biological Signal Analysis, 1.
Medizinische Klinik, Technical University, Munich; § Working Group of
Biological Signal Analysis, Deutsches Herzzentrum Mü nchen, Department
of Adult Cardiology; ¶ Department of Medicine, University of Rochester
Abstract: Epidemiological studies on air pollution and cardiovascular
function showed adverse effects of particulate air pollution on
conventional parameters of heart rate variability (HRV). Between October
2000 and April 2001 a panel study with 58 coronary heart disease patientswas conducted in Erfurt, Germany to assess the association between
cardiac function and daily variations in particulate air pollution. In the
present analysis the mean deceleration magnitude (MDM), a new
parameter to characterise fluctuations in heart rate, was applied to assess
alterations in cardiac function associated with exposure to elevated levels
of particulate air pollution. MDM was calculated from 24 hour Holter-
ECG recordings that were conducted once every four weeks in each
participant. Ambient exposure to particulate air pollution was measured at
a central site using an aerosol spectrometer covering the size range from
10 nm to 2.5 m. Elemental (EC) and organic carbon (OC) were
determined hourly from an ambient carbon monitor. Random effect linear
regression was used to model the association between individual 24-h
mean concentrations of particle mass and number concentrations and
MDM in a total of 254 recordings. Effects were adjusted for trend,
weekday, temperature, and relative humidity. Preliminary results based on% change of mean MDM show a significant decrease in MDM in
association with fine (PM2.5, 2.5 m) and ultrafine particles (UFP,
0.1 m) as well as EC and OC. MDM decreased by 14.6% (95% CI:
6.2%, 23.0%) and 12.2% (95% CI: 4.2%, 20.1%) in association
with concurrent concentrations of EC and OC per increase in interquartile
range pollutant. In contrast to EC and OC, PM2.5 and number
concentrations of UFP showed larger effects on MDM with the
cumulative mean of particle concentrations measured from the end of the
recording up to 4 days prior to the recording (14.6% (95% CI:
5.0%,24.3%) and 12.3% (95% CI: 0.3%, 24.9%)). Previous
analyses on conventional HRV parameters such as SDNN or RMSSD in
the same study population observed smaller and less significant effects of
particles in particular with respect to ultrafine particles. Most recently a
decrease in MDM has been identified as a more powerful predictor of
mortality after myocardial infarction than predictors such as HRV or left
ventricular ejection fraction. The results indicate that MDM might be a
more sensitive parameter to assess adverse effects of particulate air
pollution on cardiac function than conventionally used parameters to
assess changes in heart rate fluctuations.
We would like to acknowledge the EPA Particulate Matter Center at the
University of Rochester, which is funded by the Environmental Protection
Agency (US-EPA STAR (“Science To Achieve Results”) Center Grant R-
827354) for funding the study.
ISEE-21
AMBIENT AIR POLLUTION AND ARRHYTHMIC EVENTS IN
PATIENTS WITH IMPLANTED CARDIOVERTER DEFIBRILLATORS, 1993-2002
Kristina B. Metzger,* Paige E. Tolbert,* Mitch Klein,* Jennifer L. Peel,*
W. Dana Flanders,* James A. Mulholland †. *Rollins School of Public
Health, Emory University, Atlanta, GA; † Georgia Institute of Technology,
Atlanta, GA
Abstract: To better understand specific cardiac responses to air pollution
and roles of various pollutants, the relationship of ambient pollutant levels
and tachyarrhythmic events was studied in patients with implanted
cardioverter defi brillators (ICDs). ICDs are electronic devices implanted in
patients at high risk for sudden cardiac death, ventricular tachycardia, and
ventricular fi brillation. The device continuously monitors the heart rate for
tachyarrhythmias, emits electrical pulses or shocks to convert the heart
back to normal sinus rhythm as needed, and records and saves data on
Abstracts Epidemiology • Volume 15, Number 4, July 2004
© 2004 Lippincott Williams & WilkinsS22
8/19/2019 Air Pollution and Acute Cardio Respiratory Visits.50
9/15
each tachyarrhythmic event, including the date and time of the event and
the type of therapy delivered. The medical records of ICD patients at
three electrophysiology clinics in a major U.S. city were examined to
collect patient and event information for the study period, January 1,
1993, to December 31, 2002. We studied the occurrence of
tachyarrhythmic events in these patients in relation to speciated PM data
from August 1, 1998, to December 31, 2002, and in relation to other air
quality data for the period 1993 through 2002. In preliminary analyses
through 2000, there were 570 patients, 335 of whom experienced at least
one tachyarrhythmic event, contributing 3677 event-days in almost
300,000 total follow-up days. We used repeated-measures unconditional
logistic regression, controlling for temporal trends, meteorologic
conditions, and the occurrence of recent events. For this period, a
statistically significant association was observed between tachyarrhythmic
events and ambient levels of carbon monoxide, coarse particles, and
PM2.5 organic carbon on the same day. Abstracting of event data for the
period through December 2002 is nearly complete. We estimate that there
will be 855 patients for the full study period, with approximately 5500
event-days. Results through 2002 will be presented.
