Series
Energy and Health 2
Electricity generation and health
Anil Markandya, Paul Wilkinson
The provision of electricity has been a great benefit to society, particularly in health terms, but it also carries health Lancet 2007; 370: 979–90
costs. Comparison of different forms of commercial power generation by use of the fuel cycle methods developed in Published Online
European studies shows the health burdens to be greatest for power stations that most pollute outdoor air (those September 13, 2007
DOI:10.1016/S0140-
based on lignite, coal, and oil). The health burdens are appreciably smaller for generation from natural gas, and lower 6736(07)61253-7
still for nuclear power. This same ranking also applies in terms of greenhouse-gas emissions and thus, potentially, to
This is the second in a Series of
long-term health, social, and economic effects arising from climate change. Nuclear power remains controversial, six papers about energy and
however, because of public concern about storage of nuclear waste, the potential for catastrophic accident or terrorist health
attack, and the diversion of fissionable material for weapons production. Health risks are smaller for nuclear fusion, See Comment pages 921
but commercial exploitation will not be achieved in time to help the crucial near-term reduction in greenhouse-gas and 922
emissions. The negative effects on health of electricity generation from renewable sources have not been assessed as See Perspectives page 927
fully as those from conventional sources, but for solar, wind, and wave power, such effects seem to be small; those of See Series page 965
biofuels depend on the type of fuel and the mode of combustion. Carbon dioxide (CO2) capture and storage is University of Bath, Bath, UK,
increasingly being considered for reduction of CO2 emissions from fossil fuel plants, but the health effects associated and FEEM, Italy
(Prof A Markandya PhD); and
with this technology are largely unquantified and probably mixed: efficiency losses mean greater consumption of the London School of Hygiene and
primary fuel and accompanying increases in some waste products. This paper reviews the state of knowledge Tropical Medicine, London, UK
regarding the health effects of different methods of generating electricity. (P Wilkinson FRCP)
Introduction
Economic growth through industrialisation and rapid Key messages
technological change has produced a huge improvement • Access to electricity is pre-requisite for the achievement of health, and lack of access
in the living standards and health status of the population to it remains one of the principal barriers to the fulfilment of human potential and
of the now industrialised countries. From 1820 to 2002, wellbeing
western European countries saw their real incomes • However, electricity generation from fossil fuel—resources of which could sustain their
per head rise from US$1204 to $19 256, or 16 times.1 This continued dominant role in electricity production well beyond this century—is also a
economic growth was also accompanied by major cause of substantial adverse health burdens
improvements in health: life expectancy, for example, • Fossil-fuel use can be used with greater efficiency than it is currently, and with lower
has risen from around 40 years at the beginning of the emissions of pollutants harmful to human health. This is especially the case in
19th century in Europe to nearly 80 years today. The developing countries, and realising these efficiency gains will be increasingly
increase in life expectancy is not uniform with income important as demand for electricity increases sharply
per head of population. It increases rapidly with income • An accelerated switch to renewable sources has the potential to deliver appreciable
up to a level of $7500 and then rises more slowly with health benefits, though a major switch will pose (superable) challenges particularly in
further increases in income.2,3 The availability of modern relation to the intermittency of renewable production, land use requirements, and cost
forms of energy, especially electricity after 1900, has • The demand for valuable agricultural land will limit the role of fuel crops in future
contributed substantially to these positive developments. electricity production in Europe, but the potential contribution of such crops is greater
The replacement of traditional fuels, such as wood and in regions where crops with higher energy yields per hectare can be grown
candles, and animal power by steam power, and then by • Nuclear power has one of the lowest levels of greenhouse-gas emissions per unit
electricity and gas, has reduced the risk of fires, made the power production and one of the smallest levels of direct health effects, yet there are
air in homes cleaner and warmer in winter, and reduced understandable fears about nuclear accidents, weapons uses of fissionable material,
the risk of health hazards associated with animal waste. and storage of waste; nonetheless, it would add a substantial further barrier to the
Thus it has improved the quality of life of individuals in achievement of urgent reductions in greenhouse gases if the current 17 percent of
many ways, and continues to do so in developing world electricity generation from nuclear power were allowed to decline
countries. A 2001 World Bank study4 looked at • CO2 capture and storage could in future have an economic role in reducing CO2
demographic and health data from more than emissions from large point sources, but its effects on health are likely to be mixed
60 low-income countries and investigated the because efficiency losses mean greater consumption of the primary fuel and other
determinants of health outcomes by use of cross-country resources, and greater production of waste
data between 1985 and 1999. It found that in urban areas, • Fusion power offers some hope as a comparatively clean technology for future electricity
linking households to electricity is the only key factor generation, with environment and health risks that are substantially smaller than for
that reduced both infant mortality rate and under-5 nuclear fission. However, commercial viability is still too far away for it to make a
mortality rate, and that this effect is large, significant, significant contribution to mitigation of climate change over most of this century
and independent of incomes. In rural areas, improvement
www.thelancet.com Vol 370 September 15, 2007 979
Series
Correspondence to: of secondary education for women is crucial for reducing Electricity has also contributed to economic
Prof Anil Markandya, Department the infant mortality rate, whereas expansion of vaccination development more generally by increasing the efficiency
of Economics and International
Development, University of Bath,
coverage reduces the under-5 mortality rate. Even with with which energy is used, so that an increased level of
Bath BA2 7AY, UK allowance for the limitations of such cross-sectional production is possible with the same amount of energy.
A.Markandya@bath.ac.uk studies, the results are noteworthy and not unique. Energy use in France, Germany, and the UK increased by
4·7 times between 1840 and 1990, wheres real GDP
increased by 21·5 times.5 Thus each unit of energy now
produces more than 4·5 times as much output as it did
in 1850.
Overall, there is little doubt that electricity has had a
Emissions large positive effect on wellbeing. At the same time, new
eg, tonnes per year of SO2 problems have emerged. The burning of large amounts
of fossil fuels to produce the electricity we demand
generates emissions that are harmful to health and are a
source of climate change. Our paper focuses on these
issues. We separate the discussion into the situation in
Dispersion
Increase in ambient
developing countries and that in developed countries,
concentrations and we offer some views on emerging trends in the
eg, PM10 μg per m–3 relation between electricity use and health.
Assessment of health effects of electricity
generation
Health burden Health burden Developed countries
eg, change in number of The health effects of electricity generation can most
premature deaths easily be assessed by a bottom-up approach, in which
emissions and hazards from each stage of the power
generation cycle are measured and tracked to the
Concentration endpoints at which they cause harm to individuals. The
Cost effects are calculated for specific technology and
location—ie, for a given power station using specified
fuel sources.
The effects are referred to as external costs because the
party generating the emissions does not take full account
of these effects of his or her actions when deciding on
Figure 1 : The effect pathway approach how to generate electricity.
Methods based on this approach were first used in the
Category of impact Pollutant or burden Effects
on human health
early 1970s and have become increasingly sophisticated.
