Full text of DOI:10.1111/j.1467-842X.2002.tb00266.x

Article
The Air We Breathe
Particulate air pollution and hospital admissions
in Christchurch, New Zealand
J.A. McGowan
Abstract
Department of Mathematics and Statistics, University of Canterbury,
Christchurch, New Zealand
Aims: Winter air pollution in Christchurch is
dominated by particulate matter from solid
fuel domestic heating. The aim of the study
was to explore the relationship between
particulate air pollution and admissions to
hospital with cardio-respiratory illnesses.
Methods: Particulate air pollution statistics
(PM10) were obtained from the Canterbury
Regional Council monitoring station in the
city. The New Zealand Health Information
Service provided data on admissions to the
Princess Margaret and Christchurch
P.N. Hider
New Zealand Health Technology Assessment, Christchurch School of Medicine and
Health Sciences, University of Otago, Christchurch, New Zealand
E. Chacko
Department of Mathematics and Statistics, University of Canterbury,
Christchurch, New Zealand
G.I.Town
Hospitals for the period June 1988 through
December 1998 for both adults and
Department of Medicine, Christchurch School of Medicine and Health Sciences,
University of Otago, Christchurch, New Zealand
children with cardiac and respiratory
disorders. The relationship between PM10
and admissions was explored using a time
series analysis approach controlling for
weather variables. Missing values were
interpolated from carbon monoxide data for
the same time period, as carbon monoxide
and PM10 were highly correlated.
t has been convincingly demonstrated in below which ambient particulate can be
a number of European and North deemed to be free of health effects.5,8
I
American cities that high levels of air
To date, there has been relatively little New
pollution are associated with short-term Zealand-based research examining the rela-
adverse effects on human health.1-3 The tionship between ambient air pollution and
adverse effects have largely been associated either the morbidity or mortality associated
with exposure to elevated levels of the with respiratory or cardiovascular illnesses.9
particulate component of air pollution4 and This lack of research is surprising as parts
specifically the small sub-10 micron-sized of New Zealand have clearly been docu-
Results: There was a significant
association between PM10 levels and
cardio-respiratory admissions. For all age
groups combined there was a 3.37%
increase in respiratory admissions for each
interquartile rise in PM10 (interquartile value
14.8 mcg/m3). There was also a 1.26% rise
in cardiac admissions for each interquartile
rise in PM10. There was no relationship
between PM10 and admissions for
particles (PM10).5
mented to have a significant air pollution
The documented adverse health effects problem, most notably the Christchurch
from PM10 exposure include associated rises urban area, with ambient levels of particulate
in mortality rates, hospital admissions, emer- matter that sometimes surpass those of
gency department visits, as well as increases major North American or European cities10
in symptoms and medication use.6 Adverse and regularly exceed the local guideline of
effects have also been demonstrated on lung 50 mcg/m3 as a maximum for the mean daily
function.6 The effects of exposure to high particulate level.11 In addition, the prevalence
concentrations of inhaled pollutants have of respiratory disorders, especially asthma,
largely impacted on respiratory and cardio- and cardiovascular illnesses are high in this
vascular systems7 and have recently been country.12,13
appendicitis, the control condition selected.
Conclusions: In keeping with overseas
studies, there is evidence in Christchurch of
a relationship between ambient particulate
levels and admissions with cardiac and
respiratory illnesses. The size of the effect is
consistent with overseas data, with the
greatest impact for respiratory disorders.
Implications: These results indicate that
measures to control ambient particulate
levels have the potential to reduce hospital
admissions for cardio-respiratory illnesses.
(Aust N Z J Public Health 2002; 26: 23-9)
related to exposure to even relatively mod-
Early studies that have been conducted in
erate levels of PM10 consistently below 100 New Zealand had significant methodologi-
mcg/m3.2 Cumulative evidence from a cal weaknesses, such as a reliance on ex-
number of mortality and morbidity studies trapolated data from overseas work14 or low
now suggests that these effects are linear and power resulting from a small sample size.15
that there is no threshold concentration Recent studies have employed more robust
Correspondence to:
Submitted: July 2001
Dr P. Hider, NZHTA, Department of Public Health and General Practice, Christchurch School of
Medicine and Health Sciences, PO Box 4345, University of Otago, Christchurch, New Zealand.