ISEE-22
AIR POLLUTION AND INFLAMMATORY MARKERS IN BLOOD
Sara D. Dubowsky,* Diane R. Gold,† Joel Schwartz,* Brent Coull,*
Helen Suh*. *Harvard School of Public Health; † Channing Laboratory,
Harvard Medical School
Introduction: The adverse cardiovascular health effects of air pollution
may be caused in part by systemic inflammation triggered by oxidative
stress to the lungs. Recent studies have suggested that these associations
may be strongest for traf fic-related particles. In order to examine the
associations between traf fic-related particles and systemic inflammation,
we conducted a study in which a panel of senior citizens was exposed to
elevated traf fic-related pollutant exposures through a series of bus trips.
Methods: Repeated blood samples were collected from forty-four non-smoking participants living at one of four suburban independent senior
living facilities. Blood samples were obtained between April and June of
2002, approximately 18-hours after study subjects participated in a field
trip that included two hours aboard a diesel bus. Each subject participated
in approximately four field trips, about once every 3-weeks. Blood
samples were analyzed for red and white cell parameters using a CBC
with differential analyses. Samples were also tested for interleukin-6, c-
reactive protein, intracellular adhesion molecule – 1, fi brinogen, and von
willebrand ’s factor. Particulate exposures of the participants were assessed
using two mobile carts that followed the subjects throughout each field
trip. Micro-environmental exposure measurements included continuous
readings of fine particle mass (PM2.5), ultra-fine, fine, and coarse particle
counts, and black carbon. Carbon monoxide and nitric oxide
concentrations were also measured continuously during the bus rides.
These exposure measures were supplemented with ambient air concentrations of gases, particle mass and number, and meteorological
parameters measured at a local USEPA-funded Supersite.
Results: Among the 138 blood samples analyzed, levels of circulating
inflammatory markers were found to be comparable to the values reported
by previous studies. Evaluation of the air monitoring data indicates that
the bus trips successfully elevated hourly exposures above levels observed
at the residence facilities and central site. The mean hourly black carbon
concentration during the bus trip periods was 3,400 (SD: 1,200) ng/m3 as
compared to mean hourly concentrations of 350 (SD: 450) ng/m3 during
facility periods and 750 (SD: 370) ng/m3 at the Supersite during times
concurrent with the bus trip. Measures of PM2.5 and particle counts were
also substantially elevated during periods when subjects were aboard the
bus. The impacts of these micro-environmental changes on circulating
inflammatory markers will be presented using the results of longitudinal
mixed models.
This work is supported by funding from the NIEHS (1P01E-ES09825-01),
USEPA (R827353-01-0), and EPRI (W09207-02).
ISEE-23
SHORT-TERM COMMUNITY AIR POLLUTION EXPOSURES
RELATED TO LEVELS OF CLOTTING FACTOR VIII INDICATIVE
OF ALTERED HEMOSTASIS IN AN ELDERLY COHORT
Helene Margolis,* Paul Enright,† Dane Westerdahl,* John Robbins‡.