One major set of studies for Europe is the ExternE
Mortality PM2·5, sulphur dioxide, nitrogen Reduction in life expectancy
oxides, ozone Cancers
programme,6 which is the result of over 15 years of
Benzene benzo(a)pyrene Fatality risk from transport of materials and at research supported by the European Union (EU) and, to
1,3-butadiene, diesel particles workplace a lesser extent, the USA. ExternE is a bottom-up
Accident risk
approach of the kind described above, in which each
Morbidity PM10, sulphur dioxide, ozone Respiratory hospital admissions energy source is assessed individually and its ecological
PM2·5, ozone Restricted activity days
PM10, carbon monoxide Congestive heart failure and social footprint analysed. This approach is
Benzene, benzo(a)pyrene Cancer risk (non fatal) characterised by the so-called impact pathway, in which
1,3-butadiene, diesel particles Cerebrovascular hospital admissions emissions from a source are traced through as they
PM2·5 Cases of chronic bronchitis
Ozone Cough in asthma patients
disperse into the environment, after which the effects of
Accident risk Lower respiratory symptoms the dispersed pollutants are estimated. Finally, the
Asthma attacks health burden is valued in monetary terms where
Symptom days
possible. Figure 1 shows this pathway, and table 1
Myocardial infarction
Angina pectoris provides a description of the main effects estimated.
Hypertension Several points should be noted about the effects
Sleep disturbance assessed.
Risk of injuries from traffic and workplace accidents
Firstly, the emissions from a power source are dispersed
Data taken from ExternE.6 into the atmosphere in ways determined by the height of
the stack and by weather conditions—ie, temperature,
Table 1: Effect pathways included in analysis of the electricity sector
precipitation, and especially wind speed and direction.
980 www.thelancet.com Vol 370 September 15, 2007
Series
Deaths from accidents Air pollution-related effects
Among the public Occupational Deaths* Serious illness† Minor illness‡
Lignite30 0·02 (0·005–0·08) 0·10 (0·025–0·4) 32·6 (8·2–130) 298 (74·6–1193) 17 676 (4419–70 704)
Coal31 0·02 (0·005–0·08) 0·10 (0·025–0·4) 24·5 (6·1–98·0) 225 (56·2–899) 13 288 (3322–53 150)
Gas31 0·02 (0·005–0·08) 0·001 (0·0003–0·004) 2·8 (0·70–11·2) 30 (7·48–120) 703 (176–2813)
Oil31 0·03 (0·008–0·12) ·· 18·4 (4·6–73·6) 161 (40·4–645·6) 9551 (2388–38 204)
Biomass31 ·· ·· 4·63 (1·16–18·5) 43 (10·8–172·6) 2276 (569–9104)
Nuclear31,32 0·003 0·019 0·052 0·22 ··
Data are mean estimate (95% CI). *Includes acute and chronic effects. Chronic effect deaths are between 88% and 99% of total. For nuclear power, they include all
cancer-related deaths. †Includes respiratory and cerebrovascular hospital admissions, congestive heart failure, and chronic bronchitis. For nuclear power, they include all
non-fatal cancers and hereditary effects. ‡Includes restricted activity days, bronchodilator use cases, cough, and lower-respiratory symptom days in patients with asthma, and
chronic cough episodes. TWh=1012 Watt hours.
Table 2: Health effects of electricity generation in Europe by primary energy source (deaths/cases per TWh)
Chemistry also plays a part in determining the Because of the long-range dispersion of the pollutants,
composition and dispersion of the final product. This some effects can be felt more than 1000 km from the
dispersion can be simulated by use of complex models source. The following individual fuel cycles are worth
that take account not only of the local effects but also of noting.
the long-distance transport of the pollutants, through the
formation of particles as they are transformed into Coal and lignite
sulphates and nitrates. Long-distance effects are a The occupational health effects associated with mining
substantial proportion of total effects for air pollutants, are well known, although the rate of deaths and injuries
with the consequence that plants located away from has been declining. Nevertheless, studies have shown
centres of population can have health effects on people that up to 12% of coal miners develop one of several
living quite far away. potentially fatal diseases (pneumoconiosis, progressive
Secondly, the health burden is assessed not just for massive fibrosis, emphysema, chronic bronchitis, and
generation stage but also for the other stages of the full accelerated loss of lung function).33
cycle of the process, including the extraction of the fuel, At the generation stage the main effects arise from the
its transportation, transformation into electric energy, emissions of primary small particles (less than 2·5 µm or
disposal of the waste, and the transport of the electricity. PM2.5) and the creation of secondary small particles (less
So, for example, accidents in transportation are included. than 10 µm or PM10). Sulphur dioxide and nitrogen oxides
Thirdly, the estimates of air pollution effects are based emerge as important in this context because they
on extensive peer-reviewed epidemiological studies. Of contribute to the creation of secondary particles, in
particular importance are studies linking health effects to chemical oxidation involving atmospheric gases. Direct
concentrations of small particles and ozone health effects of sulphur dioxide and nitrogen oxides are
(webpanel 1).7–29 much less pronounced and are not included in the main See Online for webpanel 1
Fourthly, not all the effects can be valued in money estimates reported above.
terms, although the most important (ie, health) effects
have been. Although monetary valuation remains Oil and gas
controversial, especially when applied to health The health effects from gas are more than an order of
consequences such as premature mortality, methods magnitude lower than those from coal, mainly because
have been developed to make such valuations and the the effects from primary and secondary particles are
numbers are used in making decisions about investment much smaller. The technologies used in Europe and
in stricter pollution control standards through a assessed in our study are also state of the art and very
comparison of costs and benefits. In this paper, however, efficient, hence reducing emissions per unit of energy
we do not report on monetary values for health effects, generated. The health burdens associated with oil are
relying instead only on physical effects data. higher than those from gas but still much lower than for
Lastly, the scientific data on which the health effects are coal or lignite. Accidents from this fuel source are
based are not certain. This uncertainty can be seen in the estimated to be 50% higher than for gas but only 20% of
ranges of effects that are given. As new information those associated with coal and lignite.
becomes available, the values will also change and indeed
we have seen some changes in the estimates of health Biomass
effects over the past 15 years. Table 2 summarises the The biomass technologies addressed here refer to state of
main health effects that have been estimated for different the art plants that meet EU environmental standards (ie,
fuel cycles by the ExternE approach. almost all plants that were assessed for the data reported
www.thelancet.com Vol 370 September 15, 2007 981
Series
in table 2). Sources are mainly energy crops but also Safety procedures have also improved at the older reactors
some forest residues. The resulting impact, although that are not considered as safe (especially the light
substantial, is still well below that from coal and lignite. water-cooled graphite-moderated reactor of the kind that
As an indication, the resulting chronic mortality rates are was used in Chernobyl and of which 15 remain in
less than 20% of those from the lignite reference operation in Lithuania and Russia). Not all indicators
technology reported above. The most important show steady improvement. The number of unplanned
emissions are those of ozone precursors—such as automatic shutdowns (scrams) declined in the 1990s but
nitrogen oxides and volatile organic compounds. has remained stable since.34
There are also unresolved problems associated with the
Nuclear disposal of nuclear waste. The world’s 441 operating
The sources of the effects and indeed the effects themselves reactors generate more than 10 000 tonnes of heavy-metal
for the nuclear fuel cycle are very different from those for spent-fuel every year.34 A cumulative 190 000 tonnes are
the fossil fuel cycles. They can arise from occupational in storage and, although increased reprocessing will
effects (especially from mining), routine radiation during reduce the rate of growth of the stockpile, much will
generation, decommissioning, reprocessing, low-level remain and will need safe long-term storage. Finland,
waste disposal, high-level waste disposal, and accidents. Sweden, and the USA have made the most progress in
The data in table 2 show occupational deaths of around developing safe high-level waste repositories, although
0·019 per TWh, largely at the mining, milling, and none is expected to be operating before 2020. To date
generation stages. These numbers are small in the context there have been no serious incidents arising from the
of normal operations. For example, a normal reactor of high-level waste.