Tel/fax: +64 3 364 1152; e-mail: phil.hider@chmeds.ac.nz
Revision requested: October 2001
Accepted: December 2001
2002 VOL. 26 NO. 1
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
23

McGowan et al.
Article
methods and have found significant increases in respiratory symp-
toms and medication usage associated with relatively modest rises
in ambient pollution among a susceptible group of subjects with
chronic obstructive pulmonary disease16 and an increase in mor-
tality in relation to a rise in particulate level using a time series
study design that controlled for weather variables.17
Air pollution and meteorology data
Pollution levels were obtained from the Canterbury Regional
Council monitoring site in a central Christchurch site for the
period June 1988-December 1998. The site is owned by the Min-
istry of Health and operated and maintained by the Institute of
Environmental and Scientific Research Ltd. The information was
provided by the Canterbury Regional Council in the form of 24-
hour mean concentrations of the following pollutants: carbon
monoxide (CO), particulate (PM10), sulphur dioxide (SO2), and
oxides of nitrogen (NOx).
Christchurch is an ideal site for an investigation of the relationship
between outdoor air quality and respiratory/cardiovascular admis-
sions: it has one (previously two) main hospital serving the entire
city catchment with a computerised patient information system and
there is a well-established system for gathering pollution data.
Uniquely in Christchurch, more than 90% of particulate air
pollution has been estimated to come from the city’s 47,000 wood
burners and open fires that are used during the cold winters
months.18 Overseas studies have found that the smoke generated
by these sources is associated with the most deleterious effects
upon the respiratory and cardiovascular health of local inhabit-
ants.19 Furthermore, the local weather patterns during the winter
months compound the pollution problem. The development of an
inversion layer traps warmer air along with large amounts of
ambient pollutants at low altitude above the city.18 Local studies
suggest that relatively even mixing of the ambient particulate
matter occurs across the Christchurch air-shed over a 24-hour
period, implying that there is likely to be a generally homogene-
ous exposure to similar PM10 levels for all city residents.14
Finally, reliable information on important potential confounders,
largely meteorological variables,1,20 are readily available.
The aim of this study was to assess the relationship between
the number of people admitted to Christchurch hospitals each
day with respiratory or cardiovascular illness and particulate pol-
lution during the period 1988-98.
Over this time around 20% of days were missing particulate
pollution data, mostly during the summer of 1993-94 due to equip-
ment malfunction, with the remainder occuring at random. High
correlations existed between PM10 and the other pollutants, espe-
cially CO (the correlation coefficient between PM10 and CO was
0.84). A regression was performed on days when values for both
PM10 and CO existed and the relationship between the pollutants
then enabled the determination of missing PM10 data. A similar
approach was used with the other pollutants (NOx and SO2) to fill
in remaining gaps. The summer PM10 levels were in general low,
around one-quarter of the winter PM10 levels, and the readings
were also virtually constant. As a result any errors in the imputa-
tion of the summer data would be small and have limited impor-
tance in the subsequent regression calculations.
Only the relationship between particulate pollution and hospi-
tal admissions was examined. Meteorological data including wind
speed, relative humidity, and temperatures at one metre and 10
metres above ground level were also obtained from the same site.
PM10 was monitored using the Rupprecht and Patashnick Co.
Inc. Tapered Elemental Oscillating Microbalance (TEOM) device,
SO2 measurements were made by fluorescence (API 100A series,
San Diego, US), CO was measured by the gas filter correlation
method (API 300, San Diego, US). NOx monitoring was by an
API 200A (San Diego, US) and meteorological measurements
were collected by an EASI 950315 (temperature) (Dayton, US),
EASI 940919-02 (humidity) (Dayton, US) and Vector A101M
(wind speed) (Clwyd, UK). All measurements were subject to
regular quality audit using standards adopted by the Ministry of
the Environment.
Method
Study area
Christchurch has a population of 333,000 people and covers
some 452 km2 on the Canterbury Plains, which are bordered by
the Pacific Ocean to the east and the Southern Alps to the west.
Consent for the study was obtained from the two hospitals,
the Canterbury Regional Council and the Canterbury Ethics
Committee.
