Lost life expectancy due to air pollution in China

Douglas Dockery, C. Arden Pope III, 16 Jan 2014

Recent smogs in Beijing and other Chinese cities have registered almost off the scale in terms of particulate pollution. We are currently unable to measure all health costs or full losses of life due to this air pollution with absolute accuracy or certainty. Nonetheless, data suggests that poor air quality imposes a significant health burden on the urban population. When the Prime Minister quipped that living in Beijing would shorten his life by 5 years, he succinctly captured the reality of the risk air pollution poses to Chinese city dwellers. Addressing air pollution and other environmental concerns must be a priority for the Chinese authorities, both to manage public health costs and to lessen a potential impediment to effective economic growth.

Air quality in the United States and Europe has shown substantial improvements over the last couple of decades. Cleaner air has contributed to increased life expectancy in these developed countries.[1] However, in many countries in the developing world, air pollution control has been sacrificed in the name of economic development.

The severe and deteriorating air pollution situation in China is a case in point. The high particulate air pollution readings in Beijing has been documented and reported on the web by the US Embassy. The Chinese government has begun publishing real time measurements of air quality in most cities. There is increasing concern that poor air quality is not only harming the people but may also harming economic growth in China.

Winter air pollution events in the north of China are common, and the hazard is well recognised. “If I work in your Beijing, I would shorten my life at least five years,” Zhu Rongji told city officials when he was prime minister in 1999.[2] However, the government is just beginning to try and control or at least mitigate these air pollution events. In October 2013, a particularly severe air pollution episode in the northern city of Harbin was reported around the world. In blogs, people reported not being able to see their hands in front of their faces, or to see people they were speaking to. The Harbin government reported an air quality index (AQI) score of 500, the highest possible reading, and concentrations of PM2.5 — fine particulate matter that are 2.5 microns in diameter or smaller and especially harmful to health — of 1,000 micrograms per cubic meter (mg/m3).

Figure 1:  Harbin’s landmark San Sophia church was barely visible Monday as heavy pollution forced the closure of schools and highways

These anecdotal reports and quantitative measurements from Harbin (Figure 1) are remarkably similar to those from London during the 1952 Great Smog (Figure 2). Health data from Harbin have not been reported, but in London 4,000 excess deaths were attributed to this event. Recent analyses have suggested that the true number of excess deaths could be 12,000. [3]


Figure 2: A couple wearing smog masks take an afternoon stroll along The Embankment

This Harbin episode coincided with the mandated start of heating of the homes and offices. The policy of providing free coal for heating in the north has been associated with persistently high winter particulate air pollution levels in northern cities. A recent analysis [4] evaluated the effect of this policy. Outdoor ambient concentrations of particulate air pollution (Total Suspended Particulates) were found to be 55% higher and life expectancies 5½ years shorter in the north. Deaths due to cardiorespiratory causes were specifically increased.

In December 2012, the Global Burden of Disease analyses were published in The Lancet [5]. As part of that effort, average 2005 fine particle (PM2.5) air pollution was estimated across the world (Figure 3).[6]  Outdoor air pollution in China was estimated to contribute to 1.2 million premature deaths and 25 million healthy years of life lost.[7] Outdoor air pollution was ranked as the fourth leading risk for loss of life expectancy in China; and indoor air pollution from burning solid fuels for heating and cooking as the fifth leading cause.

Figure 3: Estimated 2005 annual average PM2.5 concentrations (μg/m3).(6)

These huge numbers of excess deaths and total years of lost life expectancy are compelling, but fail to communicate the risk to an individual of life-long exposure to extremely high air pollution, or the risk to visitors or temporary residents. The objective of this commentary is to provide useful, comparable effect estimates on loss of life expectancy under various exposure scenarios for exposure to air pollution and, for comparison, to cigarette smoke, a common, well studied risk. 


Our estimates of survival curves and life expectancy are derived using standard life-table techniques and are calculated using 2008 age-specific death rates for the total population of the United States.[8] The counterfactual, baseline, life expectancy for non-smokers is calculated adjusting the rates for ages 18 years and older to be 80% of rates from the total population. This provides hypothetical population-based mortality rates and estimates of life expectancy for a contemporary, healthy, non-smoking population. We estimated life expectancy for various exposure scenarios by multiplying the baseline age-specific death rates by the relative risks for each of these scenarios.

Excess risk estimates for the various air pollution exposure scenarios are based on recent literature reviews.[9, 10] Specifically, excess risk from exposure to air pollution in a mildly polluted city (15 µg/m3 mean PM2.5), a moderately polluted city (25 µg/m3 mean PM2.5), and a highly polluted city (55 µg/m3 mean PM2.5) relative to a very clean city (5 µg/m3 mean PM2.5) are estimated to be 7%, 14%, and 30%, respectively. The excess risk estimates for a highly polluted city may somewhat underestimate the effects of air pollution of Beijing for two reasons. First, average PM2.5 concentrations in Beijing are reported to be 58 µg/m3 in 2005 [6] and have been getting worse. Second, we are using more conservative risk estimates than would be obtained by linear extrapolations from U.S. cohort studies because of recent evidence that the exposure-response function flattens out at higher levels of exposure.[11] 

Results and discussion

Figure 4 illustrates differences in the life-table derived survival curves and life expectancy for the different exposure stylised scenarios. Cigarette smoking significantly adversely alters the survival curves. A lifetime of exposure to ambient air pollution in a highly polluted city has a similar, but less dramatic impact on the survival curves.  Lifetime exposure to second hand smoke (SHS) has a somewhat smaller, but similar effect (not shown in Figure 4).