the kind in operation in France would produce 5·7 TWh a Despite these mostly positive developments and several
year. Hence, more than 10 years of operations would be attempts to bridge the gap, there remains a firm divide
needed before a single occupational death could be between lay and expert estimates of the probability of
See Online for webpanel 2 attributed to the plant. Likewise, numbers of deaths nuclear accidents (webpanel 2).36–38 There also remains
through cancer, severe hereditary effects, and non-fatal the long lead time, in many cases of around 10 years, for
cancers caused by normal operations are extremely small. approval and construction of nuclear power plants,
The main sources of potential damage are accidents though construction times can be cut to as little as 4 years
and non-routine radiation, and there is a lack of for some modular designs.39
agreement between expert assessments from the industry
and the public perception of these damages. The concerns Costs of CO2 and other greenhouse gases
about safety remain high, although the safety record has The calculations reported in table 2 do not include any
been improving steadily in most respects since 1990,34 contribution from global warming. The different forms
and the new generation of reactors are widely of power generation have very different contributions to
acknowledged to be much safer than earlier ones. After CO2 emissions (figure 2).
the events of Sept 11, 2001, designs have considered safety These contributions are the result of higher summer
against the impact of a fuel-laden passenger aircraft.35 temperatures (increased costs), warmer winter
temperatures (often decreased costs), flooding, increases
1500 in vector-borne diseases, and so on. The health costs of
Indirect emissions
increased greenhouse-gas emissions are difficult to
1290
1234 Direct emissions estimate and are controversial. A WHO study has
estimated that the increase in greenhouse gases
Equivalent CO2 emissions (g/kWh)
990 since 1990 has resulted in around 150 000 excess deaths
1000 960
in 2000. Almost all of these deaths took place in countries
that are not members of the Organisation for Economic
689 Cooperation and Development, where increased risk
factors for malnutrition, diarrhoea, malaria, floods, and
500 460 cardiovascular disease are attributed to climate change.41
410
The method for deriving these results involves taking
279 linear approximations on increases in concentrations of
116 greenhouse gases in 2025 and beyond; this approach
30 75
37 9 16 30 does not lend itself to estimation of deaths per tonne of
11
0 greenhouse gases and hence per TWh emitted
Coal Oil Gas Biomass Nuclear Hydro Wind Solar between 1990 and 2000. A better way to estimate the
Primary energy source health consequences of greenhouse gases in terms of
emissions is to do a baseline run of a model with
Figure 2 : Full energy chain CO2 equivalent emissions by primary energy source greenhouse-gas emissions, and then add a small amount
Data from IAEA, 2001.40 of greenhouse gases and see the additional effect. This
982 www.thelancet.com Vol 370 September 15, 2007
Series
method was used in a study by Bosello and colleagues.42
A B
Their results, although not directly comparable with
Cases of serious illness from air pollution (TWh)
Deaths from air pollution and accidents (TWh)
those of the WHO study, paint a somewhat different 40 300 Lignite
picture. Instead of excess deaths worldwide, the net
Lignite
results were of savings of about 840 000 by 2050. 30 Coal
However, both approaches do agree that developing 200
Coal
countries face an increase in mortality. Moreover, later in
Oil
the 21st century the number of deaths might increase 20
Oil
everywhere.
100
Although these specific consequences remain 10
controversial, figure 3 shows the correlation between the Biomass
Biomass Gas
direct health effects of power generation (as quantified in Gas
table 2) and the estimated contribution of the relevant 0 Nuclear 0 Nuclear
technology to greenhouse-gas emissions (equivalent CO2 0 500 1000 1500 0 500 1000 1500
emissions per kWh of production). Figure 3 shows that Equivalent CO2 emissions (g/kWh) Equivalent CO2 emissions (g/kWh)
the modes of generation that have the greatest immediate
Figure 3 : Health effects of electricity generation per TWh
effects are also those that make the strongest contribution (A) deaths from air pollution and accidents involving workers or the public; (B) cases of serious illness attributed to
to climate change. air pollution. Data for CO2 equivalent emissions from IAEA, 2001.40
Putting the health costs of electricity generation in perspective through accelerated ageing, and a shrinking of the
The ExternE results can be viewed from two comparative probability of survival curve across the population. The
perspectives: in terms of the total health burdens caused consequence can be reported as an increase in the
by electricity generation, and relative to other sources of observed death rate, or as a change in life expectancy.
health burden, such as smoking. The total health burden Calculations made by the ExternE team indicate that the
of electricity generation will depend on which fuels are current concentrations of PM2·5 in the EU and USA of
used and the total amount generated. In the UK, for around 20 μg/m-³ result in a loss of life expectancy of
example, total generation was around 386 TWh, of which around 8 months. A UK review suggests a much smaller
34% came from coal and lignite, 37% from gas, 23% from figure for the loss of life expectancy, or between 1 month
nuclear sources, 2% from oil, and the remaining 4% from and 1 year. This amount, however, includes only the acute
hydro, wind, biomass, and other fuels.43 effects of exposure to particulate matter and is certainly
Taking the figures for health burdens by fuel type in an underestimate of the total effect (chronic and acute).45
table 2 we get the overall burden from electricity as given We should also note that the gains in air quality from
in table 3. The data indicate that about six accidental halving concentrations of particles will require reductions
deaths and 13 occupational deaths can be attributed to in emissions from not only power stations but also
the generation of electricity in the UK per year. Also, transport and other sources.46
around 3800 deaths arise from the associated air A reasonable policy goal would be to reduce the life
pollution. There are around 35 000 cases of serious illness expectancy deficit of around 8 months by half, which
a year and 1·9 million cases of minor illnesses, as defined would increase life expectancy by 4 months. By compari-
in table 2. These findings can be put in perspective by son, regular smoking is judged to cause a loss of life
comparison with general mortality and morbidity data. expectancy of 5–8 years, or about ten times the effects
There were roughly 260 000 deaths in England and Wales of PM2·5 air pollution and 20 times the effects of a
from respiratory and circulatory diseases in 2001, so, the plausible reduction in air pollution. We should also
estimates from electricity generation account for about note that the gains from reductions in air pollution
0·014% of the total. In terms of morbidity, there were apply to a much larger population than those from
around 667 000 episodes leading to hospital admission cessation of smoking.