Hospital admission data
Hospital admission information was obtained from the New
Zealand Health Information Service concerning daily counts of
acute hospitalisations for all age groups with the following pri-
mary diagnoses based on the International Classification of Dis-
eases, 9th Revision, (ICD 9) classifications: pneumonia (ICD
480-487), acute respiratory infections (460-466), chronic lung
diseases (491-492, 494-496), asthma (493), ischaemic heart dis-
ease (410-414), dysrhythmia (427) and heart failure (428). Infor-
mation on appendicitis (540-542) admissions was also included
as a control group. Until 1994 the Princess Margaret Hospital
(five kilometres south of the other hospital located in the city
centre) handled approximately half the medical admissions and
so we aggregated admission data for Christchurch Hospital (1988-
98) and Princess Margaret Hospital (1988-94) into a single dataset.
Statistical methods
The analysis was undertaken in two stages using S-PLUS soft-
ware (http://www.insightful.com/). First, the influence of the
meteorological variables on hospital admissions was determined
and then the association between the residuals and PM10 was
modelled. This approach, as successfully used in similar previ-
ous studies,1,21 is considered conservative, and expected to give
a minimum value for the pollution’s contribution to hospital
admissions.
A generalised additive model with a log link1 was used for the
first stage because, although count values are usually Poisson, the
24
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
2002 VOL. 26 NO. 1

The AirWe Breathe
Particulate air pollution and hospital admissions
mean number of admissions varied with the seasons resulting in an
overall non-Poisson distribution.1 The generalised additive model
is an analytical methodology well suited to this type of research1,21
and one that has been extensively used by researchers.1,22
conditions, an expected percentage increase or decrease in ad-
missions for an interquartile increase of PM10 was calculated ac-
cording to patients in three age groups.
Different lags also needed to be allowed for, as the effects of
particulate exposure may be delayed. From a biological perspec-
tive, previous research suggested that lag periods would not be
important for cardiac illnesses, but that a respiratory admission is
more likely a couple of days after a high pollution episode.6,38 We
confirmed this by investigating the results for different lags (from
zero to six days), and therefore all cardiac admission results have
no lag, and all respiratory results include a two-day lag.
The study was calculated to have sufficient sample size with
18,290 admissions to detect a 1% increase in cardiac admissions
for people aged over 64 years, or respiratory admissions for chil-
dren less than 15 years (20,938 admissions), in relation to a
10 mcg/m3 rise in particulate level with 80% power. These calcu-
lations did not include provision for a threshold, which is con-
sistent with previous research.6,23
This model was used to find the most likely relationship
between weather conditions and admissions, given all the data
from 1988 to 1998. The number of admissions predicted by the
model was then subtracted from the actual admission values to
leave the residual admissions, which were unrelated to the mete-
orological variables and had a normal distribution. The residuals
did not exhibit any significant periodicity and the standard errors
were not biased by dispersion. We found that the model took into
account two-thirds of the variation in admissions data, although
if we had included pollutants in this first stage that value would
be lower. This is why our final results show smaller increases
than is actually the case. The seasonality in the admissions data is
a result of seasonal variations in the other variables and is
accounted for by the model.1
A linear regression model was used for the second stage to com-
pare these newly calculated admission residuals to the PM10 con-
centrations.1 Because PM10 is significantly higher during the night,
admissions were compared on one day with the pollution from the
previous night, calculated from 9am on the previous day to 9am on
the same day. From this regression the expected increase in admis-
sions due to a given increase in PM10 could be calculated.
For each condition, as well as total respiratory and cardiac
Results
Summary statistics are given in Table 1 for the daily measure-
ments of pollutants, meteorological variables, and cardio-respi-
ratory admissions for all age groups in Christchurch between
1988-98. During the study period, the interquartile range for PM10
recordings was 14.8 mcg/m3, while the median daily total number
of respiratory and cardiac admissions were 10 and 7 respectively.
Table 1: Summary statistics for daily measurements of hospital admissions, air pollutants, and meteorological
variables, in Christchurch City (NZ), 1988-98.