*California Air Resources Board; † University of Arizona; ‡University of
California, Davis School of Medicine
Introduction: The biological mechanisms that underlie the reported air
pollution-related excess risk of cardiovascular morbidity/mortality are
unknown; hypotheses include oxidative stress, altered hemostasis and/or
changes in autonomic function. The National Heart, Lung, Blood
Institute’s Cardiovascular Health Study (CHS) is a population-based
longitudinal study of coronary heart disease and stroke among older
women and men. We initiated a CHS Ancillary Study in 1999 to evaluatethe subclinical and clinical effects of air pollution.
Objective: Evaluate relationship between short-term exposures to O3,
NO2, CO, SO2, PM10, PM2.5, PM10 – 2.5 and clotting factor VII (FVII) and
VIII (FVIII) levels.
Methods: Continuous enrollment of the main cohort between June 1989
and May 1990, and temporal variation in day-to-day pollutant levels
allows evaluation of short-term exposure effects. Baseline data were
obtained from the CHS for participants in Allegheny County, PA
(n1275), Forsyth County, NC (n1305), and Sacramento County, CA
(n1318). Air quality and meteorological data were obtained from the
U.S. EPA, CA Air Resources Board or U.S. National Climatic Data
Center. Various strategies were used to estimate exposures on days with
missing data. Gender-specific single-pollutant linear regression models
were used to evaluate effects of 24-hour average pollutant concentrations
averaged over exposure periods (EAP) of 2- days (day of exam and
previous day (L01)) or 7-days (day of exam and previous 6 days). Natural
log transformations of FVII and FVIII were required to meet model
assumptions. Final models were adjusted for age-at-exam, Community,
24-hour average temperature (L01), and (in FVII model only) 24-hour
average dew point (L01).
Results: Across the range of pollutant exposures experienced by the
cohort over the 1-year baseline period, small significant effects on
population mean levels of FVIII, but not FVII, were observed. A
10-g/m3
higher level of PM10 (7-day EAP) was associated with higher
FVIII levels among women (1.3% (95% CI: 0.12%, 2.6%)) and men
(1.5% (95% CI: 0.14%, 2.8%)). PM2.5 was also positively associated with
FVIII (women: 2.1% (95% CI: 0.6%, 3.7%); men 2.1% (95% CI: 0.5%,
3.8%)). Perhaps reflecting different mechanism(s)/kinetics of response,
among women, mean FVIII levels were inversely related to the 2-dayEAP for O3 (2.3% (95% CI: 4.3%, 0.23%) and PM10 – 2.5 (2.3%
(95% CI: 4.6%, 0.01%). PM10 – 2.5 (2d-EAP) was also inversely related
to FVIII among men (2.1% (95% CI: 4.9%, 0.73%). These results will
be discussed in the context of specific mechanistic hypotheses.
ISEE-24
ASSOCIATION OF SUBCLINICAL ATHEROSCLEROSIS (CAROTID
INTIMA MEDIA THICKNESS) WITH RESIDENTIAL AMBIENT
PM2.5 IN HEALTHY ADULTS
Nino Kuenzli,* Mike Jerrett,† Bernie Beckerman,† Wendy Mack,†
Frank Gilliland,† Duncan Thomas,† John Peters†. *University of
Southern California; † USC
Epidemiology • Volume 15, Number 4, July 2004 Abstracts
© 2004 Lippincott Williams & Wilkins S23
8/19/2019 Air Pollution and Acute Cardio Respiratory Visits.50
10/15
Abstract: Long-term exposure to fine ambient particulate matter (PM2.5)
is associated with mortality due to cardiovascular disease (CVD) in
prospective cohort studies. The largest study (1) observed a 12% (8 – 15%)
increase in CVD death per 10 g/m3 PM2.5. The pathophysiologic
mechanisms that might underlie these associations are currently unknown,
but limited experimental evidence suggests that inhalation of PM leads to
pulmonary and systemic inflammatory responses. In a rabbit model, PM
inhalation was associated with the progression of atherosclerotic lesions.
