Toxicology

Letters
ELSEVIER Toxicology Letters 82/U (1995) 317-321

Dose-effect models for ozone exposure: tool for quantitative

risk estimation

L. van Bree*, M. Marra, H.J. van Scheindelen, P.H. Fischer, S. de Loos,

E. Buringh, P.J.A. Rombout
Nattonal lnsttttcte of Publtc Health and Environment. P.O. Box 1. 3720 BA Btlthoven. The Netherlands

Abstract

Short-term ozone exposure causes lung function decrements, increased airway reactivity. airway inflammation,
increased respiratory symptoms and hospital admissions. Exposure to long-term elevated ozone levels seems to be
associated with reduced lung function (aging), increase of respiratory symptoms, exacerbation of asthma, and
airway cell and tissue changes. Health risk caused by exposure to ozone has been evaluated mainly in a qualitative
way by comparing ozone air quality data with health-based guidelines or standards. A preliminary approach to
quantifying health risk from short-term exposure to oxidant air pollution has been taken by expert judgement,
describing known or expected effects at specific levels of ozone. For quantitative assessment of the health impact of
distinct ozone exposure conditions (acute. repeated daily. chronic) specific exposure-dose-response models are
being developed which can be linked to human exposure data. Exposure-(dose-)response models using data from
epidemiological, human-clinical and animal toxicity studies are presented.

Keywords: Ozone exposure; Exposure-dose-response models; Oxidant air pollution: Lung function

1. Human exposure to oxidant air pollution
ozone
Episodes of increased photochemical activity
may last a number of consecutive days and
several such episodes may occur during a sum-
mer season. The severity of oxidant air pollution
is indicated by daily maximum I- or S-h mean
ozone concentrations [l]. During oxidant air
pollution episodes, elevated ozone concentra-
tions can occur with maximal levels ranging
between 100 and 400 pg/m (0.05-0.2 ppm; 1 h
mean). Diurnal concentration profiles of ozone
* Corresponding author
and shows broad peaks with mean ozone concen-
trations of 80-90% of the maximal l-h average
concentration during 12 h. Public health may
therefore be adversely affected by both acute
and repeated daily exposure to enhanced ozone
concentrations for several h per day. People who
live in areas frequently faced with oxidant air
pollution episodes may experience adverse
health effects from (sub)chronic daily exposure.
Rural and urban regions with relatively high
levels of photochemical activity show frequently
or continuous high levels of ozone during a
whole summer season.
It can be estimated that extremely large reduc-
tions in NO, and VOC will be needed to achieve