for respiratory and cerebrovascular diseases in England
in 2001.44 Over the same period the estimates of serious
Cases Percentage due to coal
illnesses in these two categories arising from electricity
generation in the UK amounted to about 840 (table 2). Accident-related deaths
Although the electricity data are for the UK emissions all Among the public 6 44%
over Europe and hospital admissions are for England Occupational 13 99%
alone, one can see that electricity generation accounts for Air pollution
a very small part of the total admissions in these Deaths 3778 85%
categories. Serious illness 35 186 84%
Another useful point of comparison is with the health Minor illness 1 853 152 94%
effects of smoking. In terms of excess deaths per year,
Table 3: Health burdens from electricity generation in the UK, 2001
the process by which air pollution affects mortality is
www.thelancet.com Vol 370 September 15, 2007 983
Series
Developing countries
Indoor air Lignite-based
As countries become more industrialised their use of pollution electric energy
electricity and petroleum products for transportation
Mortality (cases per year) 1962* 33†
increases, which in turn creates new environmental
Morbidity (cases per year) 502 000* 18 000‡
health problems, largely in the form of respiratory
diseases, cardiac diseases, and cancers. Although most Details of calculations available from the author. *Caused by acute respiratory
estimates of such effects of commercial energy have been infection in children and chronic obstructive pulmonary disease in women.
†Caused by occupational and public accidents in the mining and transportation of
made for developed countries, some studies are available fuel and in the generation of the electricity, and from respiratory and
for countries such as China, India, and Brazil. The large cardiovascular deaths associated with the emissions from the generation.
effects are due in some cases to lack of adequate emission ‡Respiratory hospital admissions, cerebrovascular hospital admissions, restricted
activity days, bronchodilator usage (in asthmatic adults and children), cough (in
regulations but also in many countries to ineffective
asthmatic adults and children), lower respiratory symptoms (in asthmatic adults
enforcement of existing regulations. In India, for and children), chronic bronchitis in adults and children and chronic cough in
example, concentrations of suspended particulate matter adults and children.
(roughly equivalent to total suspended particulate matter)
Table 4: Health costs of indoor air pollution against that from electric
and respirable suspended particulate matter (roughly power (per TWh of generation)
equivalent to PM10) are frequently well in excess of
national standards. In 2003, levels of suspended
particulate matter exceeded national standards in 77 of colleagues,51 whereas the health consequences of
the 91 residential monitoring stations more than 25 times electricity generated from a state-of-the-art coal plant in
and the standard respirable suspended particulate matter India are taken from ExternE (table 2). The results refer
of 120 μg/m3 (yearly average) was exceeded in most to a plant producing 1 TWh (10⁹ kW) of electricity in
cities.47 Such levels are associated with substantial health 1 year (a plant of about 150 MW would generate such an
effects in other countries, and emerging research in amount of electricity per year). This amount could
India indicates that effects are similar there.48 Thus, this provide enough electricity for basic lighting, cooking,
failure of the environmental regulations to work and other needs for 333 000 households or about
effectively is having important consequences for human 1·6 million individuals, on the basis of a household size
wellbeing. A study by the Institute of Economic Growth of five people.
in New Delhi,49 for example, has estimated yearly damage Table 4 shows that the indoor fuels cause in the region
from urban air pollution in 15 major cities in India to be of 1962 premature deaths and half a million cases of
111 billion rupees (US$2·5 billion). Much of this cost, acute respiratory illness and chronic obstructive
however, is attributable to sources such as transport and pulmonary disease in a year. If these fuels could be
industry. In some states these costs are as high as 8% of replaced by electricity then the health burden would be
the state domestic product. somewhere in the region of 33 premature deaths and
Studies in China have also revealed important health 18 000 cases of illness, ranging from severe (hospital
effects from the operations of coal-fired power stations. A admissions for respiratory failure) to mild (a cough day).
study of such plants in Shandong province estimated Even with allowance for the uncertainties that exist in
around 77 deaths per TWh from a normal coal-fired plant these estimates, the difference between the two options,
that met Chinese environmental standards.50 This which is more than an order of magnitude, makes the
estimate is much higher than that for European plants, case for a shift to cleaner commercial fuels unanswerable
indicating both a lower population density in Europe as in health terms.
well as the use of cleaner technology. Estimates of effects
on serious morbidity (respiratory and cerebrovascular Role of CO2 capture and storage
hospital admissions, congestive heart failure, and cases Because of the major contribution of fossil-fuel use in
of chronic bronchitis) are estimated at 975 per TWh electricity generation to global emissions of CO2, there
compared with around 225 per TWh in Europe.50 has been interest in the potential of CO2 capture and
Although the health consequences of commercial fuel storage to mitigate climate change. In this process, CO2
in developing countries are beginning to be felt, and from large point sources such as power plants is
studies show that the benefits of adopting cleaner captured and stored in isolation from the atmosphere.
technologies to reduce emissions from power generation The technology of capture is already commercially
and transport are almost invariably justified, we should available for large CO2 emitters, such as power plants,
recognise that even the dirtiest commercial fuels are less but the long-term storage of CO2 is mostly untested.52 In
damaging in health terms than the traditional fuels they theory, however, the capture and storage of a high
could potentially replace. A comparison is provided in proportion of CO2 emissions from large point sources of
table 4 of replacement of traditional fuels in the home fossil-fuel combustion, such as power stations, is
with electricity generated from coal. The estimates of the possible.
health costs of indoor air pollution are based on There are three main methods of capture.
epidemiological studies as summarised in Desai and Post-combustion capture, suitable for a modern
984 www.thelancet.com Vol 370 September 15, 2007
Series
pulverised coal power plant or a natural gas combined
Pulverised coal Integrated gasification Natural gas
cycle plant, entails separation of CO2 from the flue gases combined cycle system combined cycle plant
derived from the combustion of the primary fuel in air.
Rate Increase Rate Increase Rate Increase
Typically, an organic solvent such as monoethanolamine
is used to capture the CO2, which is present at Resource consumption
concentrations of 3–15% in flue gases. In pre-combustion Fuel 390 93 361 49 156 23
capture, the primary fuel is mixed with steam and air or Limestone 27·5 6·8 .. .. .. ..
oxygen to generate a synthesis gas high in carbon Ammonia 0·80 0·19 .. .. .. ..
monoxide and hydrogen, which is further treated with CO2 capture and storage 2·76 2·76 0·005 0·005 0·80 0·80
reagents
steam to yield more hydrogen and CO2 at concentrations
Solid wastes or by-products
of around 15–60% by dry volume. This method is suitable
Ash or slag 28·1 6·7 34·2 4·7 .. ..
for power plants based on integrated gasification
combined cycle technology. The third system (oxyfuel) FGD residues 49·6 12·2 .. .. .. ..
uses oxygen at 90–95% purity instead of air for Sulphur .. .. 7·53 1·04 .. ..
combustion to yield a flue gas high in CO2 and water. The Spent CO2 capture and 4·05 4·05 0·005 0·005 0·94 0·94
storage sorbent
resulting high concentration stream of CO2 is compressed
Atmospheric emissions
and dried for transportation by pipeline or tanker for
Carbon dioxide 0·07 –704 97 –720 43 –342
storage in geological formations, in deep ocean waters,
Sulphur oxides 0·001 –0·29 0·33 0·05 .. ..
or by mineral carbonation (figure 3).