Moreover, carotid artery intima-media thickness (IMT) in humans steadily
increases from childhood to death, and correlates well with all of the
major cardiovascular risk factors (smoking, LDL, blood pressure etc.), and
with coronary artery atherosclerosis. Studies also show a 5% higher
IMT to be associated with 5 – 10% higher rates of CVD events. We
hypothesize that exposure to ambient PM leads to increases in IMT and
may contribute thereby to increased mortality from CVD.
We used the baseline IMT measurements from two trials conducted in
healthy adults to investigate effects of Vitamin E and B on IMT. Between
1996 and 2003, carotid IMT was measured among 445 men and 356
women (age 40 – 89; mean: 59 yrs). Participants live across Southern
California. Residential zip codes were geo-coded to assign the zip code
centroid levels (1999) of ambient PM2.5, derived from a predictive PM2.5surface (geostatistical kriging model; assigned exposures were mostly
within / 3 g/m3 of monitored values). We determined the cross-
sectional association of log-IMT with the individually assigned residential
PM2.5, with/without adjustment (age, sex, blood pressure, LDL, smoking,
estrogen replacement, education, prescriptions).
The average (SD) IMT was 0.755 mm (0.147). Individually assigned
ambient PM2.5 estimates ranged from 5.2 to 26.9 g/m3 (mean: 20.3).
For each 10 g/m3 increase in PM2.5, IMT increased by 5.6% (95% CI:
0.7 – 10.6%) unadjusted and by 3.8 to 4.6% in various multivariate
models (p-values 0.02 – 0.07).
This cross-sectional assessment, which considers only recent exposure
provides a first test of our hypothesis. These findings require corroboration
in larger samples, taking further potential confounders and effect modi fiers
into account. Information on long-term, time-varying exposure to PM onvarious spatial scales (region, neighborhood, residence, work place,
commute) and from various sources needs to be developed. Most
importantly, the association of air pollution and IMT progression also
needs to be investigated using longitudinal data.
(1) Pope et al., Circulation 2004; 109:71 – 9.
ISEE-25
DO SHORT-TERM INCREASES IN AMBIENT AIR POLLUTANTS
TRIGGER ARRHYTHMIA? – A POPULATION-BASED STUDY
Yinkang Duan,* Duanping Liao,* Gerardo Heiss,† Hung-mo Lin,*
Zhi-jie Zheng,‡ Megan Darnell,* Xuejuan Jin*. *Penn State University
College of Medicine; † University of North Carolina at Chapel Hill;
‡Center for Disease Control and PreventionAbstract: An association between air pollution and cardiovascular disease
(CVD) mortality has been reported. One of the hypothesized
arrhythmogenic mechanisms is the impairment of cardiac autonomic
control by pollutants. We examined the short-term associations of ambient
pollutants (particulate matter 10 m in diameter PM10, O3, CO, NO2,
and SO2) with cardiac arrhythmias using data from the baseline
examination (1987 – 1989) of the population-based Atherosclerosis Risk in
Communities study. We calculated the following pollutant exposures one-
day prior to each participant’s randomly assigned examination date: PM10,
CO, NO2, and SO2 as 24-hour averages, and O3 as an 8-hour average of
the hourly measures, from the EPA’s Aerometric Information Retrieval
System (AIRS). We assessed the presence, frequency, and complexity of
arrhythmias from standardized 2-minutes EKG rhythm strips recorded
during the baseline clinical examination. After excluding missing data on
the ambient exposure measurements, the effective sample sizes for PM10,
SO2
, NO2
, CO, and O3
) were 4,001, 7,404, 7,427, 10,699, and 9,059
respectively. The prevalence of arrhythmias was 11%. Among persons
without any arrhythmia, the means (SD) of PM10
, SO2
, NO2
, CO, and O3
measured as one day prior their ECG and clinical examination were 29.5
(12.3) g/m3, 0.005 (0.004) ppm, 0.018 (0.008) ppm, 1.35 (0.60) ppm,
and 0.040 (0.017) ppm, respectively. For persons with arrhythmias, the
means (SD) were 30.0 (13.0) g/m3, 0.005 (0.003) ppm, 0.017 (0.008)
ppm, 1.33 (0.59) ppm, and 0.040 (0.017) ppm, respectively. Logistic
regression models were used to adjust for CVD risk factors, demographic
and socioeconomic variables, and relevant meteorological variables to
obtain the odds ratios (OR) and 95% confidence intervals (CI) of the
presence of arrhythmias associated with one standard deviation increase of
each ambient pollutant: the OR for PM10
, SO2
, NO2
, CO, and O3
were
1.00 (0.99, 1.01), 0.94 (0.86, 1.04), 0.98 (0.90, 1.06), 0.94 (0.84, 1.05),
and 1.00 (0.91, 1.11), respectively. The frequency and complexity of
arrhythmias were not associated with any of the pollutants analyzed. In
conclusion, no statistically significant association between short-term
increase in ambient pollutants and the presence of arrhythmias was
observed in this in healthy, middle aged population.
