1. science
2. ecology
3. pollution

Health effects of air pollution in Iceland
Respiratory health in volcanic environments
Hanne Krage Carlsen
Thesis for the degree of Philosophiae Doctor
Supervisor:
Thorarinn Gíslason
Advisor:
Anna Oudin
Doctoral committee:
Anna-Carin Olin
Bertil Forsberg
Unnur Valdimarsdóttir
October 2014
Heilbrigðisáhrif loftmengunnar á Íslandi
Öndunarfæraheilsu í eldgosaumhverfi
Hanne Krage Carlsen
Ritgerð til doktorsgráðu
Umsjónarkennari:
Þórarinn Gíslason
Leiðbeinandi:
Anna Oudin
Doktorsnefnd:
Anna-Carin Olin
Bertil Forsberg
Unnur Valdimarsdóttir
Október 2014
‚
Thesis for a doctoral degree at the University of Iceland. All right reserved.
No part of this publication may be reproduced in any form without the prior
permission of the copyright holder.
© Hanne Krage Carlsen 2014
ISBN 978-91-7601-082-2
Printing by Háskólaprent
Reykjavik, Iceland 2014
Statement of collaboration
This thesis and the work in it have been produced in collaboration between
University of Iceland and Umeå University.
The thesis will be defended and issued at both institutions.
Ágrip
Loftmengun hefur slæm áhrif á heilsu manna. Öndunarfærin eru mest útsett
og skammtímabreyting í styrk loftmengunnar hefur verið tengd við versnun á
teppusjúkdómum í lungum og aukningu á tíðni hjartaáfalla og heilablóðfalla.
Loftmengun í borgum vegna bílaumferðar er áhyggjumál þar sem mikið af
fólki er útsett. Á hinn boginn geta einnig náttúrulegar mengunaruppsprettur
eins og ryk, sandfok og aska frá eldgosum ógnað heilsu manna. Útsetning
fyrir náttúruhamförum á borð við eldgos getur einnig hótað geðheilbrigði. Ekki
hefur verið gert mikið af rannsóknum á Íslandi þrátt fyrir tilvist fleiri
náttúrulegra mengunaruppspretta og talsverðs magns bíla á götum
höfuðborgarsvæðis Íslands.
Markmið ritgerðar þessarar var að rannsaka hvort loftmengun hafi
mælanleg áhrif á heilsu íbúa Íslands. Rannsóknarnálgunin var tvíþætt.
Annars vegar voru notuð gögn úr heilbrigðisskrám sem voru skoðuð með
tímaraðagreiningaraðferð til að rannsaka skammtímaáhrif daglegra sveifla
loftmengunar og breytileika í daglegum fjölda fólks sem tók út lyfseðilsskyld
astmalyf eða leituðu sér hjálpar á bráðamóttöku vegna hjarta-,
lungnasjúkdóma eða heilablóðfalls. Hins vegar var litið á heilsuáhrif eldossins
árið 2010 í Eyjafjallajökli.
Í grein I var tímaraðagreining notuð til að rannsaka samband daglegs
fjölda fólks sem leysti út lyfseðilsskyld lyf gegn astma og magns
loftmengunar; svifryksagna minna en 10 μm (PM10), níturdíoxíðs (NO2),
ósóns (O3) og brennisteinsvetnis (H2S). Rannsóknartímabilið var frá árinu
2006 fram til 2009 og reyndist marktækt samband dagsgilda PM 10 og H2S og
fjölda þeirra sem leystu út lyf 3-5 dögum síðar. Þegar litið var á útsetningu
sem hæsta klukkutímagildi hvers dags var sambandið sterkara.
Grein II hafði fjöldi einstaklinga sem leitaði til bráðamóttöku Landspítala
Háskólasjúkrahúss vegna hjarta-, lungnasjúkdóma eða heilablóðfalla 20032009 sem meginrannsókarútkomu. Tímaraðagreiningu var beitt til að finna
bestu forspárþættina meðal PM10, NO2, og O3. Marktækt samband var á milli
styrks O3 og fjölda koma á bráðamóttöku og innlagna sama dag, sem og
einum og tveimur dögum síðar. Sambandið fannst bæði fyrir karla og konur
og meðal einstaklinga eldri en 70 ára. Einungis fannst samband á milli NO 2
og koma á bráðamóttöku og innlagna meðal eldri einstaklinga. Hins vegar
reyndist PM10 ekki marktækur forspárþáttur.
i
Í grein III var markmiðið að rannsaka hvort heilbrigðisáhrif PM 10 breyttust
eftir að aska frá gosunum í Eyjafjallajökli og Grímsvötnum bættust við PM 10 á
höfuðborgarsvæðinu. Rannsóknartímabilið var 2007-2012 og sama útkoma
var notuð og í grein II þ.e. komur á bráðamóttöku og innlagnir vegna hjarta-,
lungnasjúkdóma og blóðfalla. Tímaraðagreining sýndi að áhrif PM 10 höfðu
tilhneigingu til þess að vera meiri eftir gos en niðurstöður voru ekki
tölfræðilega marktækar. Þegar tvíþættir vísar fyrir uppsprettu PM 10 í
eldgosaösku voru bornir saman við aðrar uppsprettur sást að þegar styrkur
PM10 var hár einn dag vegna eldgosaösku var marktæk aukning á fjölda
koma á bráðamóttöku og bráðainnlagna. Aftur á móti var hár styrkur PM 10
ekki einn og sér tengdur aukningu.
Í grein IV var heilsa þeirra sem bjó á svæðinu sem var mest útsett fyrir
gosinu í Eyjafjallajökli árið 2010 könnuð þegar gosinu var að ljúka. Notast var
við blástursmælingar til að rannsaka áhrif á lungnastarfsemi, auk
læknisskoðunar og spurningalista. Lungnastarfsemi hópsins reyndist vera
svipuð og hjá samanburðarhóp frá Reykjavík en reykingar voru algengari hjá
samanburðarhópnum. Margir þjáðust af einkennum frá skynfærum, streitu og
slæmri andlegri líðan. Einstaklingar með undirliggjandi sjúkdóma höfðu mesta
aukningu á einkennum.
Grein V er spurningarannsókn gerð sex mánuðum eftir að gosi í
Eyjafjallajökli lauk. Rannsóknarhópurinn samanstóð af Sunnlendingum sem
bjuggu mislangt frá eldfjallinu ásamt samanburðarhópi frá norðurhluta
Íslands. Einkenni frá öndunarfærum og augum voru algengari hjá
Sunnlendingum eftir leiðréttingu fyrir aldri, kyni, menntun og reykingum.
Andleg vanlíðan var enn há hjá þeim sem voru mest útsettir en hafði minnkað
talsvert miðað við fyrri niðurstöður greinar IV.
Í ritgerð þessari hefur verið sýnt fram á skammtímaáhrif bæði mengunar í
bæjum og náttúrulegra agna á úttekt á astmalyfjum og koma á bráðamótttöku
og innlagna. Svifryksagnir frá eldgosum voru sérstaklega tengdar auknum
öndunarfæraeinkennum í útsettu þýði.Niðurstöður bentu til þess að samband
finnist á milli eldgosa-agna og fjölda koma á bráðamótttöku eftirfylgjanda
daga.
Lykilorð:
Öndunarfæraheilsu, loftmengun, eldfjallaösku, faraldsfræði, brennisteinsvetni.
ii
Abstract
Air pollution has adverse effects on human health. The respiratory system is
the most exposed and short-term changes in air pollution levels have been
associated with worsening of asthma symptoms and increased rates of heart
attacks and stroke. Air pollution in cities due to traffic is the major concern, as
many people are exposed. However, natural sources of air pollution such as
natural dust storms and ash from volcanic eruptions can also compromise
human health. Exposure to volcanic eruptions and other natural hazards can
also threaten mental health. Air pollution has not been extensively studied in
Iceland, in spite of the presence of several natural pollution sources and a
sizeable car fleet in the capital area.
The aim of this thesis was to determine if there was a measurable effect
on health which could be attributed to air pollution in Iceland. This aim was
pursued along two paths; time series studies using register data aimed to
determine the short-term association between daily variation in air pollution
and on one hand daily dispensing of anti-asthma medication or the daily
number of emergency room visits and emergency admissions for
cardiopulmonary causes and stroke. The other method was to investigate if
exposure to the Eyjafjallajökull volcanic eruption was associated with adverse
health outcomes, either at the end of the eruption, or 6 months later.
In paper I time series regression was used to investigate the association
between the daily number of individuals who were dispensed anti-asthma
medication and levels of the air pollutants particle matter with an
aerodynamic diameter less than 10 μm (PM10), nitrogen dioxide (NO2), ozone
(O3), and hydrogen sulfide (H2S) during the preceding days. For the study
period 2006-9, there were significant associations between the daily mean of
PM10 and H2S and the sales of anti-asthma medication 3 to 5 days later.
Giving the exposure as the highest daily one-hour mean gave more
significant results. Air pollution negatively affected the respiratory health of
asthma medication users, prompting them to refill their prescriptions before
they had originally intended to.
In paper II the main outcome was the number of individuals seeking help
at Landspitali University Hospital emergency room for cardiopulmonary
disease or stroke. Time series regression was used to identify the lag that
gave the best predictive power, and models were run for data for 2003-9
iii
pollutants PM10, NO2, and O3. O3 was significantly associated with the
number of emergency hospital visits the same day and two days later in all
models, and both for men, women and the elderly. Only emergency hospital
visits of the elderly were associated with NO 2, and there were no
associations with PM10.
In paper III the aim was to investigate if the health effects of PM10 were
affected by the addition of volcanic ash from the 2010 eruption of
Eyjafjallajökull and 2011 eruption of Grímsvötn to PM10 in the capital area.
Time series regression of emergency hospital visits and PM 10 before and
after the Eyjafjallajökull eruption showed that the effect tended to be higher
after the eruption, but the results were not significant. Analysis with a binary
indicator for high levels of PM10 from volcanic ash and other sources showed
that volcanic ash was associated with increased emergency hospital visits.
There were no associations with high levels of PM10 from other sources.
In paper IV, the health of the population exposed to the ongoing eruption
of Eyjafjallajökull in 2010 was investigated thoroughly. Lung function in adults
was similar to that in a reference group from the capital area, but smoking
was more common in the reference group. Many from the exposed group
reported sensory organ irritation symptoms and symptoms of stress and
psychological morbidity, especially those with underlying diseases.
Paper V report the results from a questionnaire study which was carried
out six months after the Eyjafjallajökull eruption. The study population
comprised a cohort of South Icelanders exposed to the eruption to varying
degrees and a reference group from North Iceland. Respiratory and eye
symptoms were more common in South Icelanders than in the reference
group, after adjusting for demographic characteristics. Psychological
morbidity rates had declined considerably.
In the studies, we found that urban air pollution and natural particles have
short-term effects on anti-asthma medication dispensing and emergency
room visits and hospital admissions. Exposure to natural particles in the form
of volcanic dust was associated with increased respiratory symptoms in a
very exposed population. There were indications that volcanic ash particles
were associated with increased emergency hospital visits in the following
days.
Keywords:
Respiratory health, air pollution, volcanic ash, epidemiology, hydrogen
sulfide.
iv
Acknowledgements
Many people have contributed to this work, and I would like to thank them all.
First of all the people of South Iceland and Skagafjörður who took the time to
answer the questionnaires and participate in the studies.
I want to thank my main supervisor Þórarinn Gislason for his
encouragement and scientific mentorship for this thesis and tireless effort in
finding support for it. I thank my advisor Anna Oudin for her support and the
effort she has put into this thesis. Thank you also to Unnur Valdimarsdóttir
who always thought I would make a good PhD candidate. I am very grateful
to my advisors and collaborators in Sweden; Anna-Carin Olin, Kadri Meister
and Bertil Forsberg for expert advice and much support.
A big thank you to my co-authors, Arna Hauksdóttir, Bryndís
Benediktsdóttir, Þórir Björn Kolbeinsson, Þröstur Þorsteinsson, Gunnlaug
Einarsdóttir, Halldór Runólfsson, Sigurður Guðmundsson & Þorsteinn
Jóhansson. Paper I would not have been publishable without the tireless
efforts of Helga Zoëga and Birgir Hrafnkelsson, who helped guided me
through decades of evolution in time trend adjustment methods. The
collaboration around the Eyjafjallajökull studies taught me a lot, especially
Gudrún Pétursdóttir and Haraldur Briem deserve much thanks for their help
with publishing the results.
Dóra Ólafsdóttir at the Centre of Public Health at University of Iceland has
been most helpful with my long-distance navigation of University of Iceland
administration.
My fellow PhD student Ragnhildur Finnbjörnsdóttir has been a good
collaborator and friend during this period, and I thank her, as well as my other
colleagues at the Centre of Public Health at the University of Iceland, Umeå
University, and AMM in Gothenburg.
The following have kindly helped me gather the data used in this thesis:
Ingibjörg Richter, Anna Rósa Böðvarsdóttir, Hildur Friðriksdóttir, Kristinn
Jónsson, Margrét Hallsdóttir and Lára Guðmundsdóttir.
The following institutions supported this project:
The University of Iceland Centre of Public Health and Umeå University
Department of Occupational and Environmental Medicine, Occupational and
Environmental Medicine at Sahlgrenska School of Public Health and
v
Community Medicine, the latter has been my work place for the last two
years, both the colleagues and the professional environment has been a big
encouragement and inspiration. Also the Sæmundur Fróði Institute for
Sustainability Studies and The Swedish Environmental Research Institute in
Gothenburg (IVL).
I thank my family for their support and encouragement, and my husband,
Steinn, my rock. Finally, I want to thank my friends; especially Aysha
Hussain, Gunnar Thorarensen & Virgile Collin-Lange who all contributed to
this thesis with valuable input.
This thesis was funded through Rannis - The Icelandic Centre for
Research PhD, the Department for Occupational and Environmental
Medicine, Umeå University, and the Icelandic Ministry of Welfare
(Eyjafjallajökull health studies). Additional funding have been provided by the
University of Iceland travel grant, The Swedish Society of Medicine Section
for Occupational and Environmental health travel grant, and the Icelandic
Association of Academics re-education grant.
vi
Contents
Ágrip .............................................................................................................. i
Abstract ....................................................................................................... iii
Acknowledgements ..................................................................................... v
Contents ..................................................................................................... vii
List of abbreviations .................................................................................. ix
List of figures............................................................................................... x
List of tables .............................................................................................. xii
List of original papers .............................................................................. xiii
Declaration of contribution...................................................................... xiv
1 Introduction ............................................................................................ 1
1.1 Air pollution ....................................................................................... 2
1.1.1 Traffic pollution .......................................................................... 2
1.1.2 Dust and sand storms ............................................................... 5
1.1.3 Geothermal areas and hydrogen sulfide ................................... 5
1.1.4 Volcanic eruptions ..................................................................... 5
1.1.5 Other sources............................................................................ 6
1.2 Air pollution effects on health ............................................................ 7
1.2.1 Pathways .................................................................................. 7
1.2.2 Evidence from epidemiological studies .................................... 11
1.3 Respiratory health and air pollution in Iceland ................................ 17
1.3.1 Mineral dust sources ............................................................... 19
1.3.2 Hydrogen sulfide ..................................................................... 19
1.3.3 Icelandic volcanoes ................................................................. 20
2 Aims ...................................................................................................... 23
3 Materials and methods ........................................................................ 25
3.1 Data sources .................................................................................. 25
3.1.1 Exposure ................................................................................. 25
3.1.1 Health outcomes ..................................................................... 27
3.1.1 Covariates ............................................................................... 29
3.1.2 Ethics approval........................................................................ 32
3.2 Methods .......................................................................................... 32
3.2.1 Short-term effect studies ......................................................... 32
3.2.2 Long-term effect studies .......................................................... 35
4 Results .................................................................................................. 39
vii
4.1
4.2
4.3
4.4
4.5
Paper I ............................................................................................ 39
Paper II ........................................................................................... 44
Paper III .......................................................................................... 44
Paper IV .......................................................................................... 45
Paper V ........................................................................................... 46
5 Discussion ............................................................................................ 49
5.1 Key results ...................................................................................... 49
5.2 Limitations and methodological considerations............................... 50
5.2.1 Exposure ................................................................................. 50
5.2.2 Outcomes ................................................................................ 52
5.2.3 Statistical methods .................................................................. 54
5.3 Interpretation................................................................................... 54
5.4 External validity............................................................................... 55
6 Conclusions .......................................................................................... 57
6.1 Limitations of the current study and future perspectives ................. 58
References ................................................................................................. 59
Original publications ........................................ Error! Bookmark not defined.
Paper 1 .............................................................. Error! Bookmark not defined.
Paper 2 .............................................................. Error! Bookmark not defined.
Paper 3 .............................................................. Error! Bookmark not defined.
Paper 4 .............................................................. Error! Bookmark not defined.
Paper 5 .............................................................. Error! Bookmark not defined.
Appendix .................................................................................................. 153
viii
List of abbreviations
ATC
Anatomical therapeutic classification system
BC
Black carbon
BS
Black smoke
CO2
Carbon dioxide
COPD Chronic obstructive pulmonary disease
ER
Emergency room
FEV1
Forced expiratory volume in 1 second
FVC
Forced vital capacity
GAM
Generalized additive model
GHQ
General health questionnaire
GLM
Generalized linear model
GP
General practitioner
ICD
International classification of diseases
H2S
Hydrogen sulfide
Lowess Locally weighted scatterplot smoothing
NO2
Nitrogen dioxide
O3
Ozone
PAH
Poly-aromatic hydrocarbons
PEF
Peak expiratory flow
PM2.5
Particle matter with an aerodynamic diameter less than 2.5 µm
PM10
Particle matter with an aerodynamic diameter less than 10 µm
(micrometer)
PTSD Post traumatic stress syndrome
SO2
Sulfur dioxide
UFP
Ultrafine particle matter with an aerodynamic diameter less than 1
µm
VOC
Volatile organic compounds
WHO
World health organization
ix
List of figures
Figure 1 Concentrations of NO2 and O3 at an urban roadside
measuring station, Reykjavík, over 4 days starting January
21, 2006. ......................................................................................... 3
Figure 2 Reykjavík seen from the north on a day with high levels of
PM due to a dust storm (top) and on a day with clear skies
(below) (image: Hanne Krage Carlsen). .......................................... 4
Figure 3 The ash plume of Eyjafjallajökull May 7 2010. The ash
contained as much as 20% PM10 (image: Hanne Krage
Carlsen)........................................................................................... 6
Figure 4 Overview of the respiratory system and particle deposition
(image: Hanne Krage Carlsen/Creative Commons). ....................... 7
Figure 5 Suggested biological pathways for health effects of air
pollution to different parts of the circulatory system (image:
Hanne Krage Carlsen/Creative commonts). .................................... 9
Figure 6 The emissions from a geothermal power plant contain H2S
(image: Hanne Krage Carlsen)...................................................... 13
Figure 7 Ash from the Eyjafjallajökull eruption causes bad visibility
mid-day May 7, 2010 (image: Hanne Krage Carlsen). .................. 16
Figure 8 Hellisheidi geothermal powerplant seen from the north
(image: Hanne Krage Carlsen)...................................................... 20
Figure 9 The Eyjafjallajökull volcano 2008 (left) and 2010 (right)
(image: Hanne Krage Carlsen)...................................................... 21
Figure 10 Composition of ambient PM10 in Reykjavík 2003 and 2013
(Image: Hanne Krage Carlsen based on Skúladóttir et al.,
2003 & Vegagerðin, 2013). ........................................................... 22
Figure 11 Reykjavík and surrounding area with main pollution
sources and measuring stations (Adapted from paper II) .............. 25
Figure 12 Ash dispersion around Eyjafjallajökull (left) and the
exposure areas based on it (right) (Image: Thorsteinsson,
from Paper IV and V). ................................................................... 26
Figure 13 Example of missing data in a multi-pollutant time series.
The shaded areas will be excluded from regression analysis
as values of one or more pollutants are missing. .......................... 35
x
Figure 14 Time series of cardiovascular, stroke, and respiratory
infectious admissions and ER visits 2003-2012 ............................ 40
Figure 15 Daily means of air pollutants (μg/m3) measured in
Reykjavik 2003-2012..................................................................... 41
Figure 16 Overview of Reykjavík postcodes and measuring stations.
Circled postcodes are re-analysed in figure 16 and table 5. ......... 42
Figure 17 The association between anti-asthma drug dispensing in
east Reykjavik and H2S. ................................................................ 43
Figure 18 Proportion of cohort reporting “some” or “a lot” of exposure
to volcanic phenomena at their residence during the
eruption in Eyjafjallajökull 2010. .................................................... 46
xi
List of tables
Table 1 Air pollution health limits in Iceland (µg/m3). .................................... 19
Table 2 Schematic overview of data used in each study .............................. 30
Table 3 Ethics approval sought for each study ............................................. 31
Table 4 Schematic overview of the methods used in each study ................. 37
Table 5 H2S and anti-asthma medication dispensing in areas within
Reykjavík lags 0-15. Adjusted for other pollutants, weather
and time trend. .............................................................................. 43
Table 6 Reply rates for Eyjafjallajökull cohort study – web and online ......... 47
xii
List of original papers
This thesis is based on the following original publications, which are referred
to in the text by their Roman numerals:
I.