Diagnostic category
Mean
SD
Min
Max
IQ range
Cardiac admissions
Ischaemic heart disease (ICD Nos: 410-14)
Heart failure (428)
4.46
1.41
0.96
6.84
2.37
1.26
1.03
3.04
0
0
0
0
18
9
3
2
1
4
Dysrhythmia (427)
9
Total cardiac admissions (410-14, 427-8)
22
Respiratory admissions
Acute respiratory infections (460-6)
Other diseases of upper tract (470-8)
Pneumonia and influenza (480-7)
Chronic airways disease (490-2, 494-6)
Asthma (493)
2.55
0.29
2.33
1.94
3.06
10.17
2.22
0.60
2.15
1.71
2.02
5.53
0
0
0
0
0
0
17
6
3
0
2
2
2
7
17
11
13
38
Total respiratory admissions (460-6, 470-8, 480-7, 490-6)
Appendicitis (540-2)
0.93
0.97
0
6
1
Pollutant concentrations
SO2 (mg/m3)
NO (mg/m3)
7.84
30.08
19.40
32.50
1.16
9.56
59.80
14.69
49.57
1.51
0
0
0
0
0
0
87
709
184
566
15.7
283
8
23
17
26
1
NO2 (mg/m3)
NOX (ppb)
CO (mg/m3)
PM10 (mg/m3)
25.17
25.49
14.8
Meteorological variables
Windspeed (m/s)
Temperature at 1m (oC)
Temperature at 10m (oC)
Relative humidity (%)
2.75
12.54
11.71
75.77
1.08
4.85
0
6.9
1.47
7.3
0.6
28.36
26.76
100.0
4.44
0.9
6.52
18.33
14.63
21.06
2002 VOL. 26 NO. 1
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
25

McGowan et al.
Article
Table 2: Estimated percentage increase (with 95% confidence intervals) in cardiac or respiratory admissions for
different age groups in relation to an interquartile rise in PM10 pollution in Christchurch between 1988-98.
Age 0-14 years
15-64 years
65 and over years
Total for all age groups
Total cardiac admissionsa
Total respiratory admissionsb
Notes:
0.21
1.33
1.22
1.26
(-14.49–4.91)
(-0.24–2.90)
(0.11–2.33)
(0.31–2.21)
3.62
3.39
2.86
3.37
(2.34–4.90)
(1.85–4.93)
(1.23–4.49)
(2.34–4.40)
(a) & (b) No lag for cardiac admission, two-day lag for respiratory admissions.
Over the study period the mean annual number of hospital ad-
missions for respiratory conditions rose (see Figure 1). Weather,
pollution, and hospital admission (for most conditions) data all
exhibited an annual cyclical pattern with peak values generally
being recorded during the winter months (see Figure 1).
A significant association was found between particulate pollu-
tion and admissions across all age groups for both cardiac and
respiratory conditions. An interquartile rise in PM10 was associ-
ated with a 3.37% (95% CI 2.34–4.40) increase in respiratory
admissions and a 1.26% (95% CI 0.31–2.21) increase in cardiac
hospitalisations (see Table 2).
increase in admissions across all age groups for appendicitis, the
control condition, (0.38%, 95% CI -1.87–2.63), or also ischae-
mic heart disease (0.70%, 95% CI -0.44–1.84) and dysrhythmia
(1.08%, 95% CI -1.24–3.40) (see Table 3). However, the study
had limited power to explore these subgroups.
For cardiac conditions the associations were strongest when
there was no lag in the pollution data (see Table 4). However, for
respiratory admissions the increase in admissions was highest
when a two-day lag was included in the analysis.
Discussion
The increase in respiratory admissions was similar across all
age groups, with children exhibiting a slightly higher increase in
hospitalisations.
The main finding of this study, that an interquartile increase of
14.8 mcg/m3 in PM10 is associated with a significant rise in hos-
pital admissions for both total respiratory (by approximately 3%)
and total cardiac (by approximately 1%) conditions, is consistent
with previous research conducted in North American and Euro-
pean cities.1,24-29 Despite the general consistency in the research,
there are difficulties with comparing the results of different stud-
Total respiratory admissions due to infectious causes; pneumo-
nia/influenza (5.32%, 95% CI 3.46–7.18) and acute respiratory
infections (4.53%, 95% CI 2.82–6.24) exhibited the largest in-
crease of all the diagnostic categories in relation to an interquartile
rise in PM10 (see Table 3). By contrast, there was no significant
Figure 1: Average daily level of PM10 and number of respiratory admissions in Christchurch, 1988-98.