ISEE-26
THE ASSOCIATION BETWEEN AMBIENT PM2.5 AND
BIOMARKERS OF AIRWAY INFLAMMATION IN PATIENTS
WITH ASTHMA
Samantha De Leon,* Kazuhiko Ito,* Hsien-Wen Hsu,† Joan Reibman,†
George Thurston*. *NYU School of Medicine, Nelson Institute of
Environmental Medicine; † NYU School of Medicine, Division of
Pulmonary & Critical Care Medicine
Abstract: Many studies have shown that individuals with asthma are
adversely affected by short-term increases in ambient Particulate Matter
(PM) levels in terms of decreased lung function, increased occurrence of respiratory symptoms (such as wheezing and coughing), and increased
hospital admissions and mortality. However, not as many studies have
considered why patients with asthma are sensitive to increases in ambient
pollution levels. Airway inflammation is one of the characteristic
symptoms of chronic asthma and asthma severity; therefore, levels of
chemokines associated with the recruitment of leukocytes, such as
eosinophils, into the lung airways were used to investigate the health
effects of ambient PM on those with asthma. Serum levels of several
chemokines (Regulated on Activation, Normal T-cells Expressed and
Secreted RANTES, Eotaxin, Thymus and Activation-Regulated
Chemokine TARC, and Interferon gamma Inducible Protein of 10 KDa
IP-10) were measured every 2 weeks, over a 3-month period in adult
subjects with asthma residing in New York City during the summer of
2001. PM2.5 (PM less than or equal to 2.5 um) data was collected over
the same 3-month period, and daily weather data were obtained from theweather station located at New York Kennedy International Airport. All
analyses were conducted using linear mixed-effects models assuming
random intercepts and slopes for each subject. Biomarker levels were
included in the regression model as raw variables (pg/ml). Of the four
chemokines examined, only RANTES showed an association with PM2.5.
There was a general positive trend between the same day PM2.5 and
RANTES levels, with one outlier that weakened the overall positive slope.
The effect of PM2.5 on RANTES levels was estimated in the mixed-
effects models adjusting for hot and humid days, day-of-week effects,
hours spent in air conditioning, number of puffs of albuterol, and serial
correlation between observations for each individual. After removal of one
outlier, RANTES levels (raw variable) were found to be increased by
11321 pg/ml (z-statistic 2.34) per Inter-Quartile Range (IQR) of PM2.5
Abstracts Epidemiology • Volume 15, Number 4, July 2004
© 2004 Lippincott Williams & WilkinsS24
8/19/2019 Air Pollution and Acute Cardio Respiratory Visits.50
11/15
(14.36 ug/m3), which corresponds to a 17% increase based on the study
average of 66640 pg/ml. The results of this study suggest that chemokines
involved in airway inflammation in asthmatics, such as RANTES, can
increase in response to short-term elevations in ambient PM2.5 levels.
While controlled exposure experiments are needed to confirm these
results, they may provide insight into the mechanism(s) underlying asthma
exacerbations associated with acute episodes of increased ambient PM
levels.