Carlsen, H. K., Zoëga, H., Valdimarsdóttir, U., Gíslason, T., &
Hrafnkelsson, B. (2012). Hydrogen sulfide and particle matter levels
associated with increased dispensing of anti-asthma drugs in
Iceland’s capital. Environmental Research, 113, 33–39.
II.
Carlsen, H. K., Forsberg, B., Meister, K., Gíslason, T., & Oudin, A.
(2013). Ozone is associated with cardiopulmonary and stroke
emergency hospital visits in Reykjavík, Iceland 2003–2009.
Environmental Health, 12(1), 28.
III.
[manuscript] Carlsen, H. K., Gislason, T., Forsberg, B., Meister, K.,
Thorsteinsson, T., Jóhansson, T., Finnbjornsdottir, R. G., Oudin, A.,
Effects of PM10 from volcanic ash, natural dust, and other sources on
emergency hospital visits in Reykjavík, Iceland.
IV.
Carlsen, H. K., Gislason, T., Benediktsdottir, B., Kolbeinsson, T. B.,
Hauksdottir, A., Thorsteinsson, T., & Briem, H. (2012). A survey of
early health effects of the Eyjafjallajökull 2010 Eruption in Iceland: A
population-based study. BMJ Open, 2(2), e000343.
V.
Carlsen, H. K. & Hauksdottir, A., Valdimarsdottir, U. A., Gíslason, T.,
Einarsdottir, G., Runolfsson, H., Briem, H., Finnbjornsdottir, R.G.,
Gudmundsson, S., Thorsteinsson, T., Pétursdóttir, G. (2012). Health
effects following the Eyjafjallajökull volcanic eruption: A cohort study.
BMJ Open, 2(6), e001851.
All papers are reprinted by kind permission of the publishers.
xiii
Declaration of contribution
In paper I, the candidate wrote the ethics application and obtained the data
from the authorities, including ethics permissions. The candidate analyzed
the data, wrote the paper in collaboration with the supervisors and reanalyzed in accordance with reviewer comments in the publication process.
For paper II, the candidate gathered the exposure data and analyzed them,
and wrote the paper in collaboration with the co-authors.
For paper III, the candidate obtained the exposure data, treated and analyzed
all data, and wrote the paper in collaboration with the supervisors.
For paper IV, the candidate participated in the preparation of the
questionnaire, and the collection of the data. The candidate then digitized
and analyzed the data. The manuscript was written by the candidate in
collaboration with the co-authors and the corresponding author, who
ultimately handled the manuscript reviewing and the publication process.
For paper V, the candidate participated in the preparation of the
questionnaire and the collection of the data. The candidate digitized and
analyzed the data. The manuscript was written in collaboration with the coauthors and the corresponding author, who ultimately handled the manuscript
submission and reviewing process.
xiv
1 Introduction
Air pollution is defined as undesired gases and particles in ambient air,
commonly known as aerosols, because of smell, nuisance or adverse health
effects (WHO, 2000, p1).
Since the early days of industrialization, air pollution has been recognized
as a hazard to human health (WHO, 2000, p16). The earliest reports of
health effects of airborne particles date back to the 15 th and 16th centuries;
the women of the Carpathian mountains (mostly in present-day Romania)
went through as many as seven husbands in their lifetime - the men who
worked in the granite mines died early from disease (silico-tuberculosis)
caused by exposure to the silica dust (Donaldson and Seaton, 2012). Falun,
Sweden, was home to the largest copper mine in Europe in the early 18 th
century. Visitors described that “Intense, black copper smoke lies dense over
the town. [The smoke] causes intense sneezing, and lung disease was more
common in this area” (Dunér, 2012, pp. 79-83).
Modern studies of air pollution and health began in the 20 th century
following the Meuse Valley disaster in Belgium 1930 (Nemery et al., 2001)
and the London fog (Logan, 1956), both events where short-term increases in
pollution levels caused increased morbidity and mortality. Results from The
Six Cities Study showed that chronic exposure to high levels of air pollution
were associated with increased morbidity and mortality rates (Dockery and
Pope, 1993).
Today, the studies of health effects of air pollution remain focused on
anthropogenic traffic pollution which affects many people who live in cities
and urban areas. In later years, increased focus has been pointed at air
pollution from natural sources such as desert dust (Morman and Plumlee,
2013), gas emitted from geothermal areas (Hansell and Oppenheimer, 2004)
or ash from volcanic eruptions (Horwell and Baxter, 2006). Air pollution is
responsible for a substantial part of respiratory mortality and morbidity (Künzli
et al., 2000).
In the introduction, I first describe types of pollution; secondly, the health
effects, and briefly review the epidemiological studies of health effects for
different types of pollution; and finally, describe the Icelandic situation and
studies of respiratory health and air pollution in Iceland.
1
1.1 Air pollution
1.1.1 Traffic pollution
In later years, abandoning of coal for space heating and the removal of lead
from gasoline has improved air quality in urban areas (Brunekreef and
Holgate, 2002). Motor traffic exhaust is the most prominent pollution source,
though local conditions, weather, building types, and topography also affect
the local pollution levels (WHO, 2013).
Gaseous pollutants from fuel combustion, mainly from traffic, contribute
greatly to air pollution in cities, for example the yellow smog clouds over
megacities like Athens and Los Angeles. These yellow clouds are acid
aerosols formed in heat-catalyzed reactions of automobile exhaust and sulfur
dioxide (SO2). SO2 is released from burning coal and other poor quality fossil
fuels. The burning of coal has been banned in many cities, other fuels are
cleaner, and SO2 has currently become less of a problem in the Western
world (Fenger, 2002; WHO 2005).
Traffic exhaust contains a large number of gases and particles which have
known health effects (Fenger, 2002, WHO, 2005). Nitrogen dioxide (NO2) is
used as an indicator for traffic pollution as it correlates well with traffic counts
when measured at the roadside. In some studies other measures of pollution
such as black carbon (BC) or black smoke (BS) are used as proxies for traffic
proximity as they also correlate well with traffic counts (WHO, 2013).
NO2 is a precursor for ozone (O3) which forms when automobile exhaust
gases, including NO2, react with the atmosphere, consuming the NO 2 to form
O3. This reaction is catalyzed by heat and sun radiation (Jenkins and
Clemitshaw, 2002). O3 levels are often lower in the presence of traffic during
the day (see Figure 1). Volatile organic compounds (VOC) and polyaromatic
hydrocarbons (PAHs) are also formed when hydrocarbon fuels are
combusted. The availability of NOx and VOC determine how high the O 3
concentration can become as described by Jenkins and Clemitshaw (2002).
In the presence of daylight (solar ultra-violet radiation, hv), the reactions
occur on a time scale of seconds according to
NO2 + hv → NO + O
Atmospheric oxygen, O2, will quickly react with the leftover O molecule
(Sillman, 2002).
2
Even though O3 tends to decrease in the presence of traffic, background
levels of O3 have largely followed the trends in fossil fuel emissions since
systematic measuring began in the 20 th century (Derwent et al., 2007;
Solberg et al., 2005).
NO2 and O3 four days
NO2
0
20
microg/m3
µg/m3
40 60 80
100
O3
1AM
1PM
1AM
1PM 1AM
1PM
1AM
1PM
1AM
Figure 1 Concentrations of NO2 and O3 at an urban roadside measuring station,
Reykjavík, over 4 days starting January 21, 2006.
Particulate matter
Particulate matter (PM) is in essence small particles that are suspended in
the air. Traffic-related PM contains a mix of materials from combustion
processes: carbon and sulfates, but also particles from soils, or mechanical
wear, mineral components and traces of heavy metals (WHO, 2005; Bérubé
et al., 2006; Donaldson et al., 2006). PM is denoted by the aerodynamic
diameter of the particles; the coarse fraction is PM10-2.5 (PM with an
aerodynamic diameter between 2.5 and 10 µm), the fine fraction is PM 0.1-2.5
(aerodynamic diameter between 0.1 and 2.5 µm), and the ultrafine fraction;
PM0.1 (aerodynamic diameter less than 0.1 µm) (Bérubé et al., 2006;
Skúladóttir et al., 2003; Akselsson et al., 1994, p 78).
In some studies, there is a distinction between primary particles or
aerosols directly emitted from the source, and secondary particles or
aerosols which are formed when primary pollutants react with the
atmosphere, or each other to form other components or larger particles. The
size fraction indicates the origin of the particle. The ultrafine fraction is formed
in anthropogenic processes, like engine combustion, whereas the fine and
coarse fraction particles are made up of primary natural aerosols or formed in
3
mechanical processes, for example from the wear of car tires on asphalt
paving (Akselsson et al., 1994; WHO, 2013).
Cocktail effects
Figure 2 Reykjavík seen from the north on a day with high levels of PM due to a dust
storm (top) and on a day with clear skies (below) (image: Hanne Krage Carlsen).
Most research has focused on identifying the health effects of single
pollutants, while adjusting for the effects of other pollutants. However,
cocktail effects are known from research in other areas of toxicology and are
beginning to appear within the field of air pollution studies. Recent research
methodology moves in the direction of a more holistic approach using
interaction models to appreciate these effects (Dominici et al., 2010), but only
a few studies have applied these methods so far, and with conflicting results
(Bobb et al., 2011; Yu et al., 2013; Anderson et al., 2012).
4
1.1.2 Dust and sand storms
Desert regions or areas with little vegetation are prone to erosion from both
wind and precipitation. Dust particles are lifted from the surface by the wind.
The size, weight and physical properties of the particles determine how easily
they are uplifted. Spores from fungi or bacteria have been found on the
surface of natural particles. These may affect ecosystems far from their home
regions, or work as allergens (Kellogg and Griffin, 2006). There are examples
of Asian dust storms that drift around the globe with the wind (Uno et al.,
2009).
Globally, the major dust contributors are the Sahara and Gobi Deserts
(Kellogg and Griffin, 2006). In recent years dust from Iceland´s glacial
outwash plains has been recognized as a significant contributor to dust in the
North Atlantic region (Prospero et al., 2012; Bullard, 2013).
1.1.3 Geothermal areas and hydrogen sulfide
Hydrogen sulfide (H2S) is a clear gas with the characteristic smell of rotten
eggs. H2S is abundant in geothermal areas. Geothermal areas with hot
springs, mud volcanoes and geysers are the surface manifestation of an
underlying cooling magma chamber. H2S is released from the harnessing of
geothermal energy (figure 6). H2S is also found as a biproduct of drilling and
refining of oil and gas and of some industrial processes like paper production
(Campagna et al., 2004).
1.1.4 Volcanic eruptions
Volcanoes are the surface manifestations of Earth’s inner thermal processes.
Solid, liquid or gaseous products, magma, from the molten part of the core
break through the earths crust and disperse as either lava, tephra, or ash.
The world's most active volcanoes are distributed along continental plate
margins (e.g. Japan, the Andes, and Indonesia in the Ring of Fire) or over hot
spots or mantle plumes that rise up from the underlying mantle (e.g. Hawaii,
Réunion). Iceland represents a mixture of the two forms of volcanism as it is
located on the oceanic plate boundary between the North American and
Eurasian plates above a mantle plume.
Volcanic ash refers to the small particles (<2mm in diameter; particles
larger than 2 mm are called tephra) emitted from volcanic eruptions. Volcanic
ash forms in eruptions of a certain magnitude as explosive energy is
necessary to produce and emit the small particles. In particular when water is
present, explosive energy is released from the molten rock to the water, and
the rock explodes into small fractions. Ash particles can have very different
5
characteristics depending on the chemical properties of the magma and the
mode of eruption (Francis and Oppenheimer, 2004). During a volcanic
eruption, fine ash remains suspended in the air and is dispersed via the
eruption column. Volcanic ash can be characterized as PM (see figure 3).
Fresh, unweathered ash particles can carry other corrosive chemicals on
the surface and are likely to have a larger surface area than dust particles
from other sources (Gíslason et al., 2011).
Figure 3 The ash plume of Eyjafjallajökull May 7 2010. The ash contained as much
as 20% PM10 (image: Hanne Krage Carlsen).
1.1.5 Other sources
The study of indoor air pollution is a field in its own right. In the Western
world, the main indoor air quality concerns include smoking, mold and
spores, particles produced during cooking, and chemical exposures e.g. from
paint. Burning biomass fuel (coal, wood, peat and dried manure) inside for
heat or cooking has been largely abandoned in the Western world, but
remain a major concern in the developing world where indoor combustion
particles constitute the second largest environmental risk factor after smoking
(Lim et al., 2012). Wild fires in nature are also a significant source of particles,
both locally and regionally (Emmanuel, 2000; Naeher et al., 2007; Künzli et al.,
2006).
6
1.2 Air pollution effects on health
1.2.1 Pathways
The main effects of air pollution on the body are found in the cardiopulmonary
system. The respiratory system is most exposed as humans breathe circa 20
m3 of air per day. The inhaled air enters the nose or mouth, goes into the
bronchi, the bronchioles and the alveoli, where inhaled oxygen (O 2) is
exchanged through the epithelium of the alveolar wall. From there O 2 goes
into the bloodstream, which delivers it to the cells in the body. Carbon dioxide
(CO2) from the cells is exchanged in the opposite direction, to the exhaled air
(WHO, 2000 p 16).
Figure 4 Overview of the respiratory system and particle deposition (image: Hanne
Krage Carlsen/Creative Commons).
The nose works as a barrier for larger particles which impact inside the
nose or get caught by the nasal hairs (Eccles, 2006) but particles smaller
than 10 µm can enter the respiratory system along with air and gases.
Further into the airways, particle deposition depends on the respiration rate,
particle sizes and chemical and physical properties of the particles, but only
the smallest particles (PM<1 µm) can reach the alveoli and become
deposited there (figure 4) (Akselsson et al., 1994). Once in the respiratory
system, the air pollution affects first the lungs, and in turn, the cardiovascular
7
system (see figure 5). Research in animals indicates that the smallest
particles also reach other organs (Kreyling et al., 2010).
Adverse effects from exposure to air pollution can be more severe in
susceptible individuals. These are often considered to be the very old and
very young, or individuals with a pre-existing condition which makes them
more sensitive to exposure to air pollution, such as obstructive lung disease,
which is described below.
Obstructive lung disease
Asthma affects a substantial proportion of the world’s population; the rates
are highest in the Western world, especially among Europeans (To et al.,
2012). Asthma is an inflammatory disorder where the airway walls react with
bronchial hyperresponsivity (BHR), restricting the diameter of the airways,
and patients experience difficulty exhaling, called obstruction (Holgate et al.,
2010). Wheezing, dyspnea, chest tightness and cough are common
symptoms (D’Amato et al., 2005), but asthma has a number of clinical
expressions (Janson et al., 2008). In allergic asthma, BHR occurs as a
reaction to stimuli such as allergens. Asthma management revolves around
quick-relief medications such as short acting β-agonists and longer acting
medications like corticosteroids. Asthma is a heterogeneous disease which
can be exacerbated by infectious disease epidemics, climate, and exercise,
but more importantly, exposure to allergens and air pollution (von Mutius and
Braun-Fahrländer, 2002). In asthma, some air pollution types can act directly
as irritants to the airways, but another possible mechanism suggests that air
pollution damages the cells in the epithelial wall of the alveoli, inducing an
inflammatory response, which increases the penetrability so that PM and
harmful gases can pass through into the blood serum (von Mutius and BraunFahrländer, 2002; Holgate et al., 2010). Asthma and chronic obstructive
pulmonary disease (COPD) share a number of risk factors such as a family
history of asthma, but COPD primarily develops in adults with a history of
smoking (de Marco et al., 2007). COPD can have a wide range of clinical
expressions, but in general the alveolar tissue breaks down, and in patients
with COPD increased levels of several inflammation markers were
associated with worse lung function after adjusting for other risk factors
(Thorleifsson et al., 2009); other studies in healthy individuals found stronger
effects in men than in women (Ólafsdóttir et al., 2007; Ólafsdóttir et al.,
2011).
The COPD diagnosis is based on results from lung function tests,
specifically the ratio between the Forced Expiratory Volume in 1 second
8
(FEV1) and Forced Vital Capacity, FVC (FEV1/FVC). Staging of COPD is
based on how low the FEV1 is. COPD is a rather common disease in older
smokers. A wide variation in prevalence was found in an international study
by Buist and colleagues (2007) and poverty and socioeconomic status also
play an important role (Burney et al., 2013).
Many COPD patients experience exacerbation with irritation symptoms
when exposed to irritants, such as high air pollution levels (Yang et al., 2005;
Halonen et al., 2008).
Figure 5 Suggested biological pathways for health effects of air pollution to
different parts of the circulatory system (image: Hanne Krage
Carlsen/Creative commonts).
9
Cardiovascular disease
Both long-term and short-term exposure to air pollution has been associated
with cardiovascular health effects of air pollution (Pope and Dockery, 2006).
The processes leading to a cardiovascular event like a cardiac infarction,
such as building up of plaque in the blood vessels, can take decades.
However, the mechanisms that trigger cardiovascular events in a susceptible
population may work in a matter of days or hours. The risk factors for the two
processes may not be the same (Künzli and Tager, 2005; Künzli et al., 2011).
Several pathways and mechanisms have been suggested for these health
effects (Sun et al., 2010). First, the immune system responds to the air
pollutants in the respiratory system. Inflammation biomarkers are increased
after exposure to air pollution (Delfino et al., 2009). Inflammation is
associated with accelerated formation of atherosclerotic plaque (Künzli et al.,
2011), and increased blood viscosity, causing thrombosis (Pope and
Dockery, 2006; Gold and Mittleman, 2013). A destabilized atherosclerotic
plaque can disrupt in the cardiac artery wall and start the formation of a clot
which obstructs the blood flow to the heart’s muscles and cause a cardiac
infarction.
Other suggested pathways for effects of air pollution include disturbance
in the cardiac autonomic function which controls the heart rate, and in the
vasomotor function that controls the widening and narrowing of blood vessels
(Sun et al., 2010).
Nearly all cardiovascular risk factors also increase the long-term risk of
developing stroke. On a shorter time scale, increases in blood viscosity and
clotting factors, along with disturbances in the heart’s autonomic function,
also increase the risk of ischemic stroke (Gold and Mittleman, 2013).
Disturbances in vasomotor function and damage to the blood vessels may
increase the risk of hemorrhagic stroke after short-term increases in air
pollution exposure levels (Yorifuji et al., 2011).
Other health outcomes
Exposure to natural hazards like volcanic eruptions is associated with
increased risk of stress, depression and fatigue (Shore et al., 1986), some
factors probably related to the loss or damage to property or livestock, crop
failure and general insecurity imposed by living near an active volcano (Tobin
et al., 2012). However, exposure to natural disasters is generally less likely to
result in post traumatic stress disorder (PTSD) than exposure to acts of
terrorism or war (Bracha, 2006).
10
In recent years more associations between chronic exposure to air
pollution and health effects outside the cardiovascular system have been
found, such as the development of dementia (Forsberg et al., 2013), and
giving birth to a child with autism spectrum disorders (Volk et al., 2014).
However, these findings are novel and it is uncertain whether they represent
correlation rather than causation.
1.2.2 Evidence from epidemiological studies
Traffic-related pollution
Short-term changes in levels of traffic-related air pollutants like PM10 are
associated with increased mortality rates (Meister et al., 2012) and
respiratory disease hospital admissions (Peng et al., 2008). Short-term
exposure to increased levels of fine particles, PM2.5, is associated with
increased hospital admissions for cardiovascular events (Pope et al., 2006;
Pope et al., 2008).
Long-term exposure, or living in an area with higher levels of total PM and
fine PM, is associated with higher mortality rates and reduced life expectancy
(Dockery et al., 1993; Pope et al., 2009), and higher risk of a cardiac
infarction attack in women (Miller et al., 2007). Exposure to air pollution from
traffic is associated with stunted lung growth in children (Gaudermann et al.,
2007).
Exposure to short-term increases in traffic pollution, PM10, O3, SO2, NO2,
exacerbates already existing asthma in adults and children (Antó et al., 2012;
Samoli et al., 2011) and is associated with increased emergency room visits
and hospital admissions for asthma (Villeneuve et al., 2007; D´Amato et al.,
2005), and COPD (Sunyer et al., 2000; Yang et al., 2005). Long-term
exposure to traffic pollution in urban areas is associated with increased risks
of developing asthma in children (Gaudermann et al., 2007; Bråbäck and
Forsberg, 2009; McConnell et al., 2010), but whether adults are also at
increased risk is debated (Anderson et al., 2013; Jacquemin et al., 2012).
Short-term increases in air pollution levels are associated with increased
mortality risk in COPD patients (Sunyer et al., 2000) and an increased risk of
being admitted to hospital (Yang et al., 2005). Long-term exposure to trafficrelated air pollution is associated with increased risk of hospitalization for
COPD (Andersen et al., 2011) and development of COPD in women
(Schikowski et al., 2014).
11
A short-term increase in PM2.5 exposure is associated with increased risk
of cardiovascular events and cardiovascular mortality (Brook et al., 2010; Sun
et al., 2010; Gold and Mittleman, 2013). Short-term exposure to gaseous
pollutants is associated with adverse changes in heart rate, arrhythmia
indicators (Devlin et al, 2003) and inflammatory markers (Sun et al., 2010;
Gold and Mittleman, 2013; Chuang et al., 2007). Long-term exposure to
higher PM2.5 levels at the residence is associated with increased risk of
developing hypertension which is associated with other adverse cardiac
outcomes (Chen et al., 2013).