0378.4274/95/$09.50 @ 1995 Elsewer Science Ireland Ltd. All rights reserved

SSDI 0378-4274(95)03484-3
318 L. van Bree et al. I Toxicology
a considerable reduction in ozone formation.
These measures are likely to be very costly and
will have a large socio-economic impact. The
public health consequences of not reducing
ozone levels and current exposures is likely to be
very serious, and possibly even more costly.
Therefore, there is an urgent need for quantita-
tive estimates of the incidence, nature, and
magnitude of ozone-related acute, repeated, and
(sub)chronic effects. Such a quantitative ap-
proach can also serve to found and justify pos-
sible effective control actions to reduce risks.
2. Health effects from acute ozone exposure
Health effects caused by acute exposure to
ozone are frequently reviewed and discussed [2].
Controlled acute ozone exposures of humans and
laboratory animals results in acute responses like
lung function impairment, airway hyperrespon-
siveness, and airway inflammation and tissue
injury [3,4]. Recently, data have been presented
suggesting that airway inflammation and tissue
injury also occur in people (children) exposed to
photochemical air pollution [5]). Typical acute
symptoms in people exposed to ozone include
coughing, tightness of the chest, nausea, difficulty
in breathing, and decreased ability to exercise.
A number of epidemiological studies have
shown associations between acute ozone expo-
sure and pulmonary function decrements, daily
increases in morbidity, exacerbation of respirato-
ry diseases like asthma, and increased hospital
admissions and daily mortality [6-91. Some of
the effects may be indicative of the onset of
long-term effects that might ultimately result in
development of persistent, accelerated decline of
lung function (aging), infections, asthma and
chronic obstructive lung diseases, following fre-
quently repeated or chronic exposures to ozone.
A large number of controlled studies now
show effects occurring even below current ozone
standards and guidelines. Because these values
for short-term adverse health effects have low
margins of safety, substantial doubt is growing in
the scientific community whether or not these
standards and guidelines still provide an
Letters 82183 (1995) 317-321
adequate degree of public health protection.
Moreover, ozone responses do not seem to have
an apparent threshold or this threshold value is
very close to background levels of ozone.
3. Critical aspects of dose-response modelling
of ozone
Multiple exposures to ozone in controlled
human and animal studies cause a diverging
pattern of health effects, i.e. attenuation of
respiratory symptoms and pulmonary function
responses and progression of tissue injury, in-
flammatory responses and increased airway re-
sponsiveness [lo-121. Epidemiological studies
have revealed some supportive data for this
divergence, showing attenuation of lung function
decrements in people residing in areas with
frequently recurring high ambient ozone levels
[13]). Controlled studies on the time course of
various health effects also show that some effects
like respiratory symptoms and lung function
responses are maximal immediately following
ozone exposure, whereas other effects, like in-
flammatory responses and tissue damage, may
peak at a later time. Therefore, health risk
analysis of acute and repeated daily ozone expo-
sure must use exposure-response relationships
selected or adjusted for specific ozone exposure
situations and measuring time points.
Individuals in the population vary greatly in
their biological susceptibility to acute responses
to ozone and oxidant air pollution [14], showing
non-responders, normal responders, and hyper-
responders. It is suggested that this variation
depends on both intrinsic (e.g. genetic back-
ground) and extrinsic factors (e.g. socio-econ-
omit status). It should be noted that this re-
sponse variation can be different for various
effects, i.e., a hyperresponder with regard to a
decline in FEV, might be a non- or normal
responder to an inflammatory reaction and vice
versa. The interindividual and intraindividual
differences in ozone-induced response may also
be a reflection of different exposures to ozone
and may therefore be related to differences in
exposure variables (concentration, exposure
 L. van Bree er al. I To.uicology Letters 8218.3 (199.T) 317-321