ISEE-27
AIR POLLUTION AND ACUTE CARDIO-RESPIRATORY VISITS IN
AN AMBULATORY CARE SETTING: TWO YEAR AND
PRELIMINARY FOUR YEAR RESULTS
Amber H. Sinclair, Dennis Tolsma. Kaiser Permanente
Introduction: There is a lack of published research concerning the
possible associations between air pollution components and acute visits to
outpatient clinics. We investigated associations between various
components of air pollution and acute visits to clinics of a health plan in a
large urban area.Methods: We used a time series analysis to determine the associations
between daily levels of suspended particulate matter and ambulatory care
acute visit rates during the 25-month period 8/1/98 to 8/31/00 and for the
four year period of 8/1/98 to 12/31/02. Acute visits of health plan
members to clinics with a respiratory diagnosis of asthma, COPD, upper
and lower respiratory infections (URI and LRI) and cardiovascular disease
were identified through electronic patient visit data. Pollutant data
included 24-hour measurements of PM2.5, coarse PM (2.5 – 10 um), PM10,
PM2.5 components (acidity, sulfates, OC, water-soluble transition metals
and elemental carbon), 10 – 100 nm PM area (ultra-fines), polar VOCs
(OHC); 8 hour maximum ozone; and 1 hour maximum NO2, CO and
SO2. Visit counts for each diagnosis group were modeled by air quality
metrics using general linear modeling, controlling for temporal trends and
meteorologic variables. Moving averages of the a priori 0 to 2 day lagged
air quality variables were investigated, as well as the 3 to 5 day average.
Results: For the 25-month data set, we found significant positive
associations for the 0 to 2 day lag for child asthma with OHC, URI with
ultrafine PM, and LRI with PM2.5 acidity and SO2. In comparison, we
found 14 significant positive associations for the 3 to 5 day lag: adult
asthma with ultrafine PM; child asthma with coarse PM, PM10, EC and
OC; URI with coarse PM; LRI with coarse PM, PM10, EC, OC and
PM2.5; and cardiovascular disease with NO2, CO and O3. There were
also a few significant negative associations. The magnitudes of the
significant risk ratios were less than 1.15. In addition to these findings,
preliminary results of the analysis of a four-year time-period of air quality
data will be presented.
Discussion: LRI and child asthma had the greatest number of significant
associations with air pollution, mainly for the 3 to 5 day lag structure, in
the 25-month data set. This study provides a unique evaluation of air quality and health effects by investigating the relationships of air pollution
to uncommonly reported but readily measurable health effects.
This research was supported by a grant from the Electric Power Research
Institute (EPRI).
ISEE-28
ASSOCIATION BETWEEN FINE PARTICLES AND RESPIRATORY/
CARDIOVASCULAR ELDERLY HOSPITAL ADMISISONS IN NEW
YORK CITY
Kazuhiko Ito, Samantha De Leon, George Thurston. Nelson Institute of
Environmental Medicine, NYU School of Medicine
Abstract: Many studies reported associations between particlate mater
(PM)/gaseous pollutants and morbidity and mortality, but a question
remains as to what emission source types are responsible for the observed
associations. We examined associations between PM2.5/gaseous pollutants
(NO2, CO, SO2, and O3) and the cardiovascular and respiratory elderly
hospital admissions in NYC for period 1999 – 2001. The relationship
among PM2.5, gaseous pollutants, and weather variables were first
examined using cross-correlation functions (CCF) across seasons after
removing long-term and seasonal cycles. The associations between air
pollution and the elderly hospital admissions were examined in four
different ways: (1) single pollutant model; (2) two-pollutant model with or
without interaction terms; (3) single pollutant model stratified with the
level of a co-pollutant; and (4) model with factor-analysis derived
composite pollution indices. Poisson Generalized Linear Model was used
to estimate pollution effects, adjusting for temporal trends, day-of-week,
and temperature (quintile indicators with lags). The temporal relationships
between PM2.5 and NO2, SO2, or CO were relatively unchanged across
seasons with moderate to high (0.5 to 0.8) correlation. The short-term
correlation between O3 and PM2.5 was positive in warm season but
negative in cold season, apparently reflecting the association between cold
sunny clear day and O3 in winter. In single pollutant models, pneumoniaadmissions were positively significantly associated with PM2.5 and CO
(percent excess 8% per 5th-to-95th percentile pollution increment for
both pollutants); COPD admissions were associated with CO (8%);
ischemic heart disease admissions were associated with NO2 and CO (6%
for both pollutants); and, heart failure admissions were associated with
PM2.5, NO2, and CO (9%, 9%, and 13%, respectively). Two pollutant
models generally resulted in attenuation of the risk estimates for one of
the pollutants while the other remaining significant without improving
model fi t. In stratified analysis, PM2.5 risk estimates were larger for high
CO days than for low CO days for both pneumonia and heart failure
admissions. The factor-analysis derived components that represented the
common variation of PM2.5, NO2, and CO were not as good predictors of
these health outcomes as the individual pollutants. Overall, the pollution
mixture that includes PM2.5, NO2, and CO, likely traf fic-related air
pollution, was associated with cardiovascular and respiratory elderly
hospital admissions in NYC.