Short-term exposure to PM2.5 and NO2 is associated with increased risk of
stroke (Yorifuji et al., 2011; Gold and Mittleman, 2013). Studies of exposure
to short-term changes in O3 show only associations with ischemic stroke in
men (Henrotin et al., 2007; Henrotin et al., 2010). Long-term exposure to
PM10 also increases the risk of stroke (Oudin et al., 2010), whereas results
regarding exposure to background NO 2 are not in agreement (Oudin et al.,
2009; Andersen et al., 2012).
Mineral dust and sand storms
Dust storms are considered a natural source of pollution, though erosion, the
main cause of dust storms, is facilitated by human activities (Morman and
Plumlee, 2013). As the health effects of the particles are highly dependent on
the origin, these studies are presented by region.
In Anchorage, Alaska, US, high levels of PM from geological sources,
some from a volcanic eruption, have been associated with increases in
bronchitis, upper respiratory infection and asthma medical visits (Choudhury
et al., 1997; Gordian et al., 1996). People under 20 had increased risk of
outpatient asthma visits and anti-asthma medication (beta-agonists) and
same day, and weekly median exposure to PM pollution (Chimonas and
Gessner, 2007). School nurse anti-asthma medication dispensing was
associated with PM levels in the preceding days (Gordian and Choudhury,
2003). Studies of dust storms in the desert regions of the US in the 1930s
have shown that there was an association between dust events and
hospitalizations for respiratory illness (Morman and Plumlee, 2013; Hefflin
and Jalaludin, 1994). Dust storms in Sydney, Australia were associated with
a 15% increase in mortality (Johnston et al., 2011).
Asian dust storms have their origin in the desert regions of China and
Mongolia and contribute to pollution in the Asian mega-cities (Chan and Yao,
2008); the dust is carried across the Pacific Ocean and has been measured
12
in the northern US (Duce et al., 1980). Associations have been found
between Asian dust storms and increased mortality in Taipei and Korean
cities (Chan and Ng, 2011; Lee et al, 2013) and increased risk of emergency
hospital admissions for COPD (Tam et al., 2012).
Saharan dust particles have been associated with increased risk of
pediatric asthma admissions, hospital admissions for meningococcal
meningitis, adverse birth outcomes, and cardiopulmonary disease mortality in
large cities in southern Europe (Karanasiou et al., 2012). Most notably, Pérez
and colleagues (2012) found that the increase in mortality was higher when
Saharan dust was present in the PM pollution, than when the PM had a
traffic-related origin (Pérez et al., 2012).
Hydrogen sulphide
The toxic effects of H2S in occupational settings are well known (WHO,
2000). H2S works as a signalling molecule in the human body and is involved
in several organ systems, notably the respiratory system (Olson, 2011), but
the mechanisms are not yet known.
Figure 6 The emissions from a geothermal power plant contain H2S (image: Hanne
Krage Carlsen).
Industrial sources
A high H2S level from a wastewater treatment facility was associated with
increased risk of asthma hospital visits the following day (Campagna et al.,
2004). Long-term exposure to H2S emitted from mostly industrial sources (oil
refinery, cheese plant, natural oil and gas wells, and a sewage plant) was
associated with adverse neuropsychological effects in residents around the
13
plants (Kilburn and Warshaw, 1995; Kilburn et al., 2010). Living in areas
exposed to H2S from pulp- and sawmills is associated with increased sensory
organ irritation symptoms and headaches (Jaakkola et al., 1996; ParttiPellinen et al., 1996; McGavran, 2001). A 2003 report (Roth and Goodwin,
2003) called for better reporting of exposure and more studies of chronic,
low-level exposure and studies on asthmatics.
Geothermal sources
The epidemiological studies of respiratory effects from exposure to H 2S from
geothermal sources have conflicting results (McGavran, 2001).
Exposure to volcanic gases, H2S among them, from geothermal areas and
volcanoes that are not erupting is associated with incidences of respiratory
cancers and respiratory illness, and respiratory mortality but not childhood
asthma or decreased lung function. Furthermore, most studies have a
moderate or high risk of bias because of the study design or methods of
surveying; e.g. self-reporting can be very biased by beliefs about the health
risks of exposure. Also, many studies lack information about important
confounders (Hansell and Oppenheimer, 2004)
Short-term exposure to increased H2S levels is not associated with
dispensing of medication for angina pectoris the following 7 days
(Finnbjornsdottir et al., 2013). In a trial where healthy volunteers were
exposed to H2S, only minor sensory and cognitive effects were found (Fiedler
et al., 2008).
Studies of long-term effects of exposure to background sulphur gases
from geothermal sources in Hawaii, New Zealand and the Azores, H 2S
among them, indicated that long-term exposure is associated with adverse
cardiopulmonary effects, both measured and self-reported (Durand and
Wilson, 2006; Longo et al., 2008; Longo and Yang, 2008; and Longo, 2009)
and increased rates of bronchitis (Amaral and Rodrigues, 2007; Longo et al.,
2008). No association was found between chronic exposure to H 2S in
residents in a geothermal field and self-reported asthma symptoms (Bates et
al., 2013). In the same cohort, no neuropsychological effects were found
(Bates, 2013).
Volcanic eruptions
Human dwellings are clustered around volcanoes in many parts of the world,
since volcanic soils can be very fertile and humans also utilize the warm
water and geothermal heat. Worldwide, 500 million people live within 100 km
14
of an active volcano (Small and Naumann, 2001). In Iceland, the
corresponding number makes up 95% of the population. During the 20th
century nearly 100,000 people were killed directly by volcanic eruptions and
more than 4.5 million people were affected or displaced (Doocy et al., 2013).
Volcanic eruptions pose immediate dangers to human health when
pyroclastic flows and lahars (mud flows) reach inhabitated areas (Hansell et
al., 2006; Heggie, 2009). Active volcanoes emit various sulfur compounds
like SO2 and H2S, CO2, and acid aerosols. Exposure to volcanic aerosols
from African, and West Indian volcanoes was associated with asthma and
COPD exacerbations in studies from before 2004 (Hansell and Oppenheimer,
2004).
Gases from volcanic eruptions
The residents of the village Miyake, Japan, were forced to evacuate from
their homes when Mt. Oyama erupted in 2000 because of the dangerous or
lethal levels of gases (Iwasawa et al., 2009). The residents began to return in
2005, and groups of volunteers came with them to help them re-establish the
community even though the volcano still emitted large amounts of SO 2. The
adult residents (n=823) were surveyed in 2004 before they began to return to
their homes and again after their return to the island in 2006. After returning,
the adults had no deterioration in lung function, but the prevalence of cough
and phlegm was higher. Those in the areas with higher levels of SO 2 had
more symptoms (Iwasawa et al., 2009). The self-reported incidence of cough,
other throat symptoms and breathlessness showed a dose-response
association with SO2 exposure in healthy volunteers working in the areas.
Women were generally more susceptible (Ishigami et al., 2008). In a followup 6 years later, those living in the areas with higher levels of SO 2 were
associated with lower than expected lung function in adults (Kochi et al.,
2013) and irritation symptoms in children (Iwasawa et al., 2013).
There are few studies of short-term exposure. A 2009 increase in volcanic
activity lasting 14 weeks in Hawaii’s Kilauea volcano provided the frame for a
natural experiment, following the increase in emissions, which were as high
as three-fold; there was an increase in respiratory complaints and
headaches, especially in the elderly and native Hawaiians (Longo et al.,
2010). In a review article summarizing results from previous studies, Longo et
al., (2009) also report that some 35% of 335 people surveyed in a
questionnaire study found their health to be affected by the eruption. Current
and former smokers and people with chronic respiratory disease were most
15
affected. In a follow-up in 2012, exposed individuals reported a more than 15fold increase in irritation symptoms while being outdoors (Longo et al., 2013).
Volcanic ash
Volcanic ash (figure 7) can damage vegetation and cause buildings to
collapse and injure or kill people (e.g. due to bad visibility, figure 7).
Inhalation of ash, skin abrasion and eye irritation are the main long-term
concerns to human health. Though fine volcanic ash can be considered a
form of PM10, it is not certain if the dose-response functions which have been
developed from traffic-related PM apply to volcanic ash (Oudin et al., 2013).
Both the physical and chemical characteristics of volcanic ash vary greatly
between volcanoes and eruptions (Horwell and Baxter, 2006).
Figure 7 Ash from the Eyjafjallajökull eruption causes bad visibility mid-day May 7,
2010 (image: Hanne Krage Carlsen).
The short-term health effects of exposure to volcanic ash range from none
to region-wide epidemics of asthma and respiratory symptoms. A similar
range was found in the clinical and toxicological studies. The major effects
seem to be short-lived and depend very much on the properties of the ash
(Horwell and Baxter, 2006).
After a 28-day long eruption of Guagur Pichincha in Ecuador in 2000,
Naumova et al. (2007) studied the pediatric hospital emergency room (ER)
16
visits in the city of Quito. In the weeks following the April eruption there was a
significant increase in diagnosed primary respiratory conditions. The risk of
lower respiratory infection was doubled and for upper respiratory infection,
the risk was increased 70%. The effect was larger in small children and the
rate of asthma diagnoses doubled.
Following a large eruption of Mt. Ruapehu in New Zealand in 1996, ash
was detected in the air near large cities 200-300 km from the volcano.
Hospital records were later examined and researchers found the highest
rates of respiratory mortality for the 1990's for these cities during one month
following the eruption. Researchers concluded that the ash was responsible
for the excess respiratory mortality (Newnham et al., 2010).
Shimizu et al. (2007) surveyed 236 adult asthma patients from an area
exposed to ash from an eruption of Mt. Asama in Japan. In the most exposed
region, where more than 100g/m2 ash fell, about half of the patients with
asthma reported worsening, their spirometry (lung function) measurements
were lower and they used more medication.
Mental health
The natural forces unleashed during a volcanic eruption mean that
experiencing a volcanic eruption can be a traumatic event as individuals fear
for their own lives and the lives of their neighbors.
The mental health effect following the eruption of Mount St. Helens
included increased ER visit rates, increased use of mental health services
and domestic violence, also increased self-reported stress and mental health
problems in those who had lost a friend or relative in the disaster; and up to
ten-fold increases in levels of anxiety, stress and major depression in people
who had experienced damage to property (Shore et al., 1986) Studies of
survivors of a deadly eruption of Mt. Unzen, Japan, had very high rates of
adverse mental health which subsided with time (Ohta et al., 2003). Societies
near volcanoes with catastrophic eruptions who are exposed over many
generations incorporate stories of the disasters into their myths and
cosmology which help future generations understand and recover in case of
disaster (Cashman and Cronin, 2008).
1.3 Respiratory health and air pollution in Iceland
Historically, asthma and allergic disease prevalence have been low in Iceland
(Gislason et al., 2002), but both asthma symptom prevalence and use of
asthma medication increased in adult Icelanders between 1990 and 2007.
However, it is unclear if this increase has been due to changes in diet,
17
smoking, higher diagnosis rates, or other factors (Sigurkarlsson et al., 2011).
The prevalence of COPD in Iceland was found to be 18% in individuals over
40 years of age (Benediktsdóttir et al., 2007),
Only a few studies have addressed the association between air pollution
and human health in Iceland. First, The European Community Respiratory
Health Study (ECRHS; Burney et al., 1994), which Iceland has participated in
for a number of years (Gíslason et al., 2002), showed that chronic bronchitis
and phlegm increased in participants exposed to high levels of NO 2 at their
residence and who lived close to major roads (Sunyer et al., 2006). A study
using pharmaco-epidemiological methods found associations between
dispensing of angina–pectoris relief medication and levels of O 3 and NO2
(Finnbjornsdottir et al., 2013).
Iceland's capital area has a reputation for being among the cleanest
capitals in the world. There is little industrial pollution and geothermal energy
has replaced the use of fossil fuels for space heating. There is considerable
traffic-related air pollution (Jóhansson, 2007; UHR, 2007).
The capital area of Iceland covers around 700 km2 (National Land Survey
of Iceland, 2012) and most of the city is characterized by low houses usually
two to three stories high. As of 2011, there were nearly 200 000 inhabitants
of which 150,000 were old enough to have a driving license (18 years of age,
Statistics Iceland, 2013a).
The air pollution regulation is based on health limits in the EU (Table 1;
Althingi, 2002; Altingi, 2003), except for H2S. H2S Health Limit Regulations
were implemented in Iceland in 2010 (Althingi, 2010) and were based on
regulation from other locations with occurrence of H 2S, such as Hawaii
(Department of Health, Hawaii, 2001).
In the capital area, there were 134,000 passenger cars in 2006 (newest
available, Statistics Iceland, 2013b), which makes 9 cars per 10 adults, one
of the highest car to people ratios in the world (Laych et al., 2009). Car use is
an ingrained part of Icelandic culture (Collin-Lange, 2013), and the
infrastructure of the capital area is very car-oriented.
The levels of NO2, a traffic indicator pollutant, are significant around major
roads (Sunyer et al., 2006). O3 levels in Iceland tend to be higher in early
spring, which is believed to be related to atmospheric conditions in the Arctic
(Solberg et al., 2005). PM10 levels in Iceland’s capital area are mainly trafficrelated, so particles are resuspended during drier periods (UHR, 2007;
Johansson, 2007). Another contributor to PM is vehicles which are driven
18
with studded tires during winter. The tire studs wear the asphalt into small
particles which mix with the salt and sand which is spread on roads and
pavements during winter to prevent skidding.
Table 1 Air pollution health limits in Iceland (µg/m3).
Act no./Year
24-hour
µg/m3
H2S
514/2010
50
NO2
251/2002
75
O3
745/2003
120
PM10
251/2002
50
a
1-hour
µg/m3
Allowed
exceedencesb
5
110
a
175
25
7
a
Maximum 8 hour mean. b Allowed exceedences is a target for the number of
times the health limit may be surpassed, 2010 is given as an example.
1.3.1 Mineral dust sources
During the warmer season, dust storms originating in South Iceland areas
with scarce vegetation are a more common source of PM 10 than urban
sources. Twenty-two % of Iceland has little or no vegetation and is
characterized as an arctic desert. Dust storms are frequent in the highlands
(Dagsson-Waldhauserowa et al., 2013). Dust storms in Iceland originate in
the sand plains of the south and affect the air quality of the capital area to the
west (see figure 2; Thorsteinsson et al., 2011). The dust is mainly basaltic
glass (Arnalds, 2010) and dust storms from the east/south-east now include a
substantial contribution of ash from the Eyjafjallajökull eruption (Arnalds et
al., 2011; Arnalds et al., 2013; Thorsteinsson et al., 2011).
1.3.2 Hydrogen sulfide
H2S is emitted naturally, but with the opening of the geothermal harnessing
sites at Hellisheiði and Nesjavellir, 25 and 35 kilometres from the capital, the
levels have risen substantially (Ólafsdóttir and Garðarsson, 2013; Ólafsdóttir
et al., 2014).
H2S levels measured in the city have not been found in levels which are
known to be toxic from studies of occupational exposure, but they routinely
surpass the WHO nuisance guideline value of 7µg/m3 (WHO, 2000) and
Iceland’s own H2S regulation.
19
Figure 8 Hellisheidi geothermal powerplant seen from the north (image: Hanne Krage
Carlsen).
1.3.3 Icelandic volcanoes
Iceland experiences a volcanic eruption every 5 years on average. Most are
small and have little impact beyond the very local environment (Thordarson
and Höskuldsson, 2002). Iceland is thankfully rather sparsely populated, so in
most cases, there is no danger to human life or health, but there are a few
exceptions which will be discussed here.
Laki/Skaftáreldar 1783-1784
The 1783-4 eruption of the Laki fissure was the most fatal in historical times
in Europe and produced the largest lave flow on Earth in historical times.
First, the acid gas and aerosol fog from the eruption and subsequently
degassing lava led to an atmospheric haze that persisted into 1786 (Lacy,
1998) and caused an increase in respiratory mortality, first and foremost in
Iceland and northern Europe. The sulfur haze lay over Europe and even Asia
and North America as well, where the gases reduced the sun-light and
caused crop failure. In Iceland, at least half of the livestock died off, and onefifth of the population died (Hansell and Oppenheimer, 2004; Tweed, 2012).
In Europe and Alaska famine disasters were described, caused by colder
climate and acid rain (Jacoby et al., 1999; Durand and Grattan, 2001; Hansell
and Oppenheimer, 2004). In England, 20,000 were estimated to have died
from the indirect effects of the Laki eruption (Witham and Oppenheimer,
20
2004). If a similar eruption happened today, it is estimated that it would cause
hundreds of thousands of deaths (Schmidt et al., 2011).
The 2010 Eyjafjallajökull eruption
The Eyjafjallajökull stratovolcano in South Iceland had been dormant since
an eruption in 1821-23 (Figure 9). After years of increasing seismic unrest,
the Eyfjafjallajökull volcano erupted on March 20, 2010 in a flank eruption
with basaltic effusive lava fountains. A magma intrusion triggered the April 14
explosive eruption of the summit crater which produced large amounts of fine
ash (Sigmundsson et al., 2010). In the first hours, the main concern was flood
risks, as the glacier melted quickly and the water came down the volcano into
the Markarfljót River. Inhabitants of low-lying areas were evacuated.
Figure 9 The Eyjafjallajökull volcano 2008 (left) and 2010 (right) (image: Hanne Krage
Carlsen).
The ash dispersed over a wide area with prevailing westerly winds. The
ash was fine during the first eruption phase, with about 20% 10 µm or less in
aerodynamic diameter (Institute of Earth Science, 2010; Thorsteinsson et al.,
2011; Stohl et al, 2011; Gislason et al., 2011; Gudmundsson et al., 2012).
The ash particles from Eyjafjallajökull had a large surface area, and had a
potential to increase inflammation markers (Horwell et al., 2013). The ash
had a high iron (Fe) content, and was found in cytotoxicological studies to
have effects on alveolar macrophage function which led to increased
bacterial growth in vitro (Monick et al., 2013). Ash from other Icelandic
volcanoes also has high iron content (Dagsson-Waldhauserowa et al., 2013).
21
In the period after the eruption some of the most extreme wind erosion
events on earth were recorded in South Iceland (Arnalds et al., 2013). PM
dust levels were higher in the area surrounding the volcano and in the capital
Figure 10 Composition of ambient PM10 in Reykjavík 2003 and 2013 (Image: Hanne
Krage Carlsen adapted from Skúladóttir et al., 2003 & Vegagerðin, 2013).
area more than 100 km from the volcano (Thorsteinsson et al., 2011;
Thorsteinsson et al., 2012). Volcanic ash is now a significant contributor to
PM10 in the capital area (Vegagerðin, 2013, Figure 10).
As we have seen, the population of Iceland is exposed to significant air
pollution both from traffic-related air pollution and pollution from natural
sources such as geothermal harnessing and volcanic eruptions. The health
effects of these remain un-investigated.
22
2 Aims
The aim of this project was to determine if, and to what extent, residents in
Iceland experience adverse health effects from air pollution. The target
populations are the residents of the Icelandic capital area, who are mainly
exposed to traffic-related air pollution, and the residents of South Iceland,
who were exposed to volcanic ash from the 2010 Eyjafjallajökull eruption.The
overarching research question of the articles is: do the residents of the capital
area or South Iceland experience adverse health effects due to air pollution
from traffic or urban sources, or the volcanic sources of ash or H 2S?
Specific aims for each paper:
I.
Were the daily number of individuals obtaining prescribed antiasthma medication in Reykjavík associated with changes in daily air pollution
levels the previous 14 days during the period 2006-2009?
II.
Were 2003-2009 emergency room visits and acute hospital
admissions (emergency hospital visits) for cardiopulmonary disease and
stroke, associated with daily air pollution levels in Iceland's capital area.
Were any subgroups (elderly, women) more vulnerable than others to the
exposure?
III.
Were the association between emergency hospital visits and PM 10
modified by the source of PM10 pollution (natural dust, volcanic ash or other),
or the introduction of volcanic ash from Eyjafjallajökull or Grímsvötn? Also,
was there an association with H2S levels?
IV.
Was the health of 207 individuals exposed to the Eyjafjallajökull
volcanic eruption surveyed with lung function measurements, questionnaires
and physician's evaluation associated with adverse health effects?
V.
Was exposure to the eruption of Eyjafjallajökull associated with
increased rates of respiratory- and mental health symptoms six months after
the eruption in the population of South Iceland compared to a
demographically matched reference group in North Iceland? Did the reported
symptom rate correspond to the level of exposure within South Iceland?
23
3 Materials and methods
The studies in this thesis fall into two categories, on one hand studies of
short-term effects where time series methods and register-data are used to
study associations between pollution levels and a health outcome. On the
other hand, there are studies of chronic exposure in South Icelanders
exposed to the Eyjafjallajökull volcanic eruption; these use survey or
questionnaire data and employ logistic regression and descriptive statistics to
analyze the data.
The following section first lists the data sources used and types (see table
2), followed by presentation of the methods used in the two types of studies
(see table 4).