Nitrogen oxides 0·77 0·18 0·10 0·01 0·11 0·02
Ammonia 0·23 0·22 .. .. 0·002 0·002
Environment and health risks and benefits
Because CO2 capture and storage is still a fairly new Values shown are rates in kg per megawatt hour for the capture plant, plus increases over the reference plant rates for
technology, our understanding of its health implications the same plant type. FGD=flue gas desulphurisation. Data from IPCC.52
is incomplete. An assessment needs to be based on a Table 5: Illustrative consequences of CO2 capture and storage energy requirements on plant-level
life-cycle approach that takes into account the extraction resource consumption and non-CO2 emission rates for three current power plant systems
and processing of the primary fuel, the generation and
distribution of electricity, and the handling and storage of
waste products. The benefits include the reduction of CO2 in solid waste products. Sulphur dioxide emissions are
emissions, which are estimated to be around 85–95% lower lower for pulverised coal, but higher for integrated
than with similar technology and no CO2 capture and gasification combined cycle technology, whereas
storage. The concentrations in the flue gas of other emissions of nitrogen oxides are higher for all three, and
substances harmful to human health are likely to be emissions of ammonia are increased in pulverised coal
similar to, or lower than, those of plants without capture and natural gas combined cycle plants. The effect of these
and storage technology. This is because the capture emission changes on human health has not been
process entails the removal of some emissions or the systematically estimated, but it would be reasonable to
upstream removal of impurities, such as sulphur assume proportionate increases in occupational and
compounds, and is required for the efficient operation of other risks associated with the increased consumption of
the technology. However, plants with such technology the primary fuel and other resources, net adverse effects
operate at lower efficiency (have higher energy relating to solid wastes, and mixed effects relating to the
requirements) than similar plants without it. The increase changes in atmospheric emissions.
in fuel consumption per kWh associated is in the Additionally, there are health and safety issues relating
range 10–40%.52 As a result, there is a need to process to the transport of concentrated CO2 by pipeline or
more of the primary fuel. Thus, even where the tanker (eg, a small risk of asphyxia with local build up of
concentration of an impurity in the flue gases is reduced, CO2, toxicity from hydrogen sulphide; figure 4), although
the overall emissions per kWh could still be higher. such risks probably differ little from those associated
A study by Rubin and colleagues,53 reported also in with hydrocarbon pipelines already in operation. Leaks
the 2005 IPCC Special Report on Carbon Dioxide Capture of CO2 after its storage in geological formations probably
and Storage, is one of the few published assessments of carries a low risk to human health comparable to that
the resource and emission consequences of common associated with current activities such as natural gas
fossil-fuel power plants (pulverised coal, integrated storage and enhanced oil recovery. The effects of the
gasification combine cycle, and natural gas combined injection of CO2 into deep ocean waters on ecosystems
cycle technology) using CO2 capture (table 5). With all remain uncertain.
three plant types, the increase in primary fuel use and Thus, whatever the potential contribution of CO2
the need for capture and storage reagents is clear. With capture and storage to mitigation of climate change, and
pulverised coal, additional amounts of limestone (for whatever the economic case (which is closely tied to the
control of sulphur dioxide) and ammonia (for control of price of the primary fuels), the effects of this technology
nitrogen oxides) is required. All three also have increases on health seem mixed. Its use does not fundamentally
www.thelancet.com Vol 370 September 15, 2007 985
Series
density—ie, the power production per square metre of
Changes to atmopheric emissions Mine or primary fuel
↓ CO2 (all) land area is low, which places constraints on large-scale
↓ SO2 (pulverised coal) production; intermittency, which means that methods
↑ NOx (all)
↑ SOx (intergrated gasification
have to be found to store their energy or to supplement it
combined cycle) by more controllable forms of power generation to
↑ NH3 (pulverised coal and NGCC) manage variations in production and demand; constraints
on their location, which is generally governed by
geological, hydrological, meteorological, and other
Power station or Solid waste factors, and which might therefore require long-distance
other facility ↑ Ash/slag (pulverised coal and transmission from the site of generation to the place of
intergrated gasification
Resource use combined cycle use; and environmental effects, aesthetic effects, or both,
↑ Primary fuel consumption ↑ FGD residues (pulverised coal) that might in part offset the broader environmental and
↑ Limestone (pulversied coal) ↑ Sulphur (intergrated gasification
↑NH3 (pulverised coal) combined cycle)
health gains derived from lower air pollution and
↑ CCS reagents (all) ↑ Spent CCS sorbent (all) greenhouse-gas emissions.
With the exception of biomass, most renewable power
systems do not rely on combustion and thus do not
Temporary CO2 storage produce notable amounts of air pollution directly. However,
Risks relating to some emissions of air pollutants can arise during
the transport manufacture and construction, such as in the production
of CO2
3 Mineral carbonation CO2 of steel for wind turbines and concrete for dams, but these
H2S are low compared with any but the cleanest system relying
Marine platform 2 Ocean storage on combustion for electricity generation.
Fixed pipeline
1 Hydroelectric
There are some 48 000 large dams in operation
Storage in worldwide, contributing to provision of drinking water,
geological CO2 escape
irrigation, flood control, and 20% of the world’s
Possible effect on formations electricity. Although apparently a clean form of electric
marine ecosystems
power, hydroelectricity is controversial because of its
social, health, and environmental costs (webpanel 3).54–59
Storage Concerns about these issues have caused a re-appraisal
formation
Dissolution Risks relating to escape of CO2 of role of hydroelectricity, and a much more
· CO2 escape to global atmosphere
· Asphyxia from CO2 pooling
cautious policy towards its further development54—
· Contamination of acquifers notwithstanding the current ambitious programmes in
Liquid CO2 lake with CO2 brine or other China, India, and elsewhere.