PM10
Respiratory
100
90
80
70
60
50
40
30
20
10
0
50
40
30
20
10
0
P
M
1
R
e
s
p
0
Jun-88 Jun-89 Jun-90 Jun-91 Jun-92 Jun-93 Jun-94 Jun-95 Jun-96 Jun-97 Jun-98
Date
26
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
2002 VOL. 26 NO. 1

The AirWe Breathe
Particulate air pollution and hospital admissions
Table 3: Estimated increase in admissions over all age groups for each diagnostic subgroup in relation to an inter-
quartile rise in PM10 between 1988-98 in Christchurch.
Admission group
Diagnostic category
(ICD Classification Nos)
Percentage increase in total admissions for each
diagnostic category (95% CI)
Cardiac (no lag)
Ischaemic heart disease (410-4 )
Heart failure (428)
0.70 (-0.44–1.84)
3.05 (1.16–4.94)
1.08 (-1.24–3.40)
Dysrhythmia (427)
Respiratory (two-day lag)
Appendicitis (no lag)
Acute respiratory infections (460-6)
Pneumonia/influenza (480-7)
Chronic lung diseases (490-2, 494-6)
Asthma (493)
4.53 (2.82–6.24)
5.32 (3.46–7.18)
3.95 (2.15–5.75)
1.86 (0.48–3.24)
0.38 (-1.87–2.63)
Appendicitis (540-20)
ies because various measures of particulate pollution (black smoke,
total suspended particles and PM10) have been used along with
different methods to present results (such as a change in admis-
sions per 10 mcg/m3, 50 mcg/m3, or interquartile change in level
of particulate). The relationships between the various measures
of particulate pollution vary with different pollution sources, and
comparisons between them can only be general.
examining the relationship between particulate exposure and pae-
diatric respiratory admissions must determine how data on respi-
ratory epidemics can be reliably integrated into the analyses.
The results of most time series studies suggest that those at risk
of increased, earlier or more severe hospital usage on higher pol-
lution days are those with pre-existing serious cardio-respiratory
ill-health.6,39 It appears likely that in people whose health is al-
ready compromised by chronic illness (and ageing), the severity
of an acute illness (such as a respiratory infection or congestive
heart failure) may be increased on days with high pollution.6,40
An exception to this mechanism may be asthma, where fine par-
ticles may directly provoke an attack of respiratory illness.5 There
is, therefore, likely to be a range of lag periods between exposure
to the exacerbating effect of pollution and the need for inpatient
treatment related to variability in the severity of the underlying
acute illness as well as variations in access to health care. Differ-
ences in the lagged responses between the two groups of admis-
sions may reflect distinct mechanisms that determine the
pulmonary and cardiac responses after exposure to PM10. Future
research should examine in more detail the mechanisms of cardio-
respiratory deterioration and pathophysiological changes that
occur after this exposure.
There was a significant increase in respiratory admissions in
relation to an interquartile rise in particulate matter for all age
groups, but the rise was highest among children. Most previous
research examining paediatric morbidity and particulate pollu-
tion has focused upon outcomes other than hospital admissions
and they have often used other study designs.30-32 Previous time
series studies have consistently documented a rise in paediatric
hospitalisations in relation to increasing particulate pollution33
and some researchers have concluded that children are particu-
larly sensitive to the adverse effects of particulate matter.34
A significant issue for analyses involving childhood respira-
tory admissions is the possibility that there may be confounding
by epidemics of viral respiratory infections that have been strongly
associated with increases in the number of hospital admissions
for children.35,36 The results from this study support the possibil-
ity that respiratory infections may act as a confounder as the high-
est increase in hospitalisations was associated with the diagnostic
categories related to respiratory infections (pneumonia and acute
respiratory infections). However, the interaction may be more
complicated: particulate exposure may be a significant risk factor
for more frequent viral respiratory infections, it may exacerbate
infectious illnesses, or a viral illness may increase individual sus-
ceptibility to the adverse effects of pollution.37 Future research
The results from the subgroup analyses examining the associa-
tions between daily PM10 levels and admissions for various diag-
nostic categories, especially when they were considered in relation
to different age groups, were less consistent with previous
findings. The increase in admissions for chronic lung disease
and asthma among all age groups were similar to the results
of Anderson41 and Sunyer42 for chronic obstructive airways
disease (COAD) and Walters25 for asthma. The increase in eld-
Table 4: Estimated percentage change in total cardiac or respiratory admissions in relation to different lags of PM10
between 1988-98 in Christchurch.