3.1 Data sources
3.1.1 Exposure
Air pollution in Reykjavík
The air pollution exposure variables used in papers I, II, and III were routinely
measured air pollution levels in Reykjavík. The data were measured at one
site in the geographical centre of the town (Figure 11). Data from other
measuring stations were not of sufficient quality and were not used.
Figure 11 Reykjavík and surrounding area with main pollution sources and measuring
stations (Adapted from paper II).
25
The data were collected, validated and corrected and provided as 30 minute
or 60 minute means by the Environmental Branch of the City of Reykjavík
(data for 2003-2008) and the Environmental Agency of Iceland (data from
2009 and later).
We chose four pollutants which have known health effects and which are
indicative of the pollution mix in Reykjavík; PM10, NO2, O3 and H2S. H2S
measurements only began in February 2006 and ozone measurements were
incorrect from 2010 and onwards (see Figure 15). As covariates, we included
weather and pollen counts.
In addition, we obtained information about sources of PM 10 from the
Reykjavík Municipality Environmental Branch and Dr. Thorsteinsson and
Rebekka Kienle, both University of Iceland (personal communication,
Thröstur Thorsteinsson, 29 oktober, 2012).
Eyjafjallajökull
Exposure to the Eyjafjallajökull volcanic eruption (papers IV and V) was multifaceted as the population was exposed first to hazards in the form of flood
risk, and after the immediate flood risk subsided the ash fall became the
major concern. At the same time noise from the explosion of the eruption
column could be heard hundreds of kilometers away in some directions.
Earthquakes were frequent during the eruption, as is to be expected. The
degree of exposure to the ash was estimated from knowledge of the ash
plume thickness and wind direction during the eruption (Figure 12).
Figure 12 Ash dispersion around Eyjafjallajökull (left) and the exposure areas based
on it (right) (Image: Thorsteinsson, from Paper IV and V).
26
3.1.1 Health outcomes
To measure health in a population it is most useful to use an outcome
measure which is both a good indicator of the outcome of interest and which
is sensitive, meaning that it has considerable variation and correlation with
the exposure. It is necessary to have both a reliable end point and a
meaningful number of events so that the study will have sufficient statistical
power.
Papers I, II, and III are register-based studies meaning that they make use
of data which are collected routinely in registers (Wallgren and Wallgren,
2007). This method is useful in studies when large amounts of data are
needed as the data are already collected, so that the data can be gathered
quickly and cheaply. A caveat is that the data were not originally intended for
research purposes; usually there are only limited possibilities of adding
variables of interest to research. An overview of the data used in each study
can be found in table 2.
Hospital admissions
Hospital admissions are rather rare events, and this outcome is also
governed by factors which are not related to the actual morbidity, such as the
availability of hospital beds, socio-economic and geographic factors like age,
or whether the patient lives far from the hospital or has relatives living nearby.
Emergency room visits are less influenced by availability, but are rather
unspecific. The differences between hospital admissions and ER visits for
asthma and other respiratory causes and in different age groups have been
investigated by Winquist et al. (2012).
The data on ER visits and admissions used in papers I and II were
collected in the Landspitali University Hospital patient data base. The data
are registered to each individual using the personal identification number in
the database. When the data were extracted from the database, the
identification number was coded into a number in such a way that an
individual could be recognized only within the data. The data for acute ER
visits and admissions for the relevant diagnoses were extracted along with
information about gender, age, day of admission/visit, duration of admission
(if admitted), and the primary and secondary diagnoses. The address was
indicated by a postal code. The ICD codes (letter and first two digits) that
were extracted were: pulmonary outcomes: J20-J22, J40-J46 and J96;
cardiac: I20-I27, I46, I48, I50; stroke: G45-G46, I61-I69.
27
Pharmaceutical dispensing
Pharmaceutical dispensing, used in paper I, of relief medication for asthma
(R03 in the anatomical therapeutic classification (ATC) is considered an
indirect indicator for respiratory health in a population. In an American study
emergency room visits for asthma were found to be well-correlated with
dispensing of asthma medication (Naureckas et al., 2005), but this may not
be true for other populations with better access to primary care. Other papers
(Hallas, 2005) and a recent review (Menichini and Mudu, 2010) have treated
the subject in detail.
A caveat of using anti-asthma medication dispensings as an indicator is
that they are given as relief for a wide range of respiratory illnesses, so there
is little specificity to asthma. Also, these medications are often dispensed in
volumes deemed to last approximately 3 months (Gíslason et al., 1997). In a
time series study looking at short-term effects, at any given day, only a small
part of the population will actually be at risk of having to refill their
prescriptions.
Databases of prescription medicine dispensing are kept in many
countries, so data are available retrospectively. As these register the sale of
the drugs in the pharmacy, rather than when the medication is actually taken,
they are less reliable than other outcomes. The Icelandic prescription drug
database goes back to 2003 and was established for surveillance purposes
(Icelandic Directorate of Health, 2012). It has since then been used in several
research papers. From these data, a time line of the daily number of
individuals who received anti-asthma drugs was created.
Spirometry
Spirometry is a measure of lung health, where the flow and volume of
exhaled air is measured. Some of the important measures are forced vital
capacity (FVC), which is the maximum amount of air exhaled after filling the
lungs as much as possible. Forced expiratory volume in 1 second, FEV 1, is
the amount of air exhaled in the first second (GOLD, 2010), which is a
measure of airway obstruction. Spirometry values depend highly on gender,
age, and height, so most often the results are reported as percent of a
predicted value (usually the average found in a general population sample,
e.g. NHANES, Hankinson et al., 1999). Low spirometry values are indicative
of lung disease, but can also be due to respiratory infections. Spirometry was
used in paper IV.
28
Self-reported health
Self-reported health outcomes were used in paper IV and V. In
questionnaires, study participants were asked to rate their own health. Most
often these include questions if the subject has experienced easily
recognizable symptoms and to what degree.
In the studies in this thesis we have used questionnaires or questionnaire
items which have already been used in other studies, which facilitates
comparisons of the results with the results from other studies. Questions
about respiratory health were drawn from the European Community Health
Survey (ECRHS, Burney et al., 1994). The General Health Questionnaire-12
question version (GHQ-12) measures psychological morbidity which has
been validated in several populations and has been used in other studies
(Goldberg et al., 1988; Goldberg et al., 1997). GHQ had previously been
used in Iceland (Tómasson and Gudmunsson, 2009), so a translated version
was available.
Other outcomes
Mortality from any cause is an indicator of health in a population, if most
deaths are due to natural causes; this outcome is useful as an indicator for
health in a population (Logan, 1956) whereas in countries with a high
mortality due to causes other than natural, it is less useful. It is preferable to
use non-accidental mortality (that is, excluding deaths due to external
causes) or cause-specific mortality where researchers typically select
cardiopulmonary causes, which is used in many studies (Dockery et al.,
1993; Halonen et al., 2010), but as this is a rare outcome it requires a
population of a certain size to be useful, so these outcomes were not used in
these studies.
3.1.1 Covariates
Weather
In studies of short-term effects, seasonal weather changes are of importance,
as temperature and humidity influences the symptoms of asthma (Vernon et
al., 2012), but also influences the levels of air pollutants (WHO, 2013). In the
studies of long-term health effects, the change of season must be
considered, not least because respiratory infections are more common in
winter.
29
Table 2 Schematic overview of the data used in each study.
Exposure
Covariates
Weather, influenza
season, pollen
Outcome data
Daily number of anti-asthma
(R03) drug dispensing, adults in PM10, NO2, O3, H2S
the capital area
Source
The Icelandic
prescription drug
database
Complete
ness
44% (620
b
/1394)
Time
period
march
I 2006-9a
Weather, influenza
season, lag 1 of the
outcome
No
2003-9
Daily number of ER visits and
admissions for cardiopulmonary
d
PM10, NO2, O3
c
disease and stroke , adults in
the capital area
II
Landspítali
University Hospital
Admissions
database
28-54%
(1032-1961 Same as paper II
/3653)
Residence near the
volcano or in
unexposed region
Having lived or
worked near the
volcano.
PM10, NO2, H2S
Age, gender, education,
smoking status
Age category, gender,
Exposure to volcanic
phenomenae
Weather, influenza
season, lag 1 of the
outcome
d
2% (2355
/2557)
Lung function, physicians
Medical examination,
report, self-reported symptoms
spirometry,
of general, mental and
questionnaires
respiratory health
e
III 2007-12
93%
(207/223)
Daily number of ER visits and
acute admissons for cardiorespiratory outcomes
31 May-7
IV June 2010
Self-reported general, mental
and respiratory health
Questionnaires
Winter
2010 -2011
72% (1656
/2315)
V
aStudy period starts 8 March 2006. bOnly multi-pollutant analysis. cICD codes for pulmonary outcomes: J20-J22, J40-J46 and J96;
cardiac: I20-I27, I46, I48 ,I50, stroke: G45-G46, I61-I69. dO3 only available until 2009 and H2S only available from February 2006, so
excluded from analysis in paper 3 and 2 respectively. eLow number for multi-pollutant analysis, high number for single-pollutant analysis.
30
Table 3 Ethics approval sought for each study
No
Data
I
Bioethics Committee, VSNb2008050023/0
Dispensing
Medicines
Data Protection
3-15,
of asthma
dispensing
Agency, National
2008080569/ÞPJ
medication
data
Directorate of Health
II+
III
IV
V
Hospital
admissions
and ER
visits
Medical
exam,
spirometry,
questionnaires
Questionnaire
Authority
Bioethics Committee,
Data Protection
Agency, Hospital
Executive Board
Ref. no.
For
VSNb2010120017/0
3.7,
Hospital
2010121176AT/,
contacts
Letter:2010/12/22
Public
Health
Survey
National Directorate of
Health
Individlevel data
a
org/used
No/no
Yes/no
Yes/yes
Bioethics Committee, VSNb2010080002/0 Collecting
Data Protection
3.7, S4878/2010
and keeping Yes/yes
Agency
(announced)
data
a
were the original data provided as individual level data?/ were individual-level data
used in the study analysis?
Pollen
Pollen levels can gravely affect symptoms in individuals with allergic asthma,
so in time series analyses, it is important to adjust for daily pollen counts.
These are measured from April to September by the Natural History Institute
of Iceland (Icelandic Institute of Natural History 2014, personal
communication Margrét Hallsdóttir November 16, 2010 and Lára
Guðmundsdóttir May 21, 2013). The samples were collected at the
Meteorological Institute of Iceland (see figure 11). In paper I, II and III, the
analyses were adjusted for pollen counts in the initial analysis, though the
results were not significant.
Influenza
When studying respiratory health in a population, information about
epidemics of infectious diseases is of importance as the susceptible
population is most at risk of experiencing worsening of their disease when
getting sick from influenza. Information about influenza epidemics is reported
31
to the Directorate of Health which reports it in a newsletter and on their
website (Chief Epidemiologists of Iceland, 2012a and b). Influenza epidemic
periods were defined as more than 300 cases reported to the Directorate of
Health per month. In the time series analyses (paper I, II and III), we adjusted
for infectious disease epidemics.
3.1.2 Ethics approval
For papers 1, 2, 3 and 5, ethics approval was sought. As the data used were
collected either for surveillance purposes (paper IV) or for use within health
care institutions (paper II and III) it is possible to argue that the data were
used beyond their originally intended purpose (table 3). Thus, extra care
must be taken that individual level data are not distributed beyond the source
and that the use of the data does not compromise the privacy of the patients.
3.2 Methods
In all papers, data were presented with descriptive statistics; other methods
were applied to comply with the type of data and the aim of the study.
A main distinction in environmental epidemiology is between studies of
short-term effects (such as whether there was an increase in pollution levels
on Monday associated with more cases of a certain disease on Tuesday?)
and long-term effects, where areas are usually compared. In these studies a
typical research question would be: does exposure to higher levels of
pollution (e.g. living in an area with higher background pollution levels, for
example, or close to a large road) increase the risk of a disease or symptom
compared to an area with lower pollution levels. The danger when comparing
different areas is that they may differ with respect to other factors than the
pollution.
In this section, methods used in the studies are presented; the methods
are presented schematically for all studies table 4 after a general discussion.
3.2.1 Short-term effect studies
Studies of short-term effects seek to find an association between temporal
changes in exposure and changes in the outcome shortly after. A temporal
association like this does not prove causality. The time from exposure to
outcome is called a lag.
The primary analytical method used is regression, with lagged exposures,
sometimes called time series regression. In brief, the health outcome,
dependent variable, is regressed on the exposure, the independent
32
variable(s), usually with a lag (delay) so that exposure during the previous
days is taken into account. The selection of the lag time to use in the model
must be built on knowledge of the physiological mechanisms and other
factors that occur between the time of exposure and registration of the
outcome of interest (Peng et al., 2008; Peng and Dominici, 2008; Dominici et
al., 2003).
It is necessary to adjust for covariates such as weather, day of week, and
season indicators. Both exposures and outcomes have seasonal trends, air
pollution levels are affected by meteorological variables; respiratory health
outcome severity can be affected by weather, infectious disease epidemics,
or pollen counts which all have seasonal patterns (Peng and Dominici 2008;
Tolbert et al., 2007).
Generalized linear models (GLM) and generalized additive models (GAM)
are both frequently used in time series analysis for air pollution epidemiology.
For a count outcome such as the number of patients admitted to hospital at
time t (for example a certain day; Yt) at a certain exposure vector Xt., GAM or
GLM take the following form:
where
denotes a measure of weather conditions;
is an
indicator of day of week. S denotes a smooth function, which is applied in
order to account for fluctuations in the outcome that are due to other causes
than the exposure (e.g. seasonal variation).
Seasonal variation may confound the effect estimates of the exposure
(Dominici et al., 2003). A smoother can be constructed using different
methods; lowess (locally weighted scatterplot smoothing) smoothers
(Vittinghoff et al., 2004 p.159), sine-cosine terms or splines (parametric in
GLM, non-parametric in GAM) (Dominici et al., 2003). Visual inspection of the
spline and the placing of knots can be used to estimate the non-linear effects
of an exposure. In paper I, II and III, splines were used to adjust for time
trends. Initial analyses of the data in paper I were done using sine-cosine
terms.
The distributed lag non-linear model (DNLM) is a newer method in testing
regression of time series data (Gasparrini et al., 2010). The model allows for
non-linear modelling (threshold, knots, splines) both across lags and
exposure values.
33
Exposure assignment
In air pollution studies, exposure is often assessed based on a geographical
measure of the subject’s residence. Other factors like occupational
exposures, the quality and indoor climate of the residence or the exposure at
the workplace, or school in the case of children, can modify the total
exposure. It is also possible to assign study subjects to carry personal
measurement devices. Personal exposure methods give a better exposure
assignment (Jerrett et al., 2004), but are impractical and expensive to use.
Inadequate exposure assignment can lead to faulty results, but in most cases
they will underestimate the effect of the exposure (Zeger et al., 2000).
Multi- and single- pollutant models
It has been a goal of the air pollution studies to determine which pollutants
are most harmful to a specific outcome, but some pollutants have high
correlation as the same factors (traffic, weather) influence their levels.
Inserting all pollutants into a model can adjust for this, but may also disguise
important associations, e.g. if two pollutants are highly correlated and are
associated with increases in health outcome measures. Intercorrelated
pollutants are at risk of violating the assumption of independence (Dominici et
al., 2003; Tolbert et al., 2007).
Missing data
In the case where time series data have gaps of missing data, using multipollutant models can mean excluding a lot of data as there must be
information about pollution levels for all pollutants for each day to be included
in the analysis (Figure 13). Using long lag periods increases the severity of
the problem, as one day of missing data means that a whole lag period must
be excluded. Single-pollutants models with short lags are less sensitive to
this problem. Results from single-pollutant models are often presented
alongside results from multi-pollutant models.
34
Figure 13 Example of missing data in a multi-pollutant time series. The shaded areas
will be excluded from regression analysis as values of one or more pollutants are
missing.
Auto-correlation
Many outcomes are correlated with themselves, so that the outcome on a
given day is highly correlated with the same outcome they day before or after
(Madsen, 2007). The autocorrelation function can give an indication of
whether there are signs of autocorrelations, and this can be often be resolved
by including the lag 1 of the outcome as a predictor variable in the regression
model. Serial autocorrelation due to seasonal trends can be adjusted with
splines (Schwartz, 2000).
Adjusting - and overadjusting
While most epidemiological studies adjust for various confounders and time
trend it is possible to go too far. Using splines to adjust for temporal trends, it
is possible to over-adjust, as the spline tries to adapt to the temporal changes
in the outcome. A spline with too many degrees of freedom can over-adjust
and is in danger of eradicating the signal caused by the outcome of interest.
A model showing signs of overadjustment will have no lags outside the 95%
confidence interval in the autocorrelation function.
3.2.2 Long-term effect studies
In studies of long-term effects, the aim is to find associations between certain
exposures and occurrence of certain diseases. There is a hierarchy in the
level of evidence inferred from each study type, and the outcomes that can
be determined from them.
An important measure in epidemiology is prevalence, i.e. how many
individuals have a given disease at a certain time. Another is incidence, the
proportion of new cases (new occurrences of the outcome) in a population
per time unit.
35
Cross sectional study
A cross-sectional study describes a moment in time; how many of the people
in the study had a certain outcome at this given moment. Prevalence is the
only measure that can be estimated from cross-sectional studies. Other
caveats of these methods are spatial confounding, and the difficulty in
distinguishing between whether an observed outcome is associated with
either a chronic (background) exposure or a recent (acute) exposure. For
these reasons, causality cannot be inferred from observational studies like
these (Rothman, 2002).
Cohort studies
In cohort studies, a group of people are selected based on a common
characteristic (most often residences within a geographic area, but other
versions are known, e.g. occupation, year of birth). Characteristics of interest
with respect to the outcome are measured at baseline (Rothman, 2002;
Dominici et al., 2003). Then the group is followed and at certain time points
the number of new occurrences of selected outcomes is measured and the
incidence of the outcome of interest can be calculated. Cohort studies are
both expensive and time-consuming, and it is hard to distinguish whether
chronic or acute exposures were the cause of a particular outcome at a given
time. Also, adjustment for residual confounding is always a concern, as both
exposure and outcome can be associated with confounders with a
geographic distribution (Dominici et al., 2003).
Methods used in studies of long-term effects
In paper IV and V several methods are used to determine if the occurrence of
a certain characteristic is significantly different between two groups (exposed
and unexposed); Χ2–tests were used for binary outcomes (Vittinghoff et al.,
2004). T-tests were used to compare means of numerical outcomes between
two groups. In paper V, we compared the occurrence of a binary outcome
between two or more groups adjusting for several predictors or covariates
using logistic regression (Daniel, 2005; Vittinghoff et al., 2004). See Table 4
for an overview of the methods used.
A logistic regression with one binary predictor is identical to a Χ2-test. The
logistic regression model takes the form;
P(x) = E [ y | x ] = β0+β1x
36
37
V
Descriptive, Logistic
regression (exposed
vs non-exposed &
exposure category)
Spline for study
period (~5 df/yr)
Spline for study
period (1 df/yr)
Spline for study
period (1 df/yr)
and season
Time trend
adjustment
Main analysis
repeated with
healthy onlyb
Exposure as 1-hr
max, β-agonist
drugsa, pollen
b
season
Stratified by
gender, age
groups
Sensitivity
analysis
0-7
0-2 (0-5)
0-15 (3day
means)
Df: Degrees of freedom. GHQ-12 – General Health Questionnaire 12 item version. a fast-working drugs. b Not reported .
South Iceland: 3
exposure levels. Nonexposed North Iceland
reference group
b
Descriptive, χ2 – test,
t-test. Stratified by
exposure
Daily air pollution
(linear), binary indicator GAM, DLNM
variables
Counts (n
reporting general,
mental and
respiratory health
symtoms
Counts
III
GAM
Daily air pollution
IV
Counts
II
GAM (GLM, DLNM)
Daily air pollution
Experiences during
eruption
Counts
I
Statistic
Expoure
Lung function
(%predicted),
n reporting
symptoms (GHQ12)
Outcome format
No
Table 4 Schematic overview of the methods used in each study
4 Results
The three time series studies used data from registers of drug dispensing
(paper I) and ER visits and acute admissions for cardiovascular and stroke
diagnoses (paper II and 3). In the data from 2003-2012 there were both
seasonal trends and a decline in the number of events during the study
period (figure 14). However, upon inspection of the hospital contacts data
stratified for when the subjects were admitted to stay at least one night in
hospital (blue triangles) or were sent home from the ER, it was clear that
most of the seasonal variation and the decline over time was due to the ER
visits. There were also seasonal trends in the levels of air pollutants 20032012 (figure 14) PM10 peaks tend to occur over the winter season, as do H2S
levels, which peak during cold weather with no precipitation; NO 2 levels have
much less variation. The highest levels occured during winter, and O 3 levels
were highest during spring, in April.
In the studies of the health effects form the Eyjafjallajökull eruptions, we
surveyed 207 individuals thoroughly, and received questionnaires from 1656
individuals. We found strong indications that there was a good contrast in
exposure, both between the exposed and non-exposed area, and the three
exposure regions of South Iceland.
In the following, the full time series for pollution data from Reykjavík, and
some of the hospital admissions are presented, as are results for each paper.