Figure 4 : Summary of the principal CO2 capture and storage processes and associated sources of environment
Despite these concerns, hydropower from large dams
and health risks has several important advantages. Among the
NGCC=natural gas combined cycle. FGD=flue gas desulphurisation. CCS=CO2 capture and storage. Adapted from non-fossil-fuel forms of power generation, it provides a
figures in the IPCC special report on Carbon Dioxide Capture and Storage.52 comparatively constant source and store of energy, which,
with large reservoir heads, can be very rapidly mobilised
See Online for webpanel 3 alter the sources of adverse health effects associated with to meet surges in demand. Also relevant is that the
fossil fuels, and it could in some cases increase immediate untapped potential for hydroelectric development is
and near-term consequences because of efficiency losses greatest in regions (Asia, sub-Saharan Africa, South
and need for additional resource use. From an America) where many of the 2 billion people currently
environment and health perspective, therefore, CO2 without access to electricity live. In these regions,
capture and storage is at best only a partial solution. hydroelectricity could have an important role in future
energy provision, provided projects follow good practice
Health effects of renewable energy guidance.54 However, the unexploited capacity is limited
In view of the evident health and environmental costs of and they are associated with appreciable emission of CO2
conventional fossil-fuel combustion, the modern debate and methane from anaerobic fermentation in the static
has appropriately turned towards other energy sources: water (webpanel 3).
the so-called renewables, derived directly or indirectly Among the range of social and health effects are the
from the energy of sunlight (direct solar, hydroelectricity, occupational risks during construction, and the low but
wind, wave, biofuel production, and surface heat), the finite chance of dam failure. The probability of such
gravitational pull of the moon (tidal), or the radioactivity events is extremely difficult to establish, since they can
of the Earth’s interior (geothermal). In varying degrees be triggered by earthquakes, wars, terrorist activity, or
these sources share four main drawbacks: low energy engineering failures.
986 www.thelancet.com Vol 370 September 15, 2007
Series
Solar rural populations in lower-income countries. Furthermore
The theoretical potential of the direct capture of solar if fuel crops are used in state of the art power-generating
energy either through photovoltaic systems or by heat plants that meet the latest EU environmental standards
generation is enormous. After allowance for energy the health consequences of bioelectricity production,
reflected by the atmosphere, around 3·9×10²⁴ J are although substantial, are still well below those from coal
incident on the Earth’s surface per year—almost and lignite.
10 000 times more than current global energy
consumption. Thus, the capture of less than 1% of Wind, wave, and geothermal
photonic energy would serve all human energy needs. Wind energy, mainly produced by horizontal-axis turbines
The limited assessment available by a full cycle analysis38 of varying sizes, is one of the more cost-effective forms of
indicates few drawbacks, though possible concerns might renewable energy with today’s technology. As with solar
arise in relation to the production, handling, and disposal power, the obvious variability of its generation raises
of the photovoltaic materials, and if battery or other questions about solutions for energy storage and so-called
technology is needed for energy storage. despatchable capacity in other parts of the electricity grid.
The constraint on much wider use is primarily With connections across a wide network, the natural
technical. With photovoltaic systems (which depend on geographical variation in wind speed could help to
quantum excitation of electrons in layered smooth out fluctuations, but this might not entirely avoid
semiconductors), the efficiency of solar capture—the the need for additional generation capacity elsewhere.
ratio of power output to the power of the incident Similar considerations obtain in relation to wave power.
radiation—is limited by, among other factors, the fact However, the balance of health risks and benefits, though
photovoltaic cells capture energy from across only a imprecisely defined, would seem strongly favourable, as
limited range of the solar electromagnetic spectrum. The it does for geothermal energy. The latter is an option only
best overall efficiency with current technology is in selected locations worldwide, though ground-source
around 10–15%. Given the time-averaged rate of incident heat pumps using surface pipes or bore holes are local
solar energy of 100–300 W, this means little more than options for space heating and offer good energy return
10–45 Wm¯² as a global average. Although thermal (around 3 W back for every 1 W of energy expenditure).
systems might be more cost-effective than photovoltaic Although geothermal generation has some local air
systems, they too capture only a fraction of the incident pollution associated with it, the effects are much smaller
energy. Thus, solar systems suffer problems of cost, large than for fossil-fuel sources and it can be considered a
requirements for land area, and intermittency. much cleaner source.41
Nonetheless, from a health perspective, the potential
benefits of direct solar capture seem very desirable. Nuclear fusion
A nuclear technology that offers some hope for future
Biomass electricity generation is fusion.62 However, its commercial
Fresh (as opposed to fossilised) biomass is a potentially development is still some way off because of major
large store of renewable energy, which can be transformed technical challenges (webpanel 4). If these issues can be See Online for webpanel 4
into useful power by combustion or by thermochemical or overcome, nuclear power offers comparatively clean
biochemical conversion to liquid (ethanol, methanol) or electricity production with little contribution to
gaseous fuels (methane, hydrogen).60 However, its greenhouse-gas emissions. It shares the range of
usefulness as a major energy source is limited by the environmental and health risks of fission technology but
inherent inefficiency of photosynthesis, which captures at generally lower levels. The main product of normal
no more than a small percentage of solar energy reaching operation is helium-4 (webpanel 4), which is an inert gas,
the Earth’s surface.61 The energy yield of even the most but the fusion reaction also requires a radioactive isotope
productive cultivated crops is therefore little higher than of hydrogen, tritium, which is difficult to capture
1 Wm¯²; the imperfect efficiency of the energy conversion completely, so some leakage is inevitable. The short
means that the power density is less than 1 Wm¯²—an half-life of just 12 years will help limit the build-up in the
order of magnitude lower than direct solar capture through environment, but there could be important effects on
photovoltaic or thermal systems, and up to four orders of health in an economy with a substantial number of
magnitude lower than fossil-fuel combustion. To substitute fusion plants. The high-energy neutrons produced in a
for even a modest fraction of current or future coal use, reactor make the structural materials surrounding the
for example, would require substantial land area to be fusion chamber radioactive, with a similar inventory of
given over to fuel crops—often in competition with food radioactive materials to a fission reactor. The half-lives of
production. However, some high-yielding crops, for the radioisotopes produced by fusion are substantially
example South American sugar cane, are already being less than those for fission, they tend to be less biologically
used successfully as fuel sources, though mainly for active, and there is potential to use low activation
transport. And bioelectricity could have an important materials that do not easily become radioactive with
function in supporting electricity needs particularly of neutron bombardment. Most materials of the core would
www.thelancet.com Vol 370 September 15, 2007 987
Series
be radioactive for around 50 years, and other low-level The study also reports estimates of the effect of nuclear
waste for another 100 years or so. Thus the difficulties of generation. The role nuclear power should have in future
handling and storing radioactive waste would be smaller energy production depends on a balance of (perceived)
than for fission. risks. It currently accounts for around 17% of global
The risk of major accident is also substantially less. electricity production and makes a small contribution to
Fusion is not a chain reaction and requires very reducing greenhouse gases; thus, a decision not to
demanding control conditions (extremes of temperature replace current nuclear capacity would correspondingly
and pressure, and magnetic containment), which, if increase the challenge of limiting greenhouse-gas
disrupted, would rapidly halt the reaction. Moreover, emissions. Such a decision would be welcome in health
unlike a fission reactor, the fusion chamber contains very terms if the nuclear plants were replaced by capacity in
little fuel—enough only to perpetuate the reaction for a renewable production additional to the level of renewable
minute or so—and stopping the supply would result in production that would otherwise occur.
rapid shut-down. Fusion also has much less overlap with Forms of renewable energy generation are still in the
weapons technology. Plutonium (needed for atomic early phases of their technological development, but
bombs) can be bred by use of the neutrons from a fusion most seem to be associated with few adverse effects on
reactor, but only with extensive redesign of the reactor— health and to contribute little to the longer-term
which would therefore be easy to monitor. environmental threat of climate change. Their rapid
Although nuclear power has promise as a future energy expansion is partly constrained by the intermittency and
technology, more than half a century of research with low density of energy production.
several experimental reactors has failed to produce net Much work is under way to improve the technologies
energy output in controlled production. The best used in electricity generation, and policymakers have
estimates are that perhaps another 50 years will pass been raising standards at all stages of the fuel cycle. The
before the technology is developed to the point of case for raising environmental standards is made on
commercially viability, which will be too late to make a grounds of cost benefit, which requires the health impact
significant contribution to mitigation of climate change to be valued in monetary terms. Although such values
over this century. Nuclear energy is not therefore a are not reported in this paper, regulatory impact studies
solution to climate change, but it remains an attractive have almost always confirmed that the benefits (which
hope for electricity production for future generations. are predominantly health related) of the higher standards
exceed the costs, in both the quickly industrialising and
Conclusion the more industrialised countries.