This research was supported by U.S. EPA STAR grant (R827997010);
U.S. EPA Particulate Matter Health Research Center (R827351); and
NIEHS Environmental Health Center (ES00260).
ISEE-29
RELATIONSHIP BETWEEN GASEOUS AIR POLLUTANTS AND
CARDIOVASCULAR ADMISSIONS: A STUDY IN 14 SPANISH
CITIES
Ferran Ballester,* Paz Rodr ı́guez,*† Santiago Pérez-Hoyos,*
Juan Bellido,‡ Federico Arribas,§ Carme Saurina, ¶ Inmaculada Aguilera,
Ana Breznes,** Margarita Taracido,†† Alvaro Cañada,‡‡Carmen Iñiguez* (EMECAS Group). *Valencian School of Studies for
Health-EVES; † Miguel Herná ndez University; ‡ Departamento de Salud. C
Valenciana; § Departamento de Salud. Aragó n; ¶ Research Group on
Statistics, Applied Economics and Health (GRECS); Escuela Andaluza de
Salud Pú blica; **Departamento de Salud. C de Madrid; †† Universidad
de Santiago; ‡‡ Departamento de Salud. Asturias
Introduction: Since year 1997, the EMECAS project evaluates the effect
of current air pollution on the health of urban population in Spain. In this
paper we present the results for the association between gaseous air
pollution and hospital admissions for cardiovascular diseases.
Methods: Daily emergency admissions for all cardiovascular diseases
(CVS) (ICD-9: 390 – 459), heart diseases (HD) (ICD-9: 410 – 414, 427,
428); ischemic heart diseases (IHD) (ICD-9: 410 – 414), and
Epidemiology • Volume 15, Number 4, July 2004 Abstracts
© 2004 Lippincott Williams & Wilkins S25
8/19/2019 Air Pollution and Acute Cardio Respiratory Visits.50
12/15
cerebrovascular diseases (Stroke) (ICD-9: 430 – 438) were collected from
hospital records (1995 – 1999). Air pollutants data were obtained from local
networks. Variables for 24 hours daily levels of SO2, and NO2; 8 hours
maximum moving average of CO and ozone; and, lastly, 1 hour maximum
of SO2, NO2 and ozone were constructed. Magnitude of association in
each city was estimated using generalized additive models (GAM) of
Poisson controlling for confusion and overdispersion. Lagged effects, up
to three days, for each cause and pollutant were examined. Data were
analysed using S-Plus GAM function with more stringent convergence
criteria. Combined estimates were obtained under a fixed effects’ model,
and, if heterogeneity, under ’random effects’ one. For ozone the analyses
were restricted to the warm period (May to October).
Results: Concurrent and one day were the most consistent lags showing
associations, only ozone showed a more delayed relationship. In the
combined estimates an increase of 10 ug/m3 in SO2 daily levels was
associated with a 1.3% (95% CI: 0.2 – 2.5%) increase in the number of
hospital admissions for CVS, and 1.7% (95% CI: 0.7 – 2.9%) for HD. The
same increase in concentrations of NO2 was significantly associated with
a 0.4% increase in CVS, and 0.9% in HD admissions. For ozone the
corresponding estimates were 0.7 in both cases. An increase in 1 mg/m3
levels of CO was associated with an increase of 2.1% in CVS, and 4.2%in HD admissions. Only ozone presented some suggestion of association
with stroke. In two pollutant models the coef ficients of SO2 and NO2
were affected after control for other pollutants. On the contrary, CO and
ozone were unaffected by the inclusion of the second pollutant.