4.1 Paper I
Only 40% of the time series 2006-2009 had complete data for all four
pollutants. Levels of all pollutants were higher in the winter season. On
average, 76 people bought a prescription drug from the pharmacy per day.
NO2 levels were significantly correlated with temperature and O 3.
In the regression analysis, exposure was given first at 24-hour means
(peak) and 1 hour daily maximum (peak) pollution. The outcomes were given
first as all anti-asthma prescription medication (R03) and secondly as shortacting adrenergic inhalants (R03A) only. Three-day moving averages of the
exposure up to lag 14 were used.
Mean H2S was associated with increased dispensing of all asthma
medication and adrenergic inhalants at lags 3-5.
39
40
100
NO2
0
50
µg m
3
NO2
O3
1. jan 2008 1. jan 2009 1. jan 2010 1. jan 2011 1. jan 2012
O3
100
1. jan 2003 1. jan 2004 1. jan 2005 1. jan 2006 1. jan 2007
50
µg/m3
Suspected faulty
3
0
H2S
1. jan 2004
1. jan 2005
1. jan 2006
1. jan 2007
1. jan 2008
1. jan 2009
1. jan 2010
1. jan 2011
1. jan 2012
31.dec 2012
50
H2S
PM10
1. jan 2003 1. jan 2004 1. jan 2005 1. jan 2006 1. jan 2007
1. jan 2008 1. jan 2009 1. jan 2010 1. jan 2011 1. jan 2012
PM10
300
200
100
0
µg/m3
400
500
600
0
µg/m3
μg/m
1. jan 2003
1. jan 2003 1. jan 2004 1. jan 2005 1. jan 2006 1. jan 2007
1. jan 2008 1. jan 2009 1. jan 2010 1. jan 2011 1. jan 2012
Figure 15 Daily means of air pollutants (μg/m3) measured in Reykjavik 20032012.
41
Both mean and peak PM10 were associated with increased dispensing of
all asthma medication and adrenergic inhalants at lags 3-5. Peak, but not
mean NO2, was associated with increased dispensing of adrenergic inhalant
drugs at lag 3-8. For O3, only peak values at lags 3-5 and 9-11 were
associated with increased dispensing of all anti-asthma drugs and adrenergic
inhalants.
Figure 16 Overview of Reykjavík postcodes and measuring stations. Circled
postcodes are re-analysed in figure 16 and table 5.
H2S comes from point sources and distributes in a different way than the
traffic-related pollution. Later research has shown that the weather conditions
that cause the highest levels of H2S in Reykjavik are a slow breeze from the
east, during cold weather, and that the resulting exposure “cloud” is only
some 3 km wide (Ólafsdóttir et al., 2013). As the study area is some 20 km
from north to south, the measured value at the station was not entirely
representative of the exposure of the whole study population.
In a further analysis, an attempt was made to improve the exposure
classification, and the analysis was repeated counting only subjects who lived
in the postcodes in line with the measuring station (figure 16). The results for
42
those who lived in the post codes areas between the exposure source and
the measuring station (east Reykjavík) are shown in table 5 and figure 17.
H2S and dispensing of asthma drugs in East Reykjavík
Effect estimates adjusted (left) and unadjusted (right) for other pollutants for 3-day distributed lag across lag0H2S and dispensing
14. of asthma drugs
0
-6
-4
-2
% change
2
4
6
East Reykjavík : Effect estimates for H2S, adjusted (left) and uadjusted (right) for 3-day lag days 0-14
Lag0-2
Lag3-5
Lag6-8
Lag9-11
Lag12-14
Figure 17 The association between anti-asthma drug dispensing in east Reykjavik
and H2S.
Table 5 The association between H2S and anti-asthma medication
dispensing in areas within Reykjavík. Adjusted for other pollutants, weather
and time trend. Results are given as percent change per 10µg/m3 pollutant
concentration at lags 0-14.
All area
In line with
Hellisheiði –
Grensásvegura
Between Hellisheiði
– Grensásvegurb
%
95% CI
%
95% CI
%
95% CI
H2S
0.66
0.10, 1.21
0.99
0.27, 1.72
1.11
0.20, 2.05
PM10
-0.00
-0.37, 0.32
-0.29
-0.61, 0.16
-0.14
-0.73, 0.43
NO2
0.49
-0.08, 1.05
0.46
-0.25, 1.19
0.23
-0.68, 1.26
O3
-0.42
-0.83, -0.01
-0.21
-0.73, 0.31
-0.58
-1.24, 0.09
a
Within circled area in figure 16. bEastern part of the circled area in figure 16.
43
Furthermore, the estimated effect of a 10 µg/m3 increase in H2S was
higher when including only the postcodes in the direct line between the
source and the measuring station. We found that the effect size increased for
H2S, whereas the results for the other pollutants did not change in any
significant way, Analysis stratified by gender revealed that the association
was stronger in women.
4.2 Paper II
The pollution data from 2003-9 were 59-92% complete. There were 9.6 ER
vists and hospital admissions per day, most were cardiac, and more than
60% were for people 60 years or older.
Only O3 was consistently associated with increased admissions to hospital
for cardiopulmonary disease or stroke, and these results were very robust
when the analysis was repeated with different lag times and covariates.
We excluded H2S from the final analysis of this paper and did not report
the results in the paper, but we found a 1% (95% CI: 0, 2%) increase in
hospital admissions and ER visits for a 5 th to 95th percentile increase in H2S,
but the increase was not statistically significant.
4.3 Paper III
The study period was 2007-2012. There were 121 days in the time series
when the PM10 values exceeded 50µg/m3. Volcanic ash contributed to the
high levels 20 times, and dust storms were a contributor 16 times. During the
study period the average number of daily emergency hospital visits was 10.5.
In a basic model, we found no association between PM10 and emergency
hospital visits; the effect estimate was 0.8% (95% CI -0.5, 2.2). To answer
the question if the effect of PM10 had changed after the volcanic eruption, we
analyzed the period before and after the eruption separately. We found that
the effect estimate for the latter period had increased to 1.6%, (95%CI -1.1,
4.3), but remained statistically insignificant.
In the indicator analysis, high PM10 levels due to any cause were not
associated with emergency hospital visits. High PM10 levels due to volcanic
ash were significantly associated with emergency hospital visits when
adjusting for other pollutants by 9.2% (95% CI -0.0, 19.6%). High PM10 levels
due to dust storms were not associated with emergency hospital visits, the
effect estimate was 4.5% (95% CI -1.6, 11.0%).
44
4.4 Paper IV
At the end of the eruption of Eyjafjallajökull in 2010 a total of 167 adults and
40 children (under 18) were investigated with questionnaires, spirometry and
a medical examination. Lung function was no worse than in a reference
group from the capital area.
In the questionnaires, the respondents reported high levels of irritation
symptoms; more than half reported symptoms from the sensory organs that
were severe enough to disturb their daily life. A high percentage reported
psychological morbidity (39% fulfilled GHQ-12 criteria), whereas only few had
symptoms of post-traumatic stress disorder. All children with asthma had
reported using more medication than normal during the eruption.
In the medical interview, a number of people reported stress; though only
a few fulfilled the PSS-SR (post traumatic stress disorder symptom scale,-self
report, Brewin et al., 2002) criteria for posttraumatic stress disorder. In the
questionnaire that was handed out, others voiced their concerns.
During the eruption while the ash was falling, my appetite was
bad, I slept badly, was nauseous and anxious. Chronic fatigue
the first week, afraid of more disasters and uncertainty about the
future.
Man, early 30s
Was surprised how large the effects were on the lungs – felt in
my lungs the minute ash began to fall here. Day or night – would
have known if I was blindfolded.
Woman, Late 30s
I seem to have phlegm in the lungs and the dust (ash) must
have gotten stuck there.
Male, 50s
I found it hard to experience, unreal. Was stressed because the
roads were closed for a while and also because of the ash fall.
Many friends and acquaintances called, empathy, yet people did
not really understand.
Female, 60s
45
Being exposed to the volcanic eruption, worrying about the safety and
health of oneself, family and neighbors, was a disturbing event, but some
also reported positive outcomes, such as having made new friends.
4.5 Paper V
Six months after the eruption of Eyjafjallajökull, questionnaires were sent to a
cohort of South Icelanders who lived between 10 and 100 km from
Eyfjafjallajökull and a reference group from North Iceland. The exposed
South Icelanders had two to three times increased risk of most respiratory
and eye symptoms 6 months after the eruption, after adjusting for
demographic factors. About half of the South Icelanders had symptoms from
eyes, nose or throat, whereas one third of the reference group had
symptoms. The sensory organ symptoms clustered in the exposed
population, so individuals with one symptom were more likely to report other
symptoms.
The proportion reporting psychological morbidity had declined from 39%
at the end of the eruption to 24% in the most exposed population, while it was
20% in the unexposed reference group. Bird and Gísladóttir (2012) polled
people exposed to the eruption and showed that after exposure to the
eruption from Eyjafjallajökull gave a more realistic expectation of what to
expect from future eruptions.
Figure 18 Proportion of cohort reporting “some” or “a lot” of exposure to volcanic
phenomena at their residence during the eruption in Eyjafjallajökull 2010.
46
While there was a dose-response association between exposure and
symptoms, we found that most individuals in the cohort had experienced ash
fall and heard rumbling from the volcano. Only a minority had felt
earthquakes. Generally, those with “medium exposure” had experienced
more volcanic phenomena (figure 18).
In the cohort study, participants were given the option of replying either
online or on paper, and while the overall reply rate for the study was 72%
(number of replies/number of people who the researchers were able to
contact) there were differences between the reply categories. In stratifying
the reply modes by exposure category (table 6), we found higher reply rates
from people who agreed to answer online. This could be in part because the
online repliers were younger, but that cannot entirely explain this difference.
The numbers in Table 6 do not correspond to the actual reply rates reported
in the paper as those only include people with accepted participation and who
were sent a questionnaire.
Table 6 Reply rates for Eyjafjallajökull cohort study – web and on paper
Exposed
Persons who received
a questionnaire
Number of valied
questionnaires returned
Reply rate by category
Non-exposed
paper
web
paper
web
666
814
215
380
448
701
175
335
67.3%
86.1%
81.4%
88.2%
47
5 Discussion
In the studies of short-term effects we found associations between increased
pollution levels and increases in adverse health indicators. There was an
association between PM10 and H2S levels and anti-asthma drug dispensing to
adults in the capital area. In the studies of the health effects following the
Eyjafjallajökull eruption, we found that lung function was not affected by
exposure to the volcanic eruption, but that irritation and respiratory symptoms
were more common in the exposed population.
5.1 Key results
Only O3 was consistently associated with emergency hospital visits due to
cardiopulmonary causes in adult males, females and the elderly over 70
years of age in 2003 to 2009. NO2 levels were only associated with increased
emergency hospital visits in those over 70 years of age, and there were no
associations between PM10.
High PM10 levels due to volcanic ash from Eyjafjallajökull were associated
with increased emergency hospital visits for cardiopulmonary causes in
adults 2007-2012. The association between PM10 and emergency hospital
visits was stronger after the eruption of Eyjafjallajökull than before, but
neither were significant.
In the studies of the South Iceland population exposed to ash from the
Eyjafjallajökull eruption, we found that at the end of the eruption more than
half of those exposed reported having experienced symptoms that were
severe enough to disturb their daily life. Many experienced stress and
psychological morbidity. Compared to a reference population from the capital
area, lung function was not lower in the exposed population; however, the
reference population included many heavy smokers.
Six months later, the exposed population had two to three times increased
risk of most respiratory and eye symptoms, and nearly half had symptoms
from the eyes, nose or throat, compared to one third of the reference
population.
49
5.2 Limitations and methodological considerations
5.2.1 Exposure
In paper I and II, on the effects of pollution from urban sources, NO2 and O3
were found on ER visits and hospital admissions, but not for dispensing of
anti-asthma medication in paper I. Effects from PM10 were only found in the
dispensing of anti-asthma drugs, but not on ER visits and hospital admissions
in 2003-2009. Data on PM2.5 were not available, which is unfortunate as
many of the known health effects of PM are for this fraction. With PM 2.5 data,
it would be possible at least to distinguish between mainly traffic related
particles from fossil fuel combustion and particles from mechanical wear and
dust storms, which tend to be larger.
In paper III, we found mainly effects with H 2S and high levels of PM10 from
volcanic ash, but O3 was not available. A lack of adjustment for O 3 is likely to
affect the effect estimates for NO2, so this is unfortunate.
The Icelandic capital area is surrounded by sea on three sides, and PM 10
has a significant salt content (3-11% Skúladóttir, 2003; Vegagerðin, 2013).
Salt is considered to relieve some asthma symptoms, as has been suggested
by other researchers (Andersen et al., 2008). The salt content could be a
contributing factor to our null result for PM10 in paper II and partially in paper
III.
In paper I, PM10 was associated with the dispensing of anti-asthma drugs,
which is similar to the findings of other studies using anti-asthma drug
dispensing as outcome (Zeghnoun et al., 1999; Pitard et al., 2004; Laurent et
al., 2008), and also for PM10 from natural sources and asthma morbidity in
children and adolescents (Chimonas and Gessner, 2007). No associations
were found with NO2 and O3 levels, but the concentrations of both were
somewhat lower in Reykjavík than in the other studies, whereas levels of
PM10 were similar to those in Laurent et al.’s study of pharmaceutical
dispensing (2008). Other studies used another metric of combustion
particles, Black Smoke (BS).
There are relatively few studies of pharmaceutical dispensing, and none
except this one considers H2S. A study from Taipei suggest that interactions
between geothermal emissions and traffic pollution may further increase the
aerosol level in urban areas (Lin et al., 2010). A study from New Zealand
found that residential exposure to background H 2S was associated with lower
risk of asthma (Bates et al., 2013).
50
In paper II, hospital emergency admissions 2003-2009 were only
associated with O3, and NO2 in the elderly. The null result for PM10 was
surprising, and possible causes for this are discussed below. No-effect levels
for NO2 have been found in experimental studies, but those levels were much
lower than in the current study (Hesterberg et al., 2009). Influenza epidemics
were not a significant predictor for increased hospitalizations or ER visits in
this study. It is possible that efforts to inoculate the susceptible population
have been sufficient to remove this risk factor.
The results of paper III indicate that the association between PM 10 from
volcanic ash and emergency hospital visits does not follow the same doseresponse functions as traffic-related PM10. These finding are similar to those
of Oudin et al., (2013) who found a bigger-than expected, but non-significant
increase in mortality the week after the ash from the Grimsvötn volcanic
eruption in 2011 passed over Sweden.
Data on O3 levels were missing or suspected faulty since 2010 due to a
fault in the measuring equipment. Therefore O 3 had to be excluded from the
analyses of the data before and after the eruption. This means that the
results are not immediately comparable to those from the 2003-2009 periods.
In paper I, II and III, the exposure measured at one station was used as a
proxy for the exposure for the entire population. This is not an unusual
method, but as the recalculation of the results from paper I illustrate (figure
17), the exposure assessment for H2S was probably improved when only
including the post codes areas near the measuring station. In these studies,
we used microgram per cubic meter air as it was readily available. Other
exposure metrics such as particle number concentration, size distribution and
surface area are used in other studies. Newer results indicate that certain
particle characteristics may be relevant to particular disease outcomes
(Andersen et al., 2008, Peng et al., 2009; Atkinson et al., 2010). In paper IV
all participants were exposed at their residence or when they were working in
the area. Some were probably more exposed that others, but the study size
or the information about the ash fall did not permit us to distinguish between
degrees of exposure.
A high level of variation (contrast) in the exposure variable is vital to the
robustness of time series methods (Dominici et al., 2003; Künzli and
Schindler, 2005). In our data, PM10 and H2S exhibited the highest degree of
variability, and we found the most robust results for exactly those two
pollutants in paper I, with only peak O 3 and NO2 significantly associated with
the outcome. In paper II, we found significant results for O 3, but not other
51
pollutants. In paper III, the contrast in PM10 increased after the
Eyfjafjallajökull volcanic eruption, which may have been a contributing factor
to the higher effect of PM after the eruption.
Some missing data meant that especially the analysis with long lags (15
days in paper I) had a lot of missing data. As we had a long time series, the
statistical power was sufficient. But stratifying further (gender, age groups)
we found no significant results.
In paper V, we estimated exposure from the participants’ area of
residence address and from satellite data estimates of the eruption plume,
but the actual exposure would have been highly sensitive to the activities of
the participants (working indoors or outdoors). Some left the exposed area for
some periods during and after the eruption to “get a break”, but such
behavior would mainly decrease the actual exposure and bias the results
towards the null hypothesis.
5.2.2 Outcomes
In paper I, we used pharmaceutical dispensing as the outcome. Use of this
outcome is increasing as several countries now register dispensing in
databases. This outcome is considered more sensitive as it reflects increased
symptoms or disease worsening which is less severe than one which causes
people to seek treatment in the emergency room or be admitted to hospital.
There are also more events, as many people take medication continuously,
which gives the results more statistical power.
Drawbacks with pharmaceutical dispensing as outcome include significant
uncertainty of the time the medication is actually used relative to the time it is
dispensed (which is the date registered in the data), or if the medication was
used at all. For anti-asthma medication a further caveat is that there is little
specificity to asthma, as the medications are used to treat symptoms of a
number of diseases.
In paper II and 3, we combined ER visits and hospital admission for
cardiopulmonary causes and stroke. These two outcomes have quite
different associations with air pollution (Winquist et al., 2012). In our data, ER
visits displayed a more temporal trend (figure 18), but we refrained from
analyzing the ER visits and hospital admissions separately as the statistical
power would have been considerably diminished.
In paper IV, standardized spirometry methods (protocol from the BOLD
study) were used to assess respiratory health, which facilitated comparisons
52
with other studies using the same methods. Unfortunately, we had no
reference group for the people under 40, and spirometry cannot be used on
children under the age of 6, so the respiratory health of these groups was not
compared to a reference group.
The self-reported health and symptom prevalence used in papers IV and
V are sensitive to the respondents’ mental state at the time of answering the
questionnaire, and there are concerns that the replies were influenced by
gender and age. The questionnaires used to assess mental health in papers
IV and V were rather robust (Goldberg et al., 1997).
In paper IV, the free text replies and reported adverse mental health
results have to be seen in light of the considerable level of uncertainty at the
time of the study, May 31 – June 11, 2010. The eruption had subsided, but
geoscientists suspected that it could start again. Eyjafjallajökull has a history
of year-long eruptions with intermittent quiescence. An even greater concern
was that nearby volcano Katla would erupt soon after an eruption of
Eyjafjallajökull as had happened several times previously (Sturkell et al.,
2010). A limitation of paper IV is the limited sample size, the lack of
knowledge about the health status before the eruption and the lack of
reference population and exposure contrast. Unfortunately, some participants
did not answer the questionnaire, which further decreased the power of the
study.
In paper V, which was based on questionnaire data, the demographically
matched reference group and the adjustment for demographic factors in the
regression analysis should minimize the risk of over- or underestimation of
symptoms in the exposed group. Still, it is impossible to eliminate a risk of
over-reporting in the exposed population.
We identified large differences in the occurrence of reported symptoms
and other complaints; and adjusting for personal characteristics should have
excluded most residual confounding.
It is a limitation of the study that the health of the populations before the
eruption is unknown, and we have little information about the non-responders
and non-participants. Both the exposed and non-exposed populations are
relatively small and tended to by rather few health professionals. Differences
in professional preference and diagnostic criteria could mean a relatively high
risk of diagnosis bias. This could possibly explain the higher asthma
prevalence reported in the non-exposed population. Our exposure categories
were based on ash fall estimates and knowledge of the area, but local
53
conditions can increase or decrease both perceived and actual exposure.
This could bias the results in either direction, but most likely towards the null.
The non-exposed population was surveyed later in winter (January-March)
when there are more respiratory infections, which would bias the results
towards the null.
5.2.3 Statistical methods
In the time series studies (paper I, II and III) we adjusted for temporal trends
using splines. Using splines with too many degrees of freedom can disguise
the true correlations. Inspection of the splines and autocorrelation plots for
signs of overfitting and a moderate selection of degrees of freedom ensured
that the splines were suitable. Remaining signs of autocorrelation at lag 1
were removed when the lag 1 of the outcome was introduced as a predictor.
Case crossover methods adjust for temporal trends by design; the exposure
on a day with an event in an individual (index) is compared to the exposure
during a day with no event (control) (Maclure, 1991; Maclure and Mittleman,
2000), but these methods have their own set of drawbacks, such as
sensitivity to selection of reference dates (see e.g. Whitaker et al., 2007), and
were not used in the current study.
The DLNM method which was used in paper III is a rather new method. It
appears highly sensitive to the selection of parameters, and while the method
has gained popularity since its introduction by Gasparrini and Armstrong
(2010), there is no best practice and little consensus as to how to best fit and
optimize these models.
In paper V, we presented the percentage reporting symptoms in each
exposure category. We used logistic regression adjusted for age, gender,
education and smoking status to obtain the risk estimates for the different
exposure categories. Other information, such as occupation, and other lifestyle factors could have been included in the model.