time, and breathing minute ventilation). There-
fore, because acute ozone responses appear to
depend to a major extent on the cumulative
exposure and inhaled dose, people subjected to
prolonged daily exposure as well as people with
high breathing minute ventilation rates due to
increased physical activity are suggested to be at
an increased risk [15,16].
Population studies indicate that exposure to
ozone during oxidant air pollution may be associ-
ated with exacerbations of asthma and increases
of asthmatic visits to hospital emergency depart-
ments [17,18]. Controlled human studies indicate
that, with similar changes in symptoms, lung
volumes, and bronchial responsiveness in both
normal and asthmatic subjects, ozone-exposed
asthmatics have greater airway obstruction. In
addition, asthmatics appear to have greater air-
way inflammation than healthy subjects [19,20].
Recent studies [21,22] also show that short-term
ozone exposure increased the bronchial allergen
responsiveness in subjects with mild allergic
asthma. Due to the irritant nature of ozone,
which is capable of inducing airway inflammation
and bronchoconstriction, asthmatics and COPD
patients are considered to be at increased risk for
ozone exposure. Ozone may contribute to acute
disease exacerbations, morbidity, and mortality.
Available human and animal toxicity data
have not conclusively demonstrated gender and
racial differences for pulmonary response to
ozone. Data on age as a susceptibility factor are
also inconclusive, although very young children
are suggested to be more responsive due to
greater thoracic and pulmonary doses [23],
whereas older adults seem to have a decreased
pulmonary function response [15].
In order to estimate the health impact of
ozone, it is important to know whether the
effects of ozone are interrelated and whether
some of the effects are causally linked with other
effects that may even be predictive in this re-
spect. A number of studies show, however, a
large dichotomy of various health effects, with
lack or even inverse relationship between ozone-
induced spirometric changes, respiratory symp-
toms, airway inflammation, and allergen respon-
siveness [22,24].
319
4. Exposure-dose-response models for short-
term ozone exposure
Health risks caused by exposure to ambient
oxidant air pollution have been mainly evaluated
in a qualitative way by comparing air quality data
with standards and health-based guidelines.
However, when guidelines are exceeded, these
are only qualitative indications of the likelihood
of adverse health effects and do not allow any
quantitative estimates of the extent and mag-
nitude of risks.
Efforts have been made by the World Health
Organization to describe air quality in terms of
indicator pollutants and to compare these data
with exposure-response relationships based on
the judgement of experts [25]. For ambient
oxidant air pollution, ozone is still considered the
most important component with regard to ad-
verse health effects in airways. Gradations have
been made for all known or expected acute
health effects caused by exposure at certain
ozone levels, and health effects have been
categorized into different classes of severity, (i.e.
mild, moderate, and severe) at specified ozone
concentrations.
There is an increasing need to develop and
implement quantitative methods and mathemati-
cal models to estimate or predict the incidence,
magnitude, and nature of health risks of ozone in
the exposed general population and in specific
(sensitive) subgroups. This approach has to be
based, on the one hand, on exposure-dose-re-
sponse models, i.e. descriptions of the relation-
ship between exposure, dose, and experienced or
predicted incidence of effects in specific human
populations. On the other hand, this approach
has to use exposure models, i.e. descriptions of
potential and actual exposure of (sub)popula-
tions. When exposure can be related to air
quality and various sources or contributive emis-
sions, this approach can be used to calculate risk
reductions dependent of improvements of air
quality.
A number of experimental animal studies with
acute ozone exposures has focused on establish-
ing exposure-time-response relationships and the
importance of cumulative dose, characterized by
320 L. van Bree et al. I Toxicology Letten 82183 (1995) 317-321