Conclusions: A short-term association between increases in daily levels
of air pollutants and the number of daily admissions for cardiovascular
diseases in the Spanish cities was found with greater estimates for heart
diseases. CO and ozone showed an independent effect of the other
pollutants.
EMECAS project has been supported by the Spanish Ministry of Health,
Fondo de Investigaciones Sanitarias (FIS 97/0051 and FIS 00/0010)
ISEE-30
URBAN PARTICULATE AIR POLLUTION AND
CARDIOVASCULAR ADMISSIONS IN SPAIN
Carmen Iniguez,* Paz Rodriguez,*† Marc Saez,‡ Antonio Daponte,§
Elena Lopez, ¶ Teresa Martı́nez, Luis Cirera,** Laura Ló pez,††Inez Aguinaga,‡ Ferran Ballester* (EMECAS Group). *Valencian School
of Studies for Health-EVES; † University Miguel Herná ndez; § Research
Group on Statistics, Applied Economics and Health (GRECS); ‡ Escuela
Andaluza de Salud Pú blica; § Departamento de Salud. Canarias;
¶ Departamento de Salud. Gobierno Vasco; Departamento de Salud.
Murcia; **Departamento de Salud. Madrid; †† Dep Salud. Ayuntamiento
de Pamplona
Introduction: The EMECAS project aims to evaluate the short-term
effects of air pollution on mortality and hospital admissions for
cardiovascular and respiratory diseases in Spain. We present the results of the association between daily levels of particulate matter and the daily
number of cardiovascular admissions.
Methods: Data from fourteen cities accounting overall for 10 million
inhabitants were analysed. The number of daily emergency admissions for
all cardiovascular diseases (CVS) (ICD-9: 390 – 459), heart diseases (HD)
(ICD-9: 410 – 414, 427, 428); ischemic heart diseases (IHD) (ICD-9:
410 – 414), and cerebrovascular diseases (Stroke) (ICD-9: 430 – 438) was
obtained from hospital records. From local Air Pollution Networks we
collected data for the available particulate indicators, i.e., black smoke
(BS), total suspended particles (TSP), and PM10, and constructed variables
for 24 hours mean levels. The period of the study was 1995 to 1999.
Magnitude of association in each city was estimated using generalized
additive models (GAM) of Poisson controlling for confusion and
overdispersion. For each cause lagged effects, up to three days, of each
pollutant were examined. Data were analysed using S-Plus GAM function
with stringent convergence criteria. Combined estimates were obtained
under a fixed effects’ model, and, if heterogeneity, under ’random
effects’ ones. Lastly, to explore potential confounding for the other
pollutants two-pollutant models were performed.
Results: Levels of PM10 ranked from 32 to 43 g/m3, BS from 18 to 80
g/m3; and TSP from 52 to 76 g/m3. Local estimates were, mostly,
positive and more consistent in lags 0 (concurrent day) and 1. Combined
estimates, expressed as the percent change in risk (and 95% confidence
interval) for an increase of 10 g/m3 particulate levels of lag 01 are
showed in the table below. In two pollutant models PM10 estimates were
not modified but BS and TSP were.
Particulate
indicator
Cities
(n) CVS HD IHD Stroke
TSP 7 0.07
(0.23; 0.36)
0.45
(0.04; 0.86)
0.29
(0.28; 0.86)
0.42
(1.03, 0.20)
Black
smoke
6 0.24
(0.18; 0.67)
0.71
(0.13; 1.29)
0.12
(0.69; 0.94)
0.49
(1.37; 0.39)PM10 5 0.91
(0.35; 1.47)
1.56
(0.82; 2.31)
1.58
(0.