5.3 Interpretation
The results from paper I indicate that dispensing of anti-asthma medication, a
proxy for population respiratory health, was affected by even moderate levels
of air pollution as found in Reykjavík 2006-9. The results with regard to H2S
are of particular interest as this was a rare exposure.
The results from paper II showed an association between O 3 and
emergency hospital visits for cardiopulmonary and stroke. O3 levels never
surpassed the health limits in Iceland during the study period.
54
The results from paper III indicated that the pathogenic effects of PM 10
increased after volcanic ash from Eyjafjallajökull was added to the urban
PM10. PM10 from volcanic ash and dust storms have substantial health
effects, and are an ignored factor in air pollution studies.
In paper IV, we found a high prevalence of self-reported irritation
symptoms to the sensory organs in the population exposed to volcanic ash
during the Eyjafjallajökull eruption. After 6 weeks of exposure, the
participants did not have lower lung function than a matched reference group.
In paper V, the results of this large cohort study suggest that exposure to
a volcanic eruption is associated with increased rates of irritation symptoms
of the sensory organs and psychological morbidity 6 months later, in spite of
a concerted disaster response and mitigation efforts.
5.4 External validity
The results from paper I, II and III are based on data from high quality
registers in a health care system with universal coverage, making the studies
in effect population based. Population-based studies are not prone to
selection bias and have a high generalizability. The results from the studies
add to the body of knowledge about exposure to moderate levels of air
pollution in near-Arctic conditions and particularly exposure to H 2S and dust
from volcanic ash, which are rare exposures that have not been extensively
studied previously.
The results from paper IV and V provide useful information about
symptom prevalence and resilience in a well-defined population exposed to a
volcanic eruption and volcanic ash. Pathogenic properties of volcanic ash
vary greatly between volcanoes and eruptions, but methods to determine
pathogenic properties of ash are constantly improving (Monick et al., 2013;
Horwell et al., 2013).
55
6 Conclusions
From the research presented in this thesis, we found that both anthropogenic
and natural sources contribute to air pollution which has significant effects on
respiratory health in Iceland.
Paper I Daily number of individuals dispensed anti-asthma drugs was
associated with the levels of PM10 and H2S 3-5 days previously, and more so
for the daily maximum than the 24-hour mean. This association was more
significant when restricting the analysis to short-acting medications, indicating
that they are a better proxy for short-term changes in respiratory health in a
population.
Paper II Daily emergency hospital admissions and ER visits for
cardiopulmonary disease and stroke were associated with O 3 levels at lags 02. Women and the elderly were more susceptible. NO 2 was also associated
with increased admissions in the elderly.
Paper III In this study of the association between emergency hospital
contacts and particle sources during days with high PM10 levels, we found
higher effect estimates when the source was volcanic, and smaller effect
estimates when PM10 levels were high because of natural dust storms. The
effect estimates of PM10 tended to be higher after the eruption of
Eyjafjallajökull, though not significantly.
Paper IV The population which lived within 40 km of Eyjafjallajökull
reported high levels of irritation symptoms in the sensory organs as a
consequence of being exposed to fresh volcanic ash for six weeks. Lung
function was not affected, and mitigation efforts had been effective.
Individuals with underlying obstructive lung disease, elderly and children
were more vulnerable.
Paper V Exposure to the Eyjafjallajökull volcanic eruption was associated
with increased risk of respiratory symptoms and psychological morbidity, and
the association followed a dose-response pattern. Psychological morbidity
was more common among those who reported more than one physical
symptom; mitigation efforts could be aimed at such subgroups.
57
Take home message
Both anthropogenic pollution and natural particles have adverse
consequences for respiratory health in Iceland. Mitigation measures in the
capital area are already installed (dust binding) to reduce PM, but the natural
source pollution has if anything been increasing and is a concern in the
future.
6.1 Limitations of the current study and future perspectives
The current studies of short-term exposure did not take the effects of
personal risk factors or underlying diseases into account; this could be the
subject of a future study, as several large cohorts with long follow-up times
are in place in Iceland.
In the studies of exposure to Eyjafjallajökull, we found that children had
been severely affected by exposure to the ash from Eyjafjallajökull, and all
children with asthma had experienced worsening of their symptoms, even
though they were not in the area during the whole eruption.
Ideas for future studies:

Short-term health effects of traffic-related air pollution in children

Short-term health effects of traffic-related pollution on susceptible
groups, such as those with underlying diseases and the elderly.

Exposure models for the capital area are being developed and
can be used to estimate the effects of long-term exposure to
traffic pollution.

In studies of other outcomes, effect modification from chronic
exposures to air pollution can be investigated using information
from the models.

Follow-up of the Eyjafjallajökulll cohort in medicines and chronic
disease registers.
58
References
Akselsson, R., Bohgard, M., Gudmundsson, A., Hansson, H.-C., Martinsson,
B., & Svenningsson, B. (1994). Aerosoler. (M. Bohgard, Ed.). Lunds
Tekniska Högskola och NOSA.
Althingi. Reglugerð um brennisteinsdíoxíð, köfnunarefnisdíoxíð og
köfnunarefnisoxíð, bensen, kolsýring, svifryk og blý í andrúmsloftinu og
upplýsingar til almennings [Act regulating the surface concentration of
sulfur dioxide, nitrogen dioxide, carbon dioxide, particle matter and lead,
and information to the public], Pub. L. No. 251 (2002). Retrieved from
http://www.reglugerd.is/interpro/dkm/WebGuard.nsf/key2/251-2002
[In
Icelandic]
Althingi. Reglugerð um styrk ósons við yfirborð jarðar [Act regulating surface
ozone concentration]. , Pub. L. No. 745 (2003). [In Icelandic] Retrieved
from http://www.reglugerd.is/interpro/dkm/WebGuard.nsf/key2/745-2003
[In Icelandic]
Althingi. Reglugerð um styrk brennisteinsvetnis í andrumslofti [‚Act regulating
the surface concentration of hydrogen sulphide]. , Pub. L. No. 514 (2010).
[In Icelandic] Retrieved from http://www.reglugerd.is/interpro/dkm
/WebGuard.nsf/58b439f05a7f412f00256a07003476bc/0cf126cdabe4eb9c
00257749004c2747?OpenDocument
Amaral, A. F. S., & Rodrigues, A. S. (2007). Chronic exposure to volcanic
environments and chronic bronchitis incidence in the Azores, Portugal.
Environ Res, 103(3), 419–23.
Andersen, Z. J., Wahlin, P., Raaschou-Nielsen, O., Ketzel, M., Scheike, T., &
Loft, S. (2008). Size distribution and total number concentration of
ultrafine and accumulation mode particles and hospital admissions in
children and the elderly in Copenhagen, Denmark. Occup Environ Med,
65(7), 458–66.
59
Andersen, Z. J., Hvidberg, M., Jensen, S. S., Ketzel, M., Loft, S., Sørensen,
M., Tjønneland, A., Overvad, K., & Raaschou-Nielsen, O. (2011). Chronic
obstructive pulmonary disease and long-term exposure to traffic-related
air pollution: A cohort study. Am J Respir Crit Care Med, 183(4), 455–61.
Andersen, Z. J., Kristiansen, L. C., Andersen, K. K., Olsen, T. S., Hvidberg,
M., Jensen, S. S., Ketzel, M., Loft, S., Sørensen, M., Tjønneland, A.,
Overvad, K., & Raaschou-Nielsen, O. (2012). Stroke and long-term
exposure to outdoor air pollution from nitrogen dioxide A cohort study.
Stroke, 43(2), 320–5.
Anderson, G. B., Krall, J. R., Peng, R. D., & Bell, M. L. (2012). Is the relation
between ozone and mortality confounded by chemical components of
particulate matter? Analysis of 7 components in 57 US communities. Am J
Epidem, 176(8), 726–32.
Anderson, H. R., Favarato, G., & Atkinson, R. W. (2013). Long-term exposure
to outdoor air pollution and the prevalence of asthma: meta-analysis of
multi-community prevalence studies. Air Qual Atmos Health, 6(1), 57–68.
Antó, J. M. (2012). Recent advances in the epidemiologic investigation of risk
factors for asthma: a review of the 2011 literature. Curr Allergy Asthma
Rep, 12(3), 192–200.
Arnalds, O. (2010, August). Dust sources and deposition of Aeolian materials
in Iceland. Icel Agric Sci, 23/2010, 3–21.
Arnalds, O., Þórarinsdóttir, E. F., Þórsson, J., & Ágústsdóttir, A. M. (2011).
Hvað varð um öskuna frá Eyjafjallajökli? [What happened to the ash from
Eyjafjallajökull?] Presented at Fræðaþing Bændafélagsins, Reykjavík. [In
Icelandic] Retrieved from www.lbhi.is/lisalib/getfile.aspx?itemid=4654
Arnalds, O., Thorarinsdottir, E. F., Thorsson, J., Waldhauserova, P. D., &
Agustsdottir, A. M. (2013). An extreme wind erosion event of the fresh
Eyjafjallajokull 2010 volcanic ash. Sci Rep, 3, p1257.
Atkinson, R., Fuller, G., Anderson, H., Harrison, R., & Armstrong, B. (2010).
Urban Ambient Particle Metrics and Health A Time-series Analysis.
Epidemiology, 21(4), 501–11.
60
Bates, M. N. (2013, September). Long-term, Ambient Hydrogen Sulfide
Exposure and Cognitive Function. Presented at the International Society
of Environmental Epidemiology, Basel, Switzerland. Retrieved from
http://ehp.niehs.nih.gov/ehbasel13/wp-content/uploads/2013/09/EHB13Abstracts.pdf
Bates, M. N., Garrett, N., Crane, J., & Balmes, J. R. (2013). Associations of
ambient hydrogen sulfide exposure with self-reported asthma and asthma
symptoms. Environ Res, 122, 81–7.
Benediktsdóttir, B., Gudmundsson, G., Jörundsdóttir, K. B., Vollmer, W., &
Gíslason, T. (2007). [Prevalence of COPD in Iceland–the BOLD study].
Laeknabladid, 93(6), 471–7.
Bérubé, K., Balharry, D., Jones, T., Moreno, T., Hayden, P., Sexton, K.,
Hicks, M., Merolla, L., Timblin, C., Shukla, A., & Mossman, B. (2006).
Characterisation of airborne particulate matter and related mechanisms of
toxicity: An experimental approach. In J. Ayres, R. Maynard, & H. Richard
(Eds.). Air Pollution and Health (Vol. 3, pp. 69–109). London: Imperial
College Press.
Bird, D., & Gísladóttir, G. (2012). Residents’ attitudes and behaviour before
and after the 2010 Eyjafjallajökull eruptions—a case study from southern
Iceland. Bull Volcanol, 74(6), 1263–79
Bobb, J. F., Dominici, F., & Peng, R. D. (2011.). A Bayesian model averaging
approach for estimating the relative risk of mortality associated with heat
waves in 105 U.S. cities. Biometrics, 67(4), 1605–16.
Bråbäck, L., & Forsberg, B. (2009). Does traffic exhaust contribute to the
development of asthma and allergic sensitization in children: findings from
recent cohort studies. Environ Health, 8(17).
Bracha, H. S. (2006). Human brain evolution and the “Neuroevolutionary
Time-depth Principle:” Implications for the reclassification of fear-circuitryrelated traits in DSM-V and for studying resilience to warzone-related
posttraumatic stress disorder. Prog Neuro Psychoph, 30(5), 827–853.
Brewin, C. R., Rose, S., Andrews, B., Green, J., Tata, P., McEvedy, C.,
Turner, S., Foa, E. B. (2002). Brief screening instrument for post-traumatic
stress disorder. Br J Psychiatr, 181(2), 158–62.
61
Brook, R. D., Rajagopalan, S., Pope, C. A., 3rd, Brook, J. R., Bhatnagar, A.,
Diez-Roux, A. V., Holguin, F., Hong, Y., Luepker, R. V., Mittleman, M. R.
Peters, A., Siscovick, D., Smith, S.C., & Kaufman, J.D. American Heart
Association Council on Epidemiology and Prevention, Council on the
Kidney in Cardiovascular Disease, and Council on Nutrition, Physical
Activity and Metabolism. (2010). Particulate matter air pollution and
cardiovascular disease: An update to the scientific statement from the
American Heart Association. Circulation, 121(21), 2331–78.
Brunekreef, B., & Holgate, S. T. (2002). Air pollution and health. Lancet,
360(9341), 1233–42.
Buist, A. S., McBurnie, M. A., Vollmer, W. M., Gillespie, S., Burney, P.,
Mannino, D. M., Menezes, A. M. B., Sullivan, S. D. Lee, T. A., Weiss, K.
B., Jensen, R. L., Marks, G. B. Gulsvik, A., & Nizankowska-Mogilnicka, E.
(2007). International variation in the prevalence of COPD (the BOLD
Study): A population-based prevalence study. Lancet, 370(9589), 741-50.
Bullard, J. E. (2013). Contemporary glacigenic inputs to the dust cycle. Earth
Surf Proc Land, 38(1), 71–89.
Burney, P. G., Luczynska, C., Chinn, S., & Jarvis, D. (1994). The European
Community Respiratory Health Survey. Eur Respir J, 7(5), 954–60.
Burney, P., Jithoo, A., Kato, B., Janson, C., Mannino, D., NizankowskaMogilnicka, E.,Studnicka, M., Tan, W. Bateman, E., Koçabas, A., Vollmer,
W. M., Gislason, T., Marks, G., Koul, P. A., Harrabi, I., Gnatiuc, L., &
Buist, S. for the Burden of Obstructive Lung Disease (BOLD) Study.
(2013). Chronic obstructive pulmonary disease mortality and prevalence:
The associations with smoking and poverty--a BOLD analysis. Thorax,
69(5):465-73
Campagna, D., Kathman, S. J., Pierson, R., Inserra, S. G., Phifer, B. L.,
Middleton, D. C., Zarus, G. M., & White, M. C. (2004). Ambient hydrogen
sulfide, total reduced sulfur, and hospital visits for respiratory diseases in
northeast Nebraska, 1998–2000. J Expo Anal Epid, 14(2), 180–7.
Cashman, K., & Cronin, S. (2008). Welcoming a monster to the world: Myths,
oral tradition, and modern societal response to volcanic disasters. J
Volcanol Geoth Res, 176(3), 407–18.
62
Chan, C. K., & Yao, X. (2008). Air pollution in mega cities in China. Atmos
Environ, 42(1), 1–42.
Chan, C.-C., & Ng, H.-C. (2011). A case-crossover analysis of Asian dust
storms and mortality in the downwind areas using 14-year data in Taipei.
Sci Tot Environ, 410–411, 47–52.
Chen, H., Burnett, R. T., Kwong, J. C., Villeneuve, P. J., Goldberg, M. S.,
Brook, R. D., van Donkelaar, A., Jerrett, M., Martin, R. V., Kopp, A.,
Brook, J. R., & Copes, R. (2013). Spatial association between ambient
fine particulate matter and incident hypertension. Circulation, 129 (5), 5629
Chief epidemiologist of Iceland. (2012a, May 4). Skráningarskyldir sjúkdómar
eftir mánuðum, 1997-2011 (xls) [Infectious diseases by month 19972011]. Government institution. Retrieved January 25, 2014, from
http://www.landlaeknir.is/servlet/file/store93/item17624/Skr%C3%A1ninga
rskyldir_1997_2011.xls [In Icelandic]
Chief epidemiologist of Iceland. (2012b, October 26). Tölur um inflúensu
[Influenza statistics]. Government institution. Retrieved January 25, 2014,
from http://www.landlaeknir.is/influensa/tolur-um-influensu/ [In Icelandic]
Chimonas, M.-A. R., & D. Gessner, B. (2007). Airborne particulate matter
from primarily geologic, non-industrial sources at levels below National
Ambient Air Quality Standards is associated with outpatient visits for
asthma and quick-relief medication prescriptions among children less than
20 years old enrolled in Medicaid in Anchorage, Alaska. Environ Res,
103(3), 397–404.
Chuang, K.-J., Chan, C.-C., Su, T.-C., Lee, C.-T., & Tang, C.-S. (2007). The
effect of urban air pollution on inflammation, oxidative stress, coagulation,
and autonomic dysfunction in young adults. Am J Respir Crit Care Med,
176(4), 370–6.
Collin-Lange, V. (2013). Socialities in motion: Automobility and car cruising in
Iceland. Mobilities, 8(3), 406–23.
D’Amato, G., Liccardi, G., D’Amato, M., & Holgate, S. (2005). Environmental
risk factors and allergic bronchial asthma. Clin Exp Allergy 35(9), 1113-24.
63
Dagsson-Waldhauserova, P., Arnalds, O., & Olafsson, H. (2013). Long-term
frequency and characteristics of dust storm events in Northeast Iceland
(1949–2011). Atmos Environ, 77, 117–27.
Daniel, W. W. (2005). Biostatistics - A Foundation for Analysis in the Health
Sciences (8th ed.). Hoboken, New Jersey: John Wiley & Sons.
Delfino, R. J., Staimer, N., Tjoa, T., Gillen, D. L., Polidori, A., Arhami, M.,
Kleinman, M. T., Vaziri, N. D., Longhurst, J., & Sioutas, C. (2009). Air
Pollution Exposures and Circulating Biomarkers of Effect in a Susceptible
Population: Clues to Potential Causal Component mixtures and
mechanisms, Environ Health Perspect 117(8), 1232–8.
Department of Health Clean Air Branch. (2001). 2001 Annual Summary
Hawaii Air Quality Data (p. 55). Honolulu, Hawaii. Retrieved from
http://health.hawaii.gov/cab/files/2013/05/aqbook_2010.pdf
Derwent, R. G., Simmonds, P. G., Manning, A. J., & Spain, T. G. (2007).
Trends over a 20-year period from 1987 to 2007 in surface ozone at the
atmospheric research station, Mace Head, Ireland. Atmos Environ,
41(39), 9091–8.
Devlin, R. B., Ghio, A. J., Kehrl, H., Sanders, G., & Cascio, W. (2003). Elderly
humans exposed to concentrated air pollution particles have decreased
heart rate variability. Eur Respir J (Suppl.), 40, 76s–80s.
Dockery, D. W., Pope, C. A., Xu, X., Spengler, J. D., Ware, J. H., Fay, M. E.,
Ferris, B. G., & Speizer, F. E. (1993). An association between air pollution
and mortality in six U.S. Cities. N Engl J Med, 329(24), 1753–9.
Dominici, F., Sheppard, L., & Clyde, M. (2003). Health effects of air pollution:
A statistical review. Int Stat Rev, 71(2), 243–76.
Dominici, F., Peng, R. D., Barr, C. D., & Bell, M. L. (2010). Protecting human
health from air pollution: Shifting from a single-pollutant to a multi-pollutant
approach. Epidemiology, 21(2), 187–94.
Donaldson, K., & Seaton, A. (2012). A short history of the toxicology of
inhaled particles. Part Fibre Toxicol, 9(13).
Donaldson, K., Stone, V., Faux, S., & MacNee, W. (2006). Testing new
particles. In J. Ayres, R. Maynard, & R. Richards (Eds.), Air Pollution and
Health (Vol. 3, pp. 163–195). London: Imperial College Press.
64
Doocy, S., Daniels, A., Dooling, S., & Gorokhovich, Y. (2013). The human
impact of volcanoes: A historical review of events 1900-2009 and
systematic literature review. PLoS Curr, 5.
Duce, R. A., Unni, C. K., Ray, B. J., Prospero, J. M., & Merrill, J. T. (1980).
Long-range atmospheric transport of soil dust from Asia to the tropical
north Pacific: Temporal variability, Science 209 (4464), 1522–4.
Dunér, D. (2012). Tankemaskinen - Polhems huvudvärd coh andra studier i
tänkandets historia. Nora: Nya Doxa. [Book in Swedish]
Durand, M., & Grattan, J. (2001). Effects of volcanic air pollution on health.
Lancet, 357(9251), 164.
Durand, M., & Wilson, J. G. (2006). Spatial analysis of respiratory disease on
an urbanized geothermal field. Environ Res, 101(2), 238–45.
Eccles, R. (2006). The role of the nose in health and disease. In J. Ayres, R.
Maynard, & H. Richard (Eds.). Air Pollution and Health (Vol. 3, pp. 17–29).
London: Imperial College Press.
Elliot, A. J., Singh, N., Loveridge, P., Harcourt, S., Smith, S., Pnaiser, R.,
Kavanagh, K., Robertson, C., Ramsay, C. N., McMenamin, J., Kibble, A.,
Murray, V., Ibbottson, S., Catchpole, M., McCloskey, B., & Smith, G. E.
(2010). Syndromic surveillance to assess the potential public health
impact of the Icelandic volcanic ash plume across the United Kingdom,
April 2010. Euro Surveill, 15(23).
Emmanuel, S. C. (2000). Impact to lung health of haze from forest fires: The
Singapore experience. Respirology, 5(2), 175–82.
Fenger, J. (2002). Urban air quality. In J. Austin, P. Brimblecombe, & W. T.
Sturges (Eds.). Air Pollution Science for the 21st Century (pp. 1–52).
Oxford: Elsevier.
Fiedler, N., Kipen, H., Ohman-Strickland, P., Zhang, J., Weisel, C.,
Laumbach, R., Kelly-McNeil, K., Olejerne, K., & Lioy, P. (2008). Sensory
and cognitive effects of acute exposure to hydrogen sulfide. Environ
Health Perspect, 116(1), 78–85.