concentration (C) and exposure time (T), for With regard to the exposure model, individual
inducing airway permeability [26] and airway exposure and inhaled dose are dependent upon a
responsiveness [27]. The role of breathing minute range of variables, such as geographical area,
volume (V,) has only been studied to a limited relationships between outdoor and indoor air
extent in laboratory animal studies, showing that quality, time-activity and physical activity pat-
physical exercise, and the concomitantly inhaled terns, and physiological variables. Mathematical
dose and dose rate, may significantly enhance models can be developed with functional mod-
pulmonary deposition and related pulmonary ules that describe individual exposure parame-
function or permeability effects [4,28]. ters. These parameters can be calculated from
A number of controlled human studies with time series of regional and urban air quality data,
determination of various measures of lung func- time-activity and physical activity patterns de-
tion have also been focused on the precise rived from population survey statistics, demo-
inhaled (effective or delivered) dose concept of graphic data, and appropriate functions and
single ozone exposure. Inhaled effective dose is distributions of variables, such as indoor-outdoor
thereby defined as the product of C, T, and V,. It relationships. From these data, population fre-
appears that the magnitude of lung function quency distributions can be estimated while
decrements, measured directly following ozone preserving individual characteristics, providing
exposure, is related to the inhaled effective the possibility to identify individuals in specific
(internal thoracic) dose [29,30]. It has been sub-populations, (e.g. the upper tail of a dis-
suggested that, both for young and old adults, the tribution). These types of exposure models are
relative ranking of contributing dose variables still in the phase of development. Ultimately,
associated with pulmonary function decline is C however, they can be used to estimate the
> V, > T [31j2]. . exposure of the population and to predict specific
Recently, a number of studies focused on health effects of interest, provided that exposurs-
establishing mathematical, non-linear models dose-response relationships are known.
that describe ozone-induced decrements in lung
function (FEV,) and increases in airway per-
References
meability as a function of total inhaled dose
(C X T X V,) and dose rate (C X V,). Studies with
Rombout. P.J.A.. Lioy. P.J. and Goldstem. B.D. (1986)
different protocols and levels of C and T, and to Rationale for an eight hour ozone standard. J. Air
a limited extent also V,, have been evaluated for Pollut. Control Assoc. 36, 913-917.
the contributive and interactive role of the vari- Lippmann, M. (1993) Health effects of tropospheric
ous dose variables in humans [26,31,33]. Studies ozone a review of recent research findings and their
lmphcations to ambient air quality standards. J. Expos
have also reported on the use of mathematical
Anal. Environ. Epidemlol. 3. 103-129.
models describing interindividual differences in Devhn, R B.. McDonnell. W.F.. Mann, R., Becker, S.,
acute pulmonary function responses following House, D.E., Schreinemachers. D. and Koren. H.S.
ozone exposure [15,16]. ( 1991) Exposure of humans to ambient levels of ozone
to 6.6 h causes cellular and lxochemical changes in the
lung. Am. J Respir. Cell Mol. Biol. 4. 72-81
5. Quantitative risk estimation
Van Bree. L.. Marra. M. and Rombout. P.J.A. (1992)
Differences in pulmonary biochemical and inflammatory
Quantitative health risk assessment of ozone responses of rats and guinea pigs resultmg from daytime
exposure can provide information on the extent or nIghttIme. single and repeated exposure to ozone.
and magnitude of effects in the general popula- Toxxol. Appl. Pharmacol. 116, 209-216.
Fnscher, Th.M.. Kuehr. J.. Pullwltt. A.. Meinert. R .
tion, specific risk groups, and individuals, if
Forster, J.. Studnicka, M. and Koren. H. (1993) Ambient
actual exposure can be estimated and exposure ozone causes upper auways mflammation in children.
or dose-response relationships can be estab- Am. Rev. Respir. DIS. 148. 961-964.
lished. Kinney, P.L and Ozkaynak. H. (1991) Assoclatlons of
 L. vnn Bree et al. I Toxicology Letters 82/X.? (19X) .il7-.?r_11 321