The generation of electricity has both health benefits and Although these are complex and rapidly evolving
costs. The health benefits of a shift away from issues, the key messages from a health perspective are
non-commercial fuels to commercial ones, particularly clear. Population health will substantially benefit from
when they are used for electricity generation, are evident improved access to electricity and from modal switch
from the evidence in developed countries in the past away from fossil fuels towards renewable sources of
century and which is still taking place in developing electricity generation where possible. The case for such
countries. Moreover, the substitution of dirty energy for switching cannot be judged purely on traditional
clean is not the only change that increases wellbeing. cost-effectiveness comparisons of current technology,
Efficient lighting, refrigeration, clothes washers, radios since investment in renewable sources and increases in
and TVs, computers, and numerous appliances that use volume of production should bring cost efficiencies to
electricity make possible those activities that otherwise newer (often the renewable) technologies; moreover, the
would not be possible. cost-benefit equation is more favourable to renewable
Although there are health costs associated with the technologies where proper account can be taken of
generation of electricity, especially from fossil fuels, they environmental and health effects. In addition to
are much smaller than those associated with indoor air increasing access to electricity (see the first paper in this
pollution from burning fuels directly in homes. The Series), our progress towards those strategic goals can be
See Online for webpanel 5 drawbacks lie mainly with their contribution to outdoor measured (webpanel 5).
air pollution, occupational risks, and greenhouse gas Acknowledgments
emissions. We thank the seminar participants at the London School of Hygiene
Assessment of the health effects of electricity and Tropical Medicine for useful comments and in particular
Kirk R Smith for detailed review of the paper. We acknowledge the work
generation should include all stages of the fuel cycle, of the whole ExternE team, whose results are reported here. Special
such as mining, transportation, and disposal of waste. mention should be made of the contributions of Ari Rabl,
Studies in Europe, based on the ExternE methods, have Fintan Hurley, and Rainer Fredrich who have worked on these issues
provided estimates of the effect, in terms of excess deaths for many years. All errors and omissions are of course our own
responsibility.
and various categories of morbidity.30–32 The effects are
not unimportant, especially from the use of coal and Conflict of interest statement
We declare that we have no conflict of interest.
lignite.
988 www.thelancet.com Vol 370 September 15, 2007
Series
References 24 Bell ML, Peng RD, Dominici F. The exposure-response curve for
1 Maddison A. The world economy. Historical perspectives. Paris: ozone and risk of mortality and the adequacy of current ozone
OECD, 2003. regulations. Environ Health Perspect 2006; 114: 532–36.
2 Deaton A. Health in an age of globalization. Research Program in 25 Ito K, De Leon SF, Lippmann M. Associations between ozone and
Development Studies. Centre for Health and Well-being. Princeton, daily mortality: analysis and meta-analysis. Epidemiology 2005; 16:
NJ: Princeton University, 2004. 446–57.
3 Woods R. The demography of Victorian England and Wales. 26 Levy JI, Chemerynski SM, Sarnat JA. Ozone exposure and
Cambridge: Cambridge University Press; 2000. mortality: an empiric bayes metaregression analysis. Epidemiology
4 Wang L. Health outcomes in poor countries and policy options: 2005; 16: 458–68.
empirical findings from demographic and health surveys. WPS 27 Ostro BD, Rothschild S. Air pollution and acute respiratory
2831. Washington DC: The World Bank, 2001. morbidity: an observational study of multiple pollutants.
5 Grubler A, Jefferson M, Nakicenovic N. Global energy perspectives: Environ Res 1989; 50: 238–47.
a summary of the joint study by the International Institute for 28 Just J, Segala C, Sahraoui F, Priol G, Grimfeld A, Neukirch F.
Applied Systems Analysis and World Energy Council. Short-term health effects of particulate and photochemical air
Technological Forecasting and Social Change 1996; 51: 237–64. pollution in asthmatic children. Eur Respir J 2002; 20: 899–906.
6 European Commission. ExternE: Externalities of Energy: 29 Hiltermann TJ, Stolk J, van der Zee SC, et al. Asthma severity and
Methodology 2005 Update. EUR 21951. Brussels: European susceptibility to air pollution. Eur Respir J 1998; 11: 686–93.
Commission, 2005. 30 ExternE. ExternE National Implementation; Germany. http://
7 Pope CA 3rd, Burnett RT, Thun MJ, et al. Lung cancer, externe.jrc.es/ger.pdf (accessed Aug 28, 2007).
cardiopulmonary mortality, and long-term exposure to fine 31 Berry JE, Holland MR, Watkiss PR, Boyd R, Stephenson W. Power
particulate air pollution. JAMA 2002; 287: 1132–41. generation and the environment—a UK perspective, vol 1.
8 Dockery DW, Pope CA 3rd, Xu X, et al. An association between air http://externe.jrc.es/uk.pdf (accessed Aug 28, 2007).
pollution and mortality in six US cities. N Engl J Med 1993; 329: 32 Dreicer M, Tort V. ExternE—Externalities of Energy: nuclear.
1753–59. European Commission report EUR 16524, Vol 5. Brussels: EC,
9 Miller KA, Siscovick DS, Sheppard L, et al. Long-term exposure to 1995.
air pollution and incidence of cardiovascular events in women. 33 Stephens C, Ahern M. Worker and community health impacts
N Engl J Med 2007; 356: 447–58. related to mining operations internationally. A rapid review of the
10 Filleul L, Rondeau V, Vandentorren S, et al. Twenty five year literature. London: Mining and Minerals for Sustainable
mortality and air pollution: results from the French PAARC survey. Development Project, 2001.
Occup Environ Med 2005; 62: 453–60. 34 International Atomic Energy Agency. Nuclear technology review.