65
Finnbjornsdottir, R. G., Zoëga, H., Olafsson, O., Thorsteinsson, T., &
Rafnsson, V. (2013). Association of air pollution and use of glyceryl
trinitrate against angina pectoris: A population-based case-crossover
study. Environ Health, 12(1), 38.
Forsberg, B. (2013, August 23). Vehicle exhaust level at home and incidence
of diagnosed dementia. Scientific conference presented at the
International Society of Environmental Epidemiology, Basel, Switzerland.
Retrieved
from
http://pp-organizers.convent-info.de/0h8f76u1jpmxva
1b3ztgf0yp2yp0/$/
Francis, P., & Oppenheimer, C. (2004). Volcanoes, 2nd edition. New York:
Oxford University press.
Gasparrini, A., & Armstrong, B. (2010). Time series analysis on the health
effects of temperature: Advancements and limitations. Environ Res,
110(6), 633–8.
Gasparrini, A., Armstrong, B., & Kenward, M. G. (2010). Distributed lag non‐
linear models. Stat Med, 29(21), 2224–34.
Gauderman, W. J., Vora, H., McConnell, R., Berhane, K., Gilliland, F.,
Thomas, D., Lurman, F., Avol. E., Künzli, N., Jerrett, M., & Peters, J.
(2007). Effect of exposure to traffic on lung development from 10 to 18
years of age: A cohort study. Lancet, 369(9561), 571–7.
Gíslason, D., Björnsdóttir, U., Blöndal, T., & Gíslason, T. (2002). European
Community Respiratory Health Survey: The main results so far with
special reference to Iceland. Laeknabladid, 88(12), 891–907.
Gislason, S. R., Hassenkam, T., Nedel, S., Bovet, N., Eiriksdottir, E. S.,
Alfredsson, H. A., Hem, C. P., Balogh, Z. I.,. Dideriksen, K., Oskarsson,
N., Sigfusson, B., Larsen, G., & Stipp, S. L. S. (2011). Characterization of
Eyjafjallajökull volcanic ash particles and a protocol for rapid risk
assessment. P Natl Acad Sci USA, 108(18), 7307 –12.
Gislason, T., Olafsson, O., & Sigvaldason, A. (1997). Users of antiasthma
drugs in Iceland: A drug utilization study. Eur Respir J, 10(6), 1230–4.
GOLD-Global Initiative for Chronic Obstructive Lung Disease. (2010). GOLD
- Spirometry Guide. Organisation. Retrieved November 25, 2013, from
http://www.goldcopd.org/other-resources-gold-spirometry-guide.html
66
Gold, D. R., & Mittleman, M. A. (2013). New insights into pollution and the
cardiovascular system: 2010 to 2012. Circulation, 127(18), 1903–13.
Goldberg, D. P., Gater, R., Sartorius, N., Ustun, T. B., Piccinelli, M., Guree,
O., & Rutter, C. (1997). The validity of two versions of the GHQ in the
WHO Study of Mental Illness in General Health Care. Psychol Med,
27(01), 191–7.
Goldberg, D., & Williams, P. (1988). A User’s Guide to the General Health
Questionnaire. Oxford: NFER-Nelson Publishing Company Ltd. Retrieved
from http://scholar.google.com/scholar?hl=en&q=goldberg+and+williams
+1988+%22A+user%27s+guide%22&btnG=Search&as_sdt=0%2C5&as_
ylo=&as_vis=1
Gordian, M. E & Choudhury, A. H. (2003). PM10 and asthma medication in
schoolchildren. Arch Environ Health, 58(1), 42–47.
Gordian, M. E., Ozkaynak, H., Xue, J., Morris, S. S., & Spengler, J. D. (1996).
Particulate air pollution and respiratory disease in Anchorage, Alaska.
Environ Health Perspect, 104(3), 290–7.
Götschi, T., Heinrich, J., Sunyer, J., & Künzli, N. (2008). Long-term effects of
ambient air pollution on lung function: A review. Epidemiology ), 19(5),
690–701.
Gudmundsson, M. T., Thordarson, T., Höskuldsson, Á., Larsen, G.,
Björnsson, H., Prata, F. J., Oddsson, B., Magnússon, E., Högnadóttir, T.,
Petersen, G. N., Hayward, C. L., Stevenson, J. A., & Jónsdóttir, I. (2012).
Ash generation and distribution from the April-May 2010 eruption of
Eyjafjallajökull, Iceland. Sci Rep, 2.
Hallas, J. (2005). Drug utilization statistics for individual-level pharmacy
dispensing data. Pharmacoepidem Dr S, 14(7), 455–63.
Halonen, J. I., Lanki, T., Tiittanen, P., Niemi, J. V., Loh, M., & Pekkanen, J.
(2010). Ozone and cause-specific cardiorespiratory morbidity and
mortality. J Epidemiol Commun Health, 64(9), 814 –20.
Halonen, J. I., Lanki, T., Yli-Tuomi, T., Kulmala, M., Tiittanen, P., &
Pekkanen, J. (2008). Urban air pollution, and asthma and COPD hospital
emergency room visits. Thorax, 63(7), 635–41.
67
Hankinson, J. L., Odencrantz, J. R., & Fedan, K. (1999). Spirometric
reference values from a sample of the general U.S. population. Am J
Respir Crit Care Med, 159(1), 179–87.
Hansell, A. L., Horwell, C. J., & Oppenheimer, C. (2006). The health hazards
of volcanoes and geothermal areas. Occup Environ Med, 63(2), 149–56,
125.
Hansell, A., & Oppenheimer, C. (2004). Health hazards from volcanic gases:
A systematic literature review. Arch Environ Health, 59(12), 628–39.
Hefflin, B. J., & Jalaludin, B. (1994). Surveillance for dust storms and
respiratory diseases in Washington State, 1991. Arch Environ Health,
49(3), 170–4.
Heggie, T. W., (2009). Geotourism and volcanoes: health hazards facing
tourists at volcanic and geothermal destinations. Travel Med Infect Dis,
7(5), 257–61.
Henrotin, J. B., Besancenot, J. P., Bejot, Y., & Giroud, M. (2007). Short-term
effects of ozone air pollution on ischaemic stroke occurrence: a casecrossover analysis from a 10-year population-based study in Dijon,
France. Occup Environ Med, 64(7), 439–45.
Henrotin, J.-B., Zeller, M., Lorgis, L., Cottin, Y., Giroud, M., & Béjot, Y.
(2010). Evidence of the role of short-term exposure to ozone on ischaemic
cerebral and cardiac events: The Dijon Vascular Project (DIVA). Heart,
96(24), 1990 –96.
Hesterberg, T. W., Bunn, W. B., McClellan, R. O., Hamade, A. K., Long, C.
M., & Valberg, P. A. (2009). Critical review of the human data on shortterm nitrogen dioxide (NO2) exposures: Evidence for NO 2 no-effect levels.
Crit Rev Toxicol, 39(9), 743–81.
Holgate, S. T., Arshad, H. S., Roberts, G. C., Howarth, P. H., Thurner, P., &
Davies, D. E. (2010). A new look at the pathogenesis of asthma. Clin Sci,
118(7), 439–50.
Horwell, C., & Baxter, P. (2006). The respiratory health hazards of volcanic
ash: A review for volcanic risk mitigation. Bull Volcanol, 69(1), 1–24.
68
Horwell, C. J., Baxter, P. J., Hillman, S. E., Calkins, J. A., Damby, D. E.,
Delmelle, P., Donaldson, K., Dunster, C., Fubini, B., Kelly, F.J., Le Blond,
J. S., Livi, K. J. T., Murphy, F., Nattrass, C., Sweeney, S., Tetley, T.D.,
Thordarson, & T.,Tomatis, M. (2013). Physicochemical and toxicological
profiling of ash from the 2010 and 2011 eruptions of Eyjafjallajökull and
Grímsvötn volcanoes, Iceland using a rapid respiratory hazard
assessment protocol. Environ Res, 127, 63–73.
Icelandic Directorate of Health. (2012, April 30). Lyfjagagnagrunnur
[Medicines registry]. Government institution. Retrieved January 25, 2014,
from
http://www.landlaeknir.is/tolfraedi-og-rannsoknir/gagnasofn/
gagnasafn/item12455/Lyfjagagnagrunnur [Database in n Icelandic]
Icelandic Institute of Natural History. (2014). Pollen monitoring in Reykjavik
and Akureyri | The Icelandic Institute of Natural History. Pollen monitoring.
Government
institution.
Retrieved
January
25,
2014,
from
http://en.ni.is/botany/pollen-monitoring/
Institute of Earth Sciences. (2010). Eyjafjallajökull 2010 - Grain size
distribution | Jarðvísindastofnun - Institute of Earth Sciences. University of
Iceland. Retrieved January 24, 2014, from http://earthice.hi.is/
eyjafjallajokul l_2010_grain_size_distribution
Ishigami, A., Kikuchi, Y., Iwasawa, S., Nishiwaki, Y., Takebayashi, T.,
Tanaka, S., & Omae, K. (2008). Volcanic sulfur dioxide and acute
respiratory symptoms on Miyakejima island. Occup Environ Med, 65(10),
701–7.
Iwasawa, S. (2013, August 23). Effects of sulfur dioxide on respiratory
symptoms in child residents of Miyakejima seven years after returning to
the island. Scientific conference presented at the International Society of
Environmental Epidemiology, Basel, Switzerland. Retrieved from http://pporganizers.convent-info.de/0h8f76u1jpmxva1b3ztgf0yp2yp0/$/
Iwasawa, S., Michikawa, T., Nakano, M., Nishiwaki, Y., Tsuboi, T., Tanaka,
S., Uemura, T., Milojevic, A., Nakashima, H., Takebayashi, T., Monkawa,
A., Maruyama, K., Kudo, S., & Omae, K. (2009). Effects of SO2 on the
respiratory systems of hild Miyakejima residents nearly 2 years after
returning to the island. Epidemiology, 20(6), S112–S113.
69
Jaakkola, J. J., Partti-Pellinen, K., Marttila, O., Miettinen, P., Vilkka, V., &
Haahtela, T. (1996). The South Karelia Air Pollution Study: Changes in
respiratory health in relation to emission reduction of malodorous sulfur
compounds from pulp mills. Arch Environ Health, 54(4), 254–63.
Jacoby, G. C., Workman, K. W., & D’Arrigo, R. D. (1999). Laki eruption of
1783, tree rings, and disaster for northwest Alaska Inuit. Quaternary Sci
Rev, 18(12), 1365–71.
Jacquemin, B., Schikowski, T., Carsin, A., Hansell, A., Krämer, U., Sunyer,
J., Probst-Hensch, N., Kauffmann, F., & Künzli, N. (2012). The role of air
pollution in adult-onset A-asthma: A review of the current evidence. Semin
Respir Crit Care, 33(06), 606–19.
Janson, C. (2008). Phenotypes of obstructive lung disease. Clin Respir J, 2,
88–91.
Jenkins, M. E., & Clemitshaw, K. (2002). Ozone and other secondary
photochemical pollutants: Chemical processes giverning their formation in
the planetary boundary layer. In J. Austin, P. Brimblecombe, & W. T.
Sturges (Eds.), Air Pollution Science for the 21st Century (pp. 285–338).
Oxford: Elsevier.
Jerrett, M., Arain, A., Kanaroglou, P., Beckerman, B., Potoglou, D.,
Sahsuvaroglu, T., Morrison, J., & Giovis, C. (2004). A review and
evaluation of intraurban air pollution exposure models. J Expo Sci Environ
Epidemiol, 15(2), 185–204.
Jóhansson, T. (2007). Svifryk í Reykjavík. Reykjavík: University of Iceland.
Retrieved from http://www2.hi.is/page/msub070b [Ms. Thesis In Icelandic]
Johnston, F., Hanigan, I., Henderson, S., Morgan, G., & Bowman, D. (2011).
Extreme air pollution events from bushfires and dust storms and their
association with mortality in Sydney, Australia 1994-2007. Environ Res,
111(6), 811–6.
Karanasiou, A., Moreno, N., Moreno, T., Viana, M., de Leeuw, F., & Querol,
X. (2012). Health effects from Sahara dust episodes in Europe: Literature
review and research gaps. Environ Int, 47, 107–14.
Kellogg, C. A., & Griffin, D. W. (2006). Aerobiology and the global transport of
desert dust. Trends Ecol Evol, 21(11), 638–44.
70
Kilburn, K. H., Thrasher, J. D., & Gray, M. R. (2010). Low-level hydrogen
sulfide and central nervous system dysfunction. Toxicol Ind Health, 26(7),
387–405.
Kilburn, K. H., & Warshaw, R. H. (1995). Hydrogen sulfide and reduced-sulfur
gases adversely affect neurophysiological functions. Toxicol Ind Health,
11(2), 185–97.
Kochi, K. (2013, August 23). The relationship between sulfur dioxide
exposure and the health of adult Miyakejima residents. Scientific
conference presented at the International Society of Environmental
Epidemiology,
Basel,
Switzerland.
Retrieved
from
http://pporganizers.convent-info.de/0h8f76u1jpmxva1b3ztgf0yp2yp0/$/
Kreyling, W. G., Hirn, S., & Schleh, C. (2010). Nanoparticles in the lung. Nat
Biotech, 28(12), 1275–6.
Künzli, N., & Tager, I. B. (2005). Air pollution: from lung to heart. Swiss Med
Wkly, 135(47-48), 697–702.
Kunzli, N., Avol, E., Wu, J., Gauderman, W. J., Rappaport, E., Millstein, J.,
Bennion, J., McConnell, R., Gilliland, F. D., Berhane, K., Lurmann, F.,
Winer, A., & Peters, J. M. (2006). Health effects of the 2003 Southern
California wildfires on children. Am J Respir Crit Care Med, 174(11),
1221–8.
Künzli, N., Kaiser, R., Medina, S., Studnicka, M., Chanel, O., Filliger, P.,
Herry, M., Horak, F., Puybonnieux-Texier, V., Quénel, P., Schneider, J.,
Seethaler, R., Vergnaud, J. C., & Sommer, H. (2000). Public-health
impact of outdoor and traffic-related air pollution: A European
assessment. Lancet, 356(9232), 795–801.
Künzli, N., Perez, L., von Klot, S., Baldassarre, D., Bauer, M., Basagana, X.,
Breton, C., Dratva, J., Elosua, R., de Faire, U., Fuks, K., de Groot, E.,
Marrugat, J., Penell, J., Seissler, J., Peters, A., & Hoffmann, B. (2011).
Investigating air pollution and atherosclerosis in humans: Concepts and
outlook. Prog Cardiovasc Dis, 53(5), 334–43.
Künzli, N., & Schindler, C. (2005). A call for reporting the relevant exposure
term in air pollution case-crossover studies. J Epidemiol Commun H ,
59(6), 527–30.
71
Lacy, T. G. (1998). Ring of Seasons: Iceland, Its Culture and History.
Reykjavík, Iceland: University of Iceland Press.
Laurent, O., Pedrono, G., Filleul, L., Segala, C., Lefranc, A., Schillinger, C.,
Riviere, E., & Bard, D. (2008). Influence of socioeconomic deprivation on
the relation between air pollution and beta-agonist sales for asthma.
Chest, 135(3), 717-23.
Laych, K., Vilen, A., & Latipov, O. (2009). IRF World Road Statistics 2009 Data 2002-2007 Booklet. Geneva, Switzerland: International Road
Federation. Retrieved from http://www.irfnet.org/files-upload/stats/2009/
wrs2009_web.pdf
Lee, H., Kim, H., Honda, Y., Lim, Y.-H., & Yi, S. (2013). Effect of Asian dust
storms on daily mortality in seven metropolitan cities of Korea. Atmos
Environ, 79, 510–7.
Lim, S. S., Vos, T., Flaxman, A. D., Danaei, G., Shibuya, K., Adair-Rohani,
H., AlMazroa, M. A., Arnann, M., Anderson, H. R., Andrews, K. G., Ayree,
M., Atkinson, C., Bacchus, L. J., Bahalim, A. N., Balakrishnan, K., BarkerCollo, S., Baxter, A., Bell, M., Blore, J. D., Blyth, F., Bonner, C., Borges,
G., Bourne, R., Boussinesq, M., Brauer, M., Brooks, P., Bruce, N. G.,
Brunekreef, B., Bryan-Hancock, C., Bucello, C., Buchbinder, R., Bull, F.,
Burnett, R. T., Byers, T. E., Calabria, B., Carapetis, J., Carnahan, E.,
Chafe, Z., Charlson, F., Chen, H., Chen, J. S., Cheng, A. T., Child, J. C.,
Cohen, A., Colson, B. C., Cowie, B. C., Darby, S., Darling, S., Davis, A.,
Degenhardt, L., et al. (2012). A comparative risk assessment of burden of
disease and injury attributable to 67 risk factors and risk factor clusters in
21 regions, 1990–2010: A systematic analysis for the Global Burden of
Disease Study 2010. Lancet, 380(9859), 2224–60.
Lin, C.-H., Mao, I.-F., Tsai, P.-H., Chuang, H.-Y., Chen, Y.-J., & Chen, M.-L.
(2010). Sulfur-rich geothermal emissions elevate acid aerosol levels in
metropolitan Taipei. Environ Res, 110(6), 536–43.
Logan, W. P. D. (1956). Mortality from fog in London, January, 1956. BMJ,
1(4969), 722–5.
Longo, B. M. (2009). The Kilauea Volcano Adult Health Study. Nurs Res,
58(1), 23–31.
72
Longo, B. M. (2013). Adverse health effects associated with increased
activity at Kīlauea Volcano: A repeated population-based survey. ISRN
Public Health, 2013, 1–10.
Longo, B. M., Rossignol, A., & Green, J. B. (2008). Cardiorespiratory health
effects associated with sulphurous volcanic air pollution. Public Health,
122(8), 809–20.
Longo, B. M., & Yang, W. (2008). Acute bronchitis and volcanic air pollution:
a community-based cohort study at Kilauea Volcano, Hawai’i, USA. J
Toxicol Environ Health, 71(24), 1565–71.
Longo, B. M., Yang, W., Green, J. B., Crosby, F. L., & Crosby, V. L. (2010).
Acute health effects associated with exposure to volcanic air pollution
(vog) from increased activity at Kilauea Volcano in 2008. J Toxicol Environ
Health, 73(20), 1370–81.
Maclure, M. (1991). The case-crossover design: A method for studying
transient effects on the risk of acute events. Am J Epidemiol, 133(2), 144–
153.
Maclure, M., and Mittleman, A, M. (2000). Should We Use a Case-Crossover
Design? Annu Rev Publ Health, 21(1), 193–221.
Madsen, H. (2007). Time Series Analysis. London: Chapman & Hall.
Marco, R. de, Accordini, S., Cerveri, I., Corsico, A., Antó, J. M., Künzli, N.,
Janson, C., Sunyer, J., Jarvis, D., Chinn, S., Vermeire, P., Svanes, C.,
Ackermann-Liebrich, U., Gislason, T., Heinrich, J., Leynaert, B., Neukirch,
F., Schouten, J. P., Wjst, M., & Burney, P. (2007). Incidence of chronic
obstructive pulmonary disease in a cohort of young adults according to
the presence of chronic cough and phlegm. Am J Respir Crit Care Med,
175(1), 32–9.
McConnell, R., Islam, T., Shankardass, K., Jerrett, M., Lurmann, F., Gilliland,
F., Gauderman, F., Avol, E., Künzli, N., Yao, L., Peters, J., & Berhane, K.
(2010). Childhood incident asthma and traffic-related air pollution at home
and school. Environ Health Perspect, 118(7), 1021–26.
McGavran, P. (2001). Literature Review of the Health Effects Associated with
the Inhalation of Hydrogen Sulfide. Idaho Department of Environmental
Quality.
Retrieved
from
http://idahodocs.contentdm.oclc.org/cdm/
compoundobject/collection/p15100coll7/id/38592/rec/1
73
Meister, K., Johansson, C., & Forsberg, B. (2012). Estimated short-term
effects of coarse particles on daily mortality in Stockholm, Sweden.
Environ Health Perspect, 120(3), 431–6.
Menichini, F., & Mudu, P. (2010). Drug consumption and air pollution: An
overview. Pharmacoepidemiol Drug Saf, 19(12), 1300–15.
Miller, K. A., Siscovick, D. S., Sheppard, L., Shepherd, K., Sullivan, J. H.,
Anderson, G. L., & Kaufman, J. D. (2007). Long-term exposure to air
pollution and incidence of cardiovascular events in women. N Engl J Med,
356(5), 447–58.
Monick, M. M., Baltrusaitis, J., Powers, L. S., Borcherding, J. A., Caraballo, J.
C., Mudunkotuwa, I., Peate, D. W., Walters, K., Thompson, J. M.,
Grassian, V. H., Gudmunsson, G., & Comellas, A. P. (2013). Effects of
Eyjafjallajökull volcanic ash on innate immune system responses and
bacterial growth in vitro. Environ Health Perspect,121(6), 691–8.
Morman, S. A., & Plumlee, G. S. (2013). The role of airborne mineral dusts in
human disease. Aeolian Res, 9, 203–12.