Admissions
Lag of PM10 in days
0
1
2
3
4
5
6
Total cardiac
1.26
(0.3-2.21)
2.52
-0.11
(-1.06–0.84)
2.56
0.71
(-0.24–1.66)
3.37
0.38
(-0.57–1.33)
3.09
0.47
-0.02
(-0.9–0.93)
3.21
-0.52
(-1.47–0.43)
3.09
(-0.48–1.42)
3.13
Total respiratory
(1.49-3.55)
(1.53–3.59)
(2.34–4.40)
(2.06–4.12)
(2.10–4.16)
(2.18–4.24)
(2.06–4.12)
2002 VOL. 26 NO. 1
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
27

McGowan et al.
Article
erly admissions for pneumonia and COAD was also consistent
with studies by Schwartz43 and Moolgavkar.44 However, in con-
trast with Pope,33 no statistically significant association was found
in Christchurch between paediatric asthma admissions and PM10.
In addition, while Schwartz28 reported a 1-1.5% increase in ad-
missions for ischaemic heart disease among the elderly, the asso-
ciation in Christchurch did not achieve statistical significance.
The absence of a statistically significant relationship between ei-
ther elderly admissions for ischaemic heart disease or paediatric
hospitalisations for asthma and PM10 may be due to an inadequate
sample size in this study. Although this study had sufficient sta-
tistical power to find an association between total cardiac or res-
piratory admissions and PM10, the number of admissions may
have been too small to find any significant associations between
individual conditions for different age groups and PM10. Biologi-
cal plausibility and the findings from previous research suggests
that associations may actually exist between admissions for some
of these conditions (e.g. dysrhythmia) and particulate pollution.6
A major issue in evaluating the association between particulate
pollution and morbidity is the relative impact of measurement
error in the assessment of exposure and outcome variables. The
relatively homogenous exposure of Christchurch residents to pol-
lution and the relatively minor contribution to pollution from non-
particulate matter reduces the potential for the study to misclassify
exposure.20 The use of 24-hour averages increases the likelihood
that a true catchment-wide measurement of exposure was obtained
but a major issue remains about whether outdoor measurements
can function as a proxy for personal exposure given that people
spend considerable amount of time indoors where they may be
exposed to particulate matter from cigarette or cooking smoke.
Although small diameter particulate matter can penetrate dwell-
ings and high correlations between environmental recordings and
personal exposures to PM10 have been recorded29,45,46 this has yet
to be systematically examined, especially in the New Zealand
setting.
determine the nature and extent of the relationships between
particulate matter and gaseous pollutants and to elucidate the toxi-
cological effects of their different combinations.
In conclusion, this study adds to the accumulating evidence
that particulate air pollution causes significant morbidity and it
augments other local research which has found that exposure to
particulate pollution is associated with significant mortality16 and
morbidity17 in Christchurch. A reduction in ambient particulate
pollution levels can therefore be expected to generate significant
public health benefits.
Acknowledgements
The following support for this study is gratefully acknowledged.
Major funding was provided by a grant from the Health Research
Council of New Zealand.Assistance with obtaining data was pro-
vided by the Canterbury Regional Council. A grant-in-aid was
received from Crown Public Health Limited.
Reference
1. Schwartz J. Air pollution and hospital admissions for respiratory disease.
Epidemiology 1996;7:20-8.
2. Schwartz J, Dockery D. Increased mortality in Philadelphia associated with
daily air pollution concentrations. Am Rev Respir Dis 1992;145:600-4.
3. Katsouyanni K, Zmirou D, Spix C, et al. Short term effects of air pollution on
health: a European approach using time series data. Eur Respir J 1995;8:1030-
9.