11 Krewski D, Burnett RT, Goldberg M, et al. Reanalysis of the Harvard Vienna: IAEA, 2006.
Six Cities Study, part I: validation and replication. Inhal Toxicol 35 Herbst AM, Hopley GW. Nuclear energy now: why the time has
2005; 17: 335–42. come for the world’s most misunderstood energy source. Hoboken,
12 Gehring U, Heinrich J, Kramer U, et al. Long-term exposure to NJ: John Wiley & Sons Inc, 2006.
ambient air pollution and cardiopulmonary mortality in women. 36 Goldemberg J, Johansson TB, eds. World Energy Assessment:
Epidemiology 2006; 17: 545–51. overview 2004 update. New York: United Nations Development
13 Hoek G, Brunekreef B, Goldbohm S, Fischer P, van den Brandt PA. Programme, 2005.
Association between mortality and indicators of traffic-related air 37 Greenpeace. The Chernobyl catastrophe. Consequences on human
pollution in the Netherlands: a cohort study. Lancet 2002; 360: health. Amsterdam, the Netherlands: Greenpeace, 2006.
1203–09. 38 Poortinga W, Pidgeon N, Lorenzoni I. Public perceptions of nuclear
14 Naess O, Nafstad P, Aamodt G, Claussen B, Rosland P. Relation power, climate change and energy options in Britain: summary
between concentration of air pollution and cause-specific mortality: findings of a survey conducted during October and November 2005.
four-year exposures to nitrogen dioxide and particulate matter Understanding Risk Working Paper 06-02. Norwich: Centre for
pollutants in 470 neighborhoods in Oslo, Norway. Am J Epidemiol Environmental Risk, Tyndall Centre for Climate Change Research,
2007; 165: 435–43. University of East Anglia, 2006.
15 Woodruff TJ, Grillo J, Schoendorf KC. The relationship between 39 Hore-Lacy I. Nuclear energy in the 21st century. London and
selected causes of postneonatal infant mortality and particulate air Burlington, MA: World Nucelar University Press, 2006.
pollution in the United States. Environ Health Perspect 1997; 105: 40 International Atomic Energy Agency. Sustainable development and
608–12. nuclear power. Vienna: IAEA, 2001.
16 Sram RJ, Binkova B, Dejmek J, Bobak M. Ambient air pollution and 41 McMichael AJ, Campbell-Lendrum D, Kovats RS. Global Climate
pregnancy outcomes: a review of the literature. Change. In: Ezzatti M, Lopez AD, Rodgers A, Murray CJ, eds.
Environ Health Perspect 2005; 113: 375–82. Comparative quantification of health risks: global and regional
17 Glinianaia SV, Rankin J, Bell R, Pless-Mulloli T, Howel D. Does burden of disease due to selected major risk factors, vol 2. Geneva:
particulate air pollution contribute to infant death? A systematic WHO, 2004: 1543–649.
review. Environ Health Perspect 2004; 112: 1365–71. 42 Bosello F, Roson R, Tol R. Economy wide estimates of the
18 APHEIS-3. Air Pollution and Health: a European Information implications of climate change: human health. Ecological Economics
System health impact assessment of air pollution and 2006; 58: 579–91.
communication strategy. Third year report, 2005. http://aphis.net 43 Eurostat. http://epp.eurostat.ec.europa.eu (accessed Aug 28, 2007).
(accessed Aug 28, 2007).
44 HESOnline. Primary diagnosis: summary. http://www.hesonline.
19 Hurley F, Hunt A, Cowie H, et al. Methodology for the cost benefit nhs.uk/Ease/servlet/ContentServer?siteID=1937&categoryID=202.
analysis for CAFE. Vol 2: health impact assessment. Didcot, Oxon,
45 UK Department of Health. Economic appraisal of the health effects
UK: AEA Technology, 2005.
of air pollution. London: Stationery Office, 1999.
20 Ostro B. Air pollution and morbidity revisited: a specification test.
46 Rabl A. Interpretation of air pollution mortality: number of
J Environ Econ Manage 1987; 14: 87–98.
deaths or years of life lost? J Air Waste Manag Assoc, 2003; 53:
21 Anderson H, Atkinson R, Peacock J, Marston L, Konstantinou K. 41–50.
Meta-analysis of time series studies and panel studies of particulate
47 TERI. TERI energy data director and yearbook. New Delhi, India:
matter (PM) and ozone (O3). Report of a WHO Task Group.
TERI Press, 2006.
Copenhagen: World Health Organization, 2004.
48 Markandya A. Power system planning in India: incorporating
22 Variations in the prevalence of respiratory symptoms, self-reported
environmental externality costs and benefits. World Bank South
asthma attacks, and use of asthma medication in the European
Asia Department working paper. Washington DC: World Bank,
Community Respiratory Health Survey (ECRHS). Eur Respir J 1996;
2007.
9: 687–95.
49 Murty M, Gulati SC, Banerjee A. Health benefits from urban air
23 Ward DJ, Ayres JG. Particulate air pollution and panel studies in
pollution abatement in the Indian subcontinent. New Delhi, India:
children: a systematic review. Occup Environ Med 2004; 61: e13.
Institute of Economic Growth, 2003.
www.thelancet.com Vol 370 September 15, 2007 989
Series
50 Eliason B, Lee Y, eds. Integrated assessment of sustainable energy 56 Keiser J, De Castro MC, Maltese MF, et al. Effect of irrigation and
systems in China. Dordrecht, Netherlands: Kluwer Academic large dams on the burden of malaria on a global and regional scale.
Publishers, 2003. Am J Trop Med Hyg 2005; 72: 392–406.
51 Desai M, Mehta S, Smith K. Indoor smoke from solid fuels. 57 Ghebreyesus TA, Haile M, Witten KH, et al. Incidence of malaria
Assessing the environmental burden of disease at national and local among children living near dams in northern Ethiopia: community
levels. Environmental Burden of Disease Series, No 4. Geneva: based incidence survey. BMJ 1999; 319: 663–66.
World Health Organization, 2004 58 Jobin W. Dams and disease. London: Routledge, 1999.
52 IPCC. Carbon dioxide capture and storage. IPCC special report 59 World Wild Fund For Nature. To dam or not to dam. Five years on
prepared by Working Group III of the Intergovernmental Panel from the World Commission on Dams. Zeist, Netherlands: WWF
in Climate Change. New York: Cambridge University Press, Global Freshwater Programme, 2005.
2005. 60 Sorensen B. Renewable energy, 3rd edn. Burlington, MA: Elsevier
53 Rubin E, Rao A, Chen C. Comparative assessments of fossil fuel Academic Press, 2004.
power plants with CO2 capture and storage. In: Rubin E, Keith D, 61 Miyamoto K, ed. Renewable biological systems for alternative
Gilboy C, eds. Proceedings of 7th International Conference on sustainable energy production (FAO Agricultural Services
Greenhouse Gas Control Technologies, 2005. Oxford, UK: Elsevier Bulletin—128). Rome: Food and Agriculture Organization of the
Science 2005: 285–94. United Nations, 1997.
54 World Commission on Dams. Dams and development. A new 62 EURATOM/UKAEA Fusion Association. Fusion power. Energy for
framework. The report of the World Commission on Dams. London the future. Culham: UK AEA, 2007.
and Sterling, VA: Earthscan Publications Ltd, 2000.
55 Smil V. Energy at the crossroads. Global perspectives and
uncertainties. Cambridge, Massachusetts: MIT Press, 2005: 317–49.
990 www.thelancet.com Vol 370 September 15, 2007
Anil Markandya