Naeher, L. P., Brauer, M., Lipsett, M., Zelikoff, J. T., Simpson, C. D., Koenig,
J. Q., & Smith, K. R. (2007). Woodsmoke health effects: A review. Inhal
Toxicol, 19(1), 67–106.
National Land Survey of Iceland. (2012). Municipalities web page. National
Land Survey of Iceland. Government institution. Retrieved January 14,
2014, from https://atlas.lmi.is/sveitarfelog/
Naumova, E. N., Yepes, H., Griffiths, J. K., Sempértegui, F., Khurana, G.,
Jagai, J. S., Játiva, E., & Estrella, B. (2007). Emergency room visits for
respiratory conditions in children increased after Guagua Pichincha
volcanic eruptions in April 2000 in Quito, Ecuador observational study:
Time series analysis. Environ Health, 6, 21.
Naureckas, E. T., Dukic, V., Bao, X., & Rathouz, P. (2005). Short-acting betaagonist prescription fills as a marker for asthma morbidity. Chest, 128(2),
602–8.
Nemery, B., Hoet, P. H., & Nemmar, A. (2001). The Meuse Valley fog of
1930: An air pollution disaster. Lancet, 357(9257), 704–8.
74
Newnham, R., Dirks, K., & Samaranayake, D. (2010). An investigation into
long-distance health impacts of the 1996 eruption of Mt Ruapehu, New
Zealand. Atmos Environ 44(12), 1568–78.
Ohta, Y., Araki, K., Kawasaki, N., Nakane, Y., Honda, S., & Mine, M. (2003).
Psychological distress among evacuees of a volcanic eruption in Japan: A
follow-up study. Psychiatry Clin. Neurosci, 57(1), 105–11.
Olafsdóttir, I. S., Gíslason, T., Thjódleifsson, B., Olafsson, I., Gíslason, D.,
Jõgi, R., & Janson, C. (2007). Gender differences in the association
between C-reactive protein, lung function impairment, and COPD. Int J
Chron Obstruct Pulmon Dis, 2(4), 635–42.
Ólafsdóttir, I. S., Gislason, T., Gudnason, V., Benediktsdóttir, B., Ólafsson, I.,
Aspelund, T., Thjódleifsson, B., & Janson, C. (2011). Systemic
inflammation and lung function. Clin Respir Jal, 5, 7–8.
Olafsdottir, S., & Gardarsson, S. M. (2013). Impacts of meteorological factors
on hydrogen sulfide concentration downwind of geothermal power plants.
Atmos Environ, 77, 185–92.
Olafsdottir, S., Gardarsson, S. M., & Andradottir, H. O. (2014). Spatial
distribution of hydrogen sulfide from two geothermal power plants in
complex terrain. Atmos Environ, 82, 60–70.
Olson, K. R. (2011). The therapeutic potential of hydrogen sulfide: Separating
hype from hope. Am J Physiol Regul Integr Comp Physiol, 301(2), R297–
R312.
Oudin, A., Carlsen, H. K., Forsberg, B., & Johansson, C. (2013). Volcanic ash
and daily mortality in Sweden after the Icelandic volcano eruption of May
2011. Int J Environ Res Public Health, 10(12), 6909–19.
Oudin, A., Stroh, E., Strömberg, U., Jakobsson, K., & Björk, J. (2009). Longterm exposure to air pollution and hospital admissions for ischemic stroke:
A register-based case-control study using modelled NO(x) as exposure
proxy. BMC Public Health, 9, 301.
Oudin, A., Strömberg, U., Jakobsson, K., Stroh, E., & Björk, J. (2010).
Estimation of short-term effects of air pollution on stroke hospital
admissions in Southern Sweden. Neuroepidemiology, 34(3), 131–42.
75
Partti-Pellinen, K., Marttila, O., Vilkka, V., Jaakkola, J. J. K., Jäppinen, P., &
Haahtela, T. (1996). The South Karelia Air Pollution Study: Effects of lowlevel exposure to malodorous sulfur compounds on symptoms. Arch
Environ Health, 51(4), 315–20.
Peng, R. D., Bell, M. L., Geyh, A. S., McDermott, A., Zeger, S. L., Samet, J.
M., & Dominici, F. (2009). Emergency admissions for cardiovascular and
respiratory diseases and the chemical composition of fine particle air
pollution. Environ Health Perspect, 117(6), 957–63.
Peng, R. D., Chang, H. H., Bell, M. L., McDermott, A., Zeger, S. L., Samet, J.
M., & Dominici, F. (2008). Coarse particulate matter air pollution and
hospital admissions for cardiovascular and respiratory diseases among
Medicare patients. J Am Med Assoc, 299(18), 2172–9.
Peng, R. D., & Dominici, F. (2008). Statistical Methods for Environmental
Epidemiology with R - A Case Study in Air Pollution and Health. New
York: Springer.
Pérez, L., Tobías, A., Pey, J., Pérez, N., Alastuey, A., Sunyer, J., & Querol,
X. (2012). Effects of local and Saharan particles on cardiovascular
disease mortality: Epidemiology, 23(5), 768–69.
Pitard, A., Zeghnoun, A., Courseaux, A., Lamberty, J., Delmas, V., Fossard,
J. L., & Villet, H. (2004). Short-term associations between air pollution and
respiratory drug sales. Environ Res, 95(1), 43–52.
Pope, C. A., & Dockery, D. W. (2006). Health effects of fine particulate air
pollution: Lines that connect. J Air Waste Ma, 56(6), 709–42.
Pope, C. A., Ezzati, M., & Dockery, D. W. (2009). Fine-particulate air pollution
and life expectancy in the United States. N Engl J Med, 360(4), 376–86.
Pope, C. A., Muhlestein, J. B., May, H. T., Renlund, D. G., Anderson, J. L., &
Horne, B. D. (2006). Ischemic heart disease events triggered by shortterm exposure to fine particulate air pollution. Circulation, 114(23), 2443–8.
Pope, C. A., Renlund, D. G., Kfoury, A. G., May, H. T., & Horne, B. D. (2008).
Relation of heart failure hospitalization to exposure to fine particulate air
pollution. Am J Cardiol, 102(9), 1230–4.
76
Prospero, J. H., Bullard, J. E., & Hodgkins, R. (2012). High-latitude dust over
the North Atlantic: Inputs from Icelandic proglacial dust storms. Science,
335(2012), 1078–82.
Roth, S. H., & Goodwin, V. M. (2003). Health Effects of Hydrogen Sulphide:
Knowledge Gaps. Edmonton, Alberta: Alberta Environment. Retrieved
from http://environment.gov.ab.ca/info/library/6708.pdf
Rothman, K. (2002). Epidemiology: An Introduction. Oxford: Oxford
University press.
Retrieved
from http://library.wur.nl/WebQuery/
clc/1691458
Samoli, E., Nastos, P. T., Paliatsos, A. G., Katsouyanni, K., & Priftis, K. N.
(2011). Acute effects of air pollution on pediatric asthma exacerbation:
Evidence of association and effect modification. Environ Res, 111(3),
418–24.
Schikowski, T., Adam, M., Marcon, A., Cai, Y., Vierkötter, A., Carsin, A. E.,
Jacquemin, B., Al Kanani, Z., Beelen, R., Birk, M., Bridevaux, P-O.,
Brunekreef, B., Burney, P., Cirach, M., Cyrys, J., de Hoogh, K., de Marco,
R., de Nazelle, A., Declercq, C., Forsberg, B., Hardy, R., Heinrich, J.,
Hoek, G., Jarvis, D., Keidel, D., Kuh, D., Kuhlusch, T., Micliore, E.,
Mosler, G., Nieuwenhuijsen, M. J., Phuleria, H., Rochat, T., Schindler, C.,
Villani, S., Tsai, M-Y., Zemp, E., Hansell, A., Kauffmann, F., Sunyer, J.,
Probst-Hensch, N., Krämer, U., & Künzli, N. (2014). Association of
ambient air pollution with the prevalence and incidence of COPD. Eur
Respir J, 44(3), 614–26.
Schmidt, A., Ostro, B., Carslaw, K. S., Wilson, M., Thordarson, T., Mann, G.
W., & Simmons, A. J. (2011). Excess mortality in Europe following a future
Laki-style Icelandic eruption. Proc Natl Acad Sci USA, 108(38), 15710–5.
Schwartz, J. (2000). Assessing confounding, effect modification, and
thresholds in the association between ambient particles and daily deaths.
Environ Health Perspect, 108(6), 563–68.
Shimizu, Y., Dobashi, K., Hisada, T., Ono, A., Todokoro, M., Iijima, H.,
Utsugi, M., Kakegawa, S., Iizuka, K., Ishizuka, T., Morikawa, A., & Mori,
M. (2007). Acute impact of volcanic ash on asthma symptoms and
treatment. Int J Immunopathol Pharmacol, 20(2 Suppl 2), 9–14.
77
Shore, J. H., Tatum, E. L., & Vollmer, W. M. (1986). Evaluation of mental
effects of disaster, Mount St. Helens eruption. Am J Public Health,
76(Suppl), 76–83.
Sigmundsson, F., Hreinsdottir, S., Hooper, A., Arnadottir, T., Pedersen, R.,
Roberts, M. J., Oskarsson, N., Auriac, A., Decriem, J., Einarsson, P.,
Geirsson, H., Hensch, M., Ofeigsson, B. G., & Sturkell, E., Sveinbjorsson,
H., Feigl, K. L. (2010). Intrusion triggering of the 2010 Eyjafjallajokull
explosive eruption. Nature, 468(7322), 426–30.
Sigurkarlsson, S., Clausen, M., Gislason, T., & Gislason, D. (2011).
[Prevalence of respiratory symptoms and use of asthma drugs are
increasing among young adult Icelanders]. Laeknabladid, 97(9), 463–7. [in
Icelandic]
Sillman, S. (2002). The relationship between ozone, NOx and hydrocarbons
in urban and polluted rural environments. In J. Austin, P. Brimblecombe, &
W. T. Sturges (Eds.), Air Pollution Science for the 21st Century (pp. 339–
386). Oxford:Elsevier.
Skúladóttir, B., Larssen, S., Bjarnason, G., Hallsdóttir, B., Gustafsson, L., &
Guðmundsdóttir, Á. (2003). Method for determining the composition of
airborne particle pollution in Reykjavík. Reykjavík, Iceland: Icelandic
Technological Institute. Retrieved from english.ust.is/media/skyrslur2003/
uppruni_svifryks_eng.pdf
Small, C., & Naumann, T. (2001). The global distribution of human population
and recent volcanism. Global Environ Change, 3(3-4), 93–109.
Solberg, S., Bergström, R., Langner, J., Laurila, T., & Lindskog, A. (2005).
Changes in Nordic surface ozone episodes due to European emission
reductions in the 1990s. Atmos Environ, 39(1), 179–92.
Statistics Iceland. (2013a). Statistics Iceland - Statistics » Tourism, transport
and information technology » Transport. Retrieved November 20, 2013,
from
:
http://www.statice.is/?PageID=1237&src=https://rannsokn.hagstofa.is/pxe
n/Dialog/varval.asp?ma=SAM03103%26ti=Motor+vehicles+by+registered
+domicile+of+owner+2001%2D2006++++%26path=../Database/ferdamal/
Okutaeki/%26lang=1%26units=Number
78
Statistics Iceland. (2013b, November 20). Statistics Iceland - Statistics »
Population » Urban nuclei and zip codes. Population by urban nuclei, age
and sex 1 January 2011-2013. Government institution. Retrieved
November 20, 2013, from http://www.statice.is/?PageID=1173&src=
https://rannsokn.hagstofa.is/pxen/Dialog/varval.asp?ma=MAN03105%26ti
=Population+by+locality%2C+age+and+sex+1+January+2011%2D2013%
26path=../Database/mannfjoldi/Byggdakjarnar/%26lang=1%26units=Data
Stohl, A., Prata, A. J., Eckhardt, S., Clarisse, L., Durant, A., Henne, S.,
Kristiansen, N. I., Minikin, A., Schumann, U., Seibert, P., Stebel, K.,
Thomas, H. E., Thorsteinsson, T., & Weinzierl, B. (2011). Determination of
time- and height-resolved volcanic ash emissions for quantitative ash
dispersion modeling: The 2010 Eyjafjallajökull eruption. Atmos Chem
Phys Discuss, 11(2), 5541–88.
Sturkell, E., Einarsson, P., Sigmundsson, F., Hooper, A., Ófeigsson, B. G.,
Geirsson, H., & Ólafsson, H. (2010). 2 Katla and Eyjafjallajökull
Volcanoes. In A. Schomacker, J. Krüger, & K. H. Kjær (Eds.), The
Mýrdalsjökull Ice Cap, Iceland. Glacial processes, sediments and
landforms on an active volcano. Developments in Quaternary Sciences
(Vol.
13,
pp.
5–21).
London:
Elsevier.
Retrieved
from
http://www.sciencedirect.com /science/article/pii/S1571086609013025
Sun, Q., Hong, X., & Wold, L. E. (2010). Cardiovascular Effects of Ambient
Particulate Air Pollution Exposure. Circulation, 121(25), 2755–2765.
Sunyer, J., Jarvis, D., Gotschi, T., Garcia-Esteban, R., Jacquemin, B.,
Aguilera, I., Ackerman, U., de Marco, R., Forsberg, B., Gislason, T.,
Heinrich, J., Norbäck, D., Villani, S., & Künzli, N. (2006). Chronic
bronchitis and urban air pollution in an international study. Occup Environ
Med, 63(12), 836–43.
Sunyer, J., Schwartz, J., Tobías, A., Macfarlane, D., Garcia, J., & Antó, J. M.
(2000). Patients with chronic obstructive pulmonary disease are at
increased risk of death associated with urban particle air pollution: A casecrossover analysis. Am J Epidemiol, 151(1), 50 –56.
Tam, W. W. S., Wong, T. W., Wong, A. H. S., & Hui, D. S. C. (2012). Effect of
dust storm events on daily emergency admissions for respiratory
diseases. Respirology 17(1), 143–8.
79
The Icelandic Institute of Natural History. (2014). Pollen monitoring in
Reykjavik and Akureyri | The Icelandic Institute of Natural History. Pollen
monitoring. Government institution. Retrieved January 25, 2014, from
http://en.ni.is/botany/pollen-monitoring/
Thordarson, T., & Hoskuldsson, A. (2002). Iceland. Classic Geology in
Europe Vol 3. Dunedin Academic Press, Ltd., Edinburgh, UK.
Thorleifsson, S. J., Margretardottir, O. B., Gudmundsson, G., Olafsson, I.,
Benediktsdottir, B., Janson, C., Buist, S., & Gislason, T. (2009). Chronic
airflow obstruction and markers of systemic inflammation: Results from
the BOLD study in Iceland. Respir Med, 103(10), 1548–53.
Thorsteinsson, T., Gísladóttir, G., Bullard, J., & McTainsh, G. (2011). Dust
storm contributions to airborne particulate matter in Reykjavík, Iceland.
Atmos Environ, 45(32), 5924–33.
Thorsteinsson, T., Jóhannsson, T., Stohl, A., & Kristiansen, N. I. (2012). High
levels of particulate matter in Iceland due to direct ash emissions by the
Eyjafjallajökull eruption and resuspension of deposited ash. J Geophys
Res, 117, 9 PP.
To, T., Stanojevic, S., Moores, G., Gershon, A. S., Bateman, E. D., Cruz, A.
A., & Boulet, L.-P. (2012). Global asthma prevalence in adults: Findings
from the cross-sectional world health survey. BMC Public Health, 12(1),
204.
Tobin, G. A., & Whiteford, L. M. (2012). Provisioning capacity: A critical
component of vulnerability and resilience under chronic volcanic
eruptions. In K. Pfeifer & N. Pfeifer (Eds.), Forces of Nature and Cultural
Responses (pp. 139–66). Springer Netherlands.
Tolbert, P. E., Klein, M., Peel, J. L., Sarnat, S. E., & Sarnat, J. A. (2007).
Multipollutant modeling issues in a study of ambient air quality and
emergency department visits in Atlanta. J Expo Sci Environ Epidemiol,
17(S2), S29–S35.
Tómasson, K., & Gudmundsson, G. (2009). [Mental health and wellbeing in
Icelandic farmers]. Laeknabladid, 95(11), 763–9.
Tweed, F. S. (2012). “Now that the dust has settled…” The impacts of
Icelandic volcanic eruptions. Geol Today, 28(6), 217–23.
80
UHR –Umhverfis og heilbrigðissvið Reykjavíkurborg [The Reykjavík
Municipality Environmental Branch]. (2007). [Measurements of air
pollutants
in
Reykjavík
2006].
Retrieved
from
http://www.reykjavik.is/Portaldata/1/Resources/umhverfissvid/myndir/skyrl
sur/_RSSK_R SLA_2006_1.pdf [Report In Icelandic]
Uno, I., Eguchi, K., Yumimoto, K., Takemura, T., Shimizu, A., Uematsu, M.,
Liu, Z., Want, Z., Hara, Y., & Sugimoto, N. (2009). Asian dust transported
one full circuit around the globe. Nat Geosci, 2(8), 557–60.
Vegagerðin. (2013). Samsetning svifryks í Reykjavík (Research report) (p.
21). Reykjavík, Iceland: The Icelandic Road Administration. Retrieved
from
http://www.vegagerdin.is/Vefur2.nsf/Files/Samsetning_svifryks_Rvk/$file/
Samsetning%20svifryks%20%C3%AD%20Reykjav%C3%ADk.pdf
[Resport In Icelandic]
Vernon, M. K., Wiklund, I., Bell, J. A., Dale, P., & Chapman, K. R. (2012).
What do we know about asthma triggers? A review of the literature. J
Asthma, 49(10), 991–8.
Villeneuve, P., Chen, L., Rowe, B., & Coates, F. (2007). Outdoor air pollution
and emergency department visits for asthma among children and adults:
A case-crossover study in northern Alberta, Canada. Environ Health, 6(1),
40.
Vittinghoff, E., Glidden, D. V., Shiboski, S. C., & McCulloch, C. E. (2004).
Regression Methods in Biostatistics. New York: Springer.
Volk, H. E., Kerin, T., Lurmann, F., Hertz-Picciotto, I., McConnell, R., &
Campbell, D. B. (2014). Autism spectrum disorder: Interaction of air
pollution with the MET receptor tyrosine kinase gene. Epidemiology,
25(1), 44–7.
von Mutius, E., & Braun-Fahrländer, C. (2002). Environmental predicors of
allergic disease and effects of air pollution in children. In J. D‘Amato G., &
Holgate, S. T. (Eds.), The Impact of Air Pollution on Respiratory Health
(Vol. 7, pp. 17–29). Huddersfield: European Respiratory Society.
Wallgren, A., & Wallgren, B. (2007). Register-based Statistics –
Administrative Data for Statistical Purposes (Vol. 553). Hoboken, New
Jersey: John Wiley & Sons.
81
Whitaker, H. J., Hocine, M. N., & Farrington, C. P. (2007). On case-crossover
methods for environmental time series data. Environmetrics, 18(2), 157–
71.
WHO. (2000). Air Quality Guidelines for Europe (No. 91). Copenhagen:
WHO.
Retrieved
from
http://www.euro.who.int/__data/assets/
pdf_file/0005/74732/E71922.pdf
WHO. (2005). Health Aspects of Air Pollution. Copenhagen: World Health
Organisation.
Retrieved
from
http://www.euro.who.int/document/
E83080.pdf
WHO Regional Office for Europe. (2013). Review of Evidence on Health
Aspects of Air Pollution – REVIHAAP Project Technical Report (Technical
Report) (p. 67). Copenhagen: WHO.
Winquist, A., Klein, M., Tolbert, P., Flanders, W. D., Hess, J., & Sarnat, S. E.
(2012). Comparison of emergency department and hospital admissions
data for air pollution time-series studies. Environ Health, 11(1), 70.
Yang, Q., Chen, Y., Krewski, D., Burnett, R. T., Shi, Y., & McGrail, K. M.
(2005). Effect of short-term exposure to low levels of gaseous pollutants
on chronic obstructive pulmonary disease hospitalizations. Environ Res,
99(1), 99–105.
Yorifuji, T., Kawachi, I., Sakamoto, T., & Doi, H. (2011). Associations of
outdoor air pollution with hemorrhagic stroke mortality. J Occup Environ
Med, 53(2), 124–6.
Yu, I. T., Qiu, H., Wang, X., Tian, L., & Tse, L. A. (2013). Synergy between
particles and nitrogen dioxide on emergency hospital admissions for
cardiac diseases in Hong Kong. Int J Cardiol, 168(3), 2831–6.
Zeger, S. L., Thomas, D., Dominici, F., Samet, J. M., Schwartz, J., Dockery,
D., & Cohen, A. (2000). Exposure measurement error in time-series
studies of air pollution: Concepts and consequences. Environ Health
Perspect, 108(5), 419–26.
Zeghnoun, A., Beaudeau, P., Carrat, F., Delmas, V., Boudhabhay, O.,
Gayon, F., Guincètre, & D.,Czernichow, P. (1999). Air pollution and
respiratory drug sales in the City of Le Havre, France, 1993-1996. Environ
Res, 81(3), 224–30.
82
Appendix
153
<≪<
153
154