datlv mortality and au polltton in Los Angeles County. [19] Basha, M.A.. Gross. K.B., Gwtzdala. C.J . Hatdar. A.H.
Environ. Res. 54, 99-120. and Popovich, J. (1994) Bronchoalveolar lavage neutro-
[7] Hoek, G.. Fischer, P., Brunekreef, B.. Lebret. E., phtha m asthmattc and healthy volunteers after con-
Hofschreuder. P and Mennen, M.G. (1993) Acute trolled exposure to ozone and filtered purified air. Chest
effects of ambient ozone on pulmonary function of 106. 175771765.
children in the Netherlands. Am. Rev. Respir. Dis. 147. [20] McBride, D.E.. Koemg. J.Q.. Luchtel. D.L.. Willlams.
111-117 P.V. and Henderson. W.R. (1994) Inflammatory effects of
[8] Thurston. G D., Ito, K., Lippmann. M.. Hayes, C.G. and ozone m the upper airways of subjects with asthma. Am.
Bates. D.V (1994) Respiratory hospttal admissions and J. Resptr. Crtt. Care Med. 149, 1192-1197.
summertime haze au pollution m Toronto. Ontario: [21] Molfino, N.A , Wright, S.C., Katz, I.. Tarlo. S., Silver-
considerations of the role of acid aerosols. Environ. Res. man. F., M&lean, P.A.. Szalai. J.P.. Raizenne. M..
65. 271-290. Slutsky. AS. and Zamel. N. (1991) Effect 01 low
[9] Burnett, R.T.. Dales, R.E.. Raizenne. M.E.. Krewski. D , concentrattons of ozone on inhaled allergen responses in
Summers. P.W.. Roberts, G.R., Raad-Young. M., Dan. asthmattc subjects. Lancet 338. 199-203.
T. and Brook. J. (1994) Effects of low ambient levels of [22] Jorrcs. R 1 Nowak. D. and Magnusscn, H. (1995) The
ozone and sulfates on the frequency of respiratory effect of ozone exposure on allergen responsiveness in
admisstons to Ontario hospitals. Environ. Res. 6.5. 172- subjects with asthma or rhmitis. Am J. Resptr. Crit.
194. Care Med. (m press).
[lo] Tepper, J.S.. Costa, D.L.. Lehmann, J.R., Weber. M.F. [23] Overton. J.H. and Graham. R.C. (1989) Predictions of
and Hatch, G.E. (1989) Unattenuated structural and ozone absorption rn human lungs from newborn to
biochemical alterations m the rat lung during functtonal adult. Health Phys. 57. 29936
adaptation to ozone. Am. Rev. Respir. Dis 140. 493- [24] Schelegle. ES.. Siefkin. A.D. and McDonald. R.J.
501. (1991) Time course of ozone-induced neutrophilia in
[ll] Van Bree. L.. Rombout. P.J.A., Rietjens, I.M.C.M.. normal sublects. Am. Rev. Respn Dis. 143. 1353-13.5X.
Dormans, J.A.M.A. and Marra. M. (1989) Pathobioch- [2S] WHO (1992) Acute Effects on Health of Smog ELI-
emical effects m rat lung related to episodic ozone sodes World Health Organization Regional Puhhcations
exposure. In: T. Schneider et al (Eds.), Proceedings of Series No. 43.
the Third US-Dutch International Symposium on At- [26] Highfill. J.W. and Costa, D L (1995) Statistical response
mospheric Ozone Researh and its Policy Imphcattons. models for ozone exposure: then generality when rp-
Elsevier. Amsterdam. pp. 723-732. plied to human sptrometric and animal permeabihty
[12] Fohnsbec, L.J.. Horstman. D.H.. Kerhl. H.R., Harder. functions of the lung. J. An Waste Manage Assoc. 45.
S 1 Ahdul-Salaam, S. and Ives, P.J. (1994) Respiratory 95-107.
responses to repeated prolonged ozone exposure to 0.12 [27] Ntshtkawa, M.. Suzukt, S . Ikeda. H.. Fukuda. T., Suzukt.
ppm ozone. Am. J. Respir. Crit. Care Med. 149. 98-105. J. and Okubo. T. (1990) Dose-response relationshtp of
[13] Hackney. J.D.. Linn, W.S.. Mohler. J.G. and Collier. C.R. ozone-induced airway hyperresponsiveness in unanes-
(1977) Adaptation to short-term respiratory effects of thetized gumea pigs J Toxic01 Envtron Health 30.
ozone in men exposed repeatedly. J. Appl. Physioi. 43, 13-134
82-85. [28] Mautz, W.J., McClure. T.R . Reischl, P. Phalen. R.F and
[14] McDonnell. W.F. (1991) IntersubJect variability in Crockcr, T.T. (1985) Enhancement of ozone-induced
human acute ozone responses. Pharmacogenetics 1. 1 lo- lung injury by exercise. J Tox~col. Environ Health 16,
113. 841-851.
[lS] McDonnell, WF.. Muller, K.E.. Bromberg. P.A. and Shy. [29] Adams. W.C., Savm. W.M and Chrism. A E. (1981)
C.M. (1993) Predictors of mdividual differences m acute Detectton of ozone toxicity during contmuou, exercise
response to ozone exposure. Am. Rev. Respn. DIS. 147. via the effective dose concept. J Appl Physiol. 51.
818-825 41.5~322.
[16] Gerrity. T.R., McDonnell, W.F. and House, D.E. (1994) [30] Klemman. M.T. (1991) Effects ot ozone on pulmonary
The relationship between delivered ozone dose and tunction: the relattonship of response to dose. J. Expos.
functional responses m humans, Toxlcol Appl. Phar- Anal. Environ. Epidemiol 1. 3099325
macol. 124. 275-283. [31] Hazucha. M (1987) Relationship between ozone expo-
[17] Cody. R.P., Weisel. C.P., Birnbaum. G. and L~oy. P.J. sure and pulmonary function changes J. Appl Physiol.
( 1992) The effect of ozone associated with summertime 62. 1671-1680.
photochemical smog on the frequency of asthma vistts to [32] Drechsler-Parks. D.M.. H orvath, S.M and Bcdi. J.F.
hospital emergency departments. Environ. Res. 58. 1844 (1990) The etfective dose concept in older adults ex-
194 posed to ozone. Exp. Gerontol. 25, 107-I 15
[18] White. M.C.. Etzel, R.A.. WIICOX. W.D. and Lloyd, C. [33] McDonnell. W.F. and Smith. M.V (1994) Description of
(1994) Exacerbations of childhood asthma and ozone acute ozone response as a functton of exposure rate and
pollution in Atlanta. Environ. Res. 65. 56-68. total mhalcd dose. J. Appl Physiol 76. 27762783