4. Pope C, Dockery D, Schwartz J. Review of epidemiologic evidence of health
effects of particulate air pollution. Inhal Toxicol 1995;7:1-18.
5. Environmental and Occupational Assembly of the American Thoracic Soci-
ety Committee. Health effects of outdoor air pollution. Am J Respir Crit Care
Med 1996;153:3-50.
6. Non-biological Particles and Health. London: HMSO, 1995.
7. Lipfert F, Wyzga R. Air pollution and mortality: Issues and uncertainties.
J Air Waste Manage Assoc 1995;45:949-66.
8. Pope C, Bates D, Raizenne M. Health effects of particulate air pollution:
Time for reassessment? Environ Health Perspect 1995;103:472-80.
9. Crane J. Asthma and The Indoor Environment. Wellington: Public Health
Commission, 1996.
10. Progress on Health Outcome Targets. The State of the Public Health in New
Zealand. Wellington: Public Health Commission, 1995.
11. Lets Clear the Air: Issues and Options for Air Quality Management in Can-
terbury and Background Information. Christchurch: Canterbury Regional
Council, 1993.
12. Crane J, Lewis S, Slater T, et al. The self reported prevalence of asthma symp-
toms amongst adult New Zealanders. N Z Med J 1994;107:417-21.
13. Our Future Our Health: Hauora Pakari, Koiora Roa: The State of the Public
Health in New Zealand. Wellington: Public Health Commission, 1993.
14. Health Effects of Suspended Particulate. Christchurch: Canterbury Regional
Council, 1996.
15. Dawson K, Allan J, Fergusson D. Asthma, air pollution and climate: A
Christchurch Study. N Z Med J 1983;96:165-7.
16. Harré E, Price P, Ayrey R, et al. The respiratory effects of air pollution in
chronic obstructive pulmonary disease: A three month prospective study.
Thorax 1997;52:1040-4.
17. Hales S, Town G, Kjellstrom T, et al. Daily mortality in relation to weather
and air pollution in Christchurch, New Zealand. Aust N Z J Public Health
2000;24:89-91.
18. Christchurch Inventory of Home Heating and Motor Vehicle Emissions. Wel-
lington: National Institute of Weather Atmospherics, 1997.
19. Schwartz J. Air pollution and daily mortality: A review and meta analysis.
Environ Res 1994;64:36-52.
20. Vedal S. Ambient particles and health: lines that divide. J Air Waste Manage
Assoc 1997;47:551-81.
21. Hastie T, Tibshirana R. Generalised Additive Models. London: Chapman and
Hall, 1990.
22. Schwartz J. The use of generalised additive models in epidemiology. In: In-
ternational Biometrics Conference; 1994; Hamilton, Ontario, Canada.
23. Ostro B. A search for a threshold in the relationship of air pollution to mortal-
This study addressed the relationship between particulate lev-
els and hospital admissions and did not consider the role of other
pollutants. A notable omission was any data on ozone exposure,
despite overseas research that has found a clear association be-
tween ozone levels and admission rates for respiratory ill-
nesses.33,43 However, it should be noted that the association has
principally been found in conjunction with summertime, photo-
chemical smog rather than particulate pollution.42,43 Studies that
have concurrently examined the effect of particulate and ozone
levels have typically reported that the effect of each is separate.6
Christchurch is considered unlikely to experience significant
effects from ozone, especially during the winter when particulate
levels are highest.18 Although SO2 and NO2 have been shown to
be related to respiratory admissions, the relationship is neither as
consistent, nor as strong, as that with particulate levels, suggest-
ing that these oxides may be exerting an effect mainly through
confounding with airborne particulate.6,33 By contrast, CO and
PM10 may each exert an independent and additive effect upon
cardiac admissions.28 There is a need for further research to
28
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
2002 VOL. 26 NO. 1

The AirWe Breathe
Particulate air pollution and hospital admissions
ity: A reanalysis of data on London winters. Environ Health Perspect
1984;58:397-9.
24. Dab W, Medina S, Quenel Y. Short term respiratory health effects of ambient
air pollution: Results of the APHEA project in Paris. J Epidem Community
Health 1996;50(Suppl):S42-46.
25. Walters S, Phupinyokul M, Ayres J. Hospital admission rates for asthma and
respiratory disease in the West Midlands: Their relationship to air pollution
levels. Thorax 1995;50:948-54.
26. Schwartz J, Morris R. Air pollution and hospital admissions for cardiovascu-
lar disease in Detroit, Michigan. Am J Epidemiol 1995;142:23-35.
27. Burnett R, Krewski D. Air pollution effects on hospital admission rates: A
random effects modelling approach. Can J Stat 1994;22:441-58.
28. Schwartz J. Air pollution and hospital admissions for cardiovascular disease
in Tucson. Epidemiology 1997;8:371-7.
35. Johnston S, Pattemore P, Sanderson G, et al. The relationship between upper
respiratory infections and hospital admissions for asthma:A time-trend analy-
sis. Am J Respir Crit Care Med 1996;154:654-60.
36. Lamm S, Hall T, Engel A, et al. PM10 particulates: Are they the major deter-
minant of pediatric respiratory admissions in Utah County, Utah? Ann Occup
Hyg 1994;28(Suppl):969-72.
37. Lipfert F. A critical review of studies of the association between demands for
hospital services and air pollution. Environ Health Perpect
1993;101(Suppl):229-68.
38. Morris R, Naumova E, Munasinghe R. Ambient air pollution and hospitalisa-
tion for congestive heart failure among elderly people in seven large US cit-
ies. Am J Public Health 1995;85:1361-5.
39. Lipfert F, Hammerstorm T. Temporal patterns in air pollution and hospital
admissions. Environ Res 1992;59:374-99.
29. Thomas K, Pellizzari E, Clayton C, et al. Particle Total Exposure Assessment
Methodology (PTEAM): Method performance and data quality for personal,
indoor and outdoor monitoring. J Expo Anal Environ Epidemiol 1993;3:203-
26.
30. Romieu I, Menses F, Ruiz S, et al. Effects of air pollution on the respiratory
health of asthmatic children living in Mexico City. Am J Respir Crit Care
Med 1996;154:300-7.
31. Gold D, Damokosh A, Pope A, et al. Particulate and ozone pollutant effects
on the respiratory function of children in Southwest Mexico City. Epidemiol-
ogy 1999;10:8-16.
32. Peters A, Dockery D, Heinrich J, et al. Short term effects of particulate air
pollution on respiratory morbidity in asthmatic children. Eur Resp J
1997;10:872-9.
40. Bates D, Baker-Anderson M, Sizto R. Asthma attack periodicity: A study of
hospital emergency visits in Vancouver. Environ Res 1990;51:51-70.
41. Anderson HR, Spix C, Medina S. Air pollution and daily admissions for
chronic obstructive pulmonary disease in 6 European cities: Results from the
APHEA project. Eur Respir J 1997;10:1064-71.
42. Sunyer J, Saez M, Murillo C. Air pollution and emergency room admissions
for chronic obstructive pulmonary disease: A 5 year study. Am J Epidemiol
1993;137:701-5.
43. Schwartz J. PM10 ozone, and hospital admissions for the elderly in
Minneapolis-St. Paul, Minnesota. Arch Environ Health 1994;49:366-74.
44. Moolgavkar S, Luebeck E, Anderson E. Air pollution and hospital admis-
sions for respiratory causes in Minneapolis-St Paul and Birmingham. Epide-
miology 1997;8:364-70.
33. Pope C, Dockery D, Spengler J, et al. Respiratory health and PM10 pollu-
tion; a daily time series analysis. Am Rev Respir Dis 1991;144:668-74.
34. Dockery D, Pope C. Acute respiratory effects of particulate air pollution. Ann
Rev Public Health 1994;15:107-32.
45. Wallace L, Pellizzari E, Hartwell T. The TEAM study: Personal exposures to
toxic substances in air, drinking water and breath of 400 residents of New
Jersey, North Carolina and North Dakota. Environ Res 1987;43:290-307.
46. Brauer M, Koutrakis P, Spengler J. Personal exposure to acidic aerosols and
gases. Environ Sci Technol 1989;23:1408-12.
2002 VOL. 26 NO. 1
AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH
29