Figure 4: Survival curves 18-100 years and estimated life expectancy (LE) for alternative excess risk assumptions

Table 1 presents the estimated years of life expectancy and the estimated reduction in estimated life expectancy, relative to the baseline.  Long-term active smoking clearly has a substantial impact on life expectancy—4½ to 12½ years lost, depending on the level of smoking.  The loss of life expectancy is substantially reduced for smokers that quit smoking. How much loss of life expectancy will occur depends on various factors including, level of smoking, the age when began and stopped smoking, and the lagged or residual excess risk from the smoking upon cessation. For an ex-smoker who smoked from age 18-40, life expectancy would be almost two years less than if he/she had never smoked, but nearly 6 years longer than if he/she had continued smoking.

Table 1:  Life-table derived estimates of reduced life expectancy from different exposures to cigarette smoke and ambient fine particulate matter air pollution

As can be seen in Table 1, living with a smoker throughout adult life time could reduce life expectancy by up to 2½ years. On the other hand, working with a smoker between 18 and 65 years, was estimated to reduce life expectancy by only 1 year, assuming the increased risk does not persist once exposure stops. Because of the relatively low baseline risks of mortality for children, exposure to SHS as a child results in a reduction in life expectancy of only about 74 days. If, however, the increased risks of childhood exposure to SHS persist, the reduction in life expectancy may be substantially larger.

The estimated reduction in life expectancy from a lifetime of exposure to ambient air pollution clearly depends on the level of pollution (Table 1). For example, lifetime exposure to air pollution in a mildly polluted city (15 µg/m3 mean PM2.5) or a moderately polluted city (25 µg/m3 mean PM2.5) relative to a clean city (5 µg/m3 mean PM2.5) results in an estimate of 0.8 and 1.6 years reduction in life expectancy. Life-time exposure to ambient air pollution, in a highly polluted city (comparable to Beijing, China) may result in an estimated loss of life expectancy of approximately 3 years.

Thus a lifetime of exposure to air pollution either from outdoor air pollution, indoor air pollution from SHS, or personal smoking can lead to years of lost life expectancy. Living in a highly polluted city has estimated effects comparable to or even greater than that from living with a smoker. Smoking, however, is a personal choice and only a fraction of the population engages in this voluntary exposure. On the other hand, breathing is not. The entire population is exposed to ambient air pollution. The net effect on population of a 3.1 year reduction in life expectancy across everyone breathing ambient air pollution is much larger than a 7.8 year reduction only among those smoking.

It is useful to compare these risks in terms of the incremental effect of each year of exposure. This helps us appreciate the effect of potential changes in exposures or behaviours. It also provides insights into the comparative risk for a worker or student who temporarily moves to such an environment.

To illustrate, Table 2 presents the estimated reduction in life-expectancy for a 50-year old non-smoker who spends one year in various modelled cites with mild, moderate, and high PM2.5 air pollution. Because the incremental reductions in estimated life expectancy for each year of exposure are small, we report these as days lost per year. One year of living in an elevated air pollution environment could result in as much as a few days to a few weeks of shorter life expectancy per year, depending on the levels of pollution and age at time of exposure. If the increased risks from a one-year exposure were to persist, even only in part, the reduction in life expectancy would be larger.  Because the baseline risk goes up with ageing, the impact of each year of exposure to high pollution on life expectancy is much larger at age 65 (21 days lost life expectancy) than it is at age 50. Likewise, the impact in lost life expectancy is less at age 35 years (5 days).

Table 2:  Estimates of reduced life expectancy for 1 year exposures to cigarette smoke and ambient fine particulate matter air pollution at 50 years of age

What does this mean for the individual?

These estimates are for the population life expectancy and do not provide specific estimates how much any individual’s life is shortened by one year of exposure in a polluted city. In fact in this modelling exercise, we assume that if the person survives this experience, they would go on with the normal expectation of death with no increased residual risk. Evidence from smoking cessation studies suggest that risk of fatal cardiovascular events (ischemic heart or cerebrovascular), the primary cause of death from these exposures, begins returning to near normal within a few weeks/months of cessation of smoking and is only somewhat elevated after a few years. For respiratory conditions, air pollution exposure can contribute to accelerated, irreversible loss of lung capacity. It may take months to years to return to normal risk, and indeed there may be permanent but small elevated risk.

For an individual, the implication of these results is not that their life is measurably shortened. Rather these estimates reflect the increase probability of death in each year. Again to illustrate, among one thousand (1,000) non-smoking 50 year olds, we would expect 3½ to die within a year. If all of them smoked, we would expect an additional 3½ to die. (Note, only a fraction of them would be expected to take up this behaviour.) Alternatively if the non-smokers experience PM2.5 air pollution of approximately 55 mg/m3 for a year we would expect one additional death. (Note; in this case everyone is at risk.)

These individuals would most likely die from acute events such as a myocardial infarction, stroke, asthma attack, or traffic accident. Note that we assume that once the air pollution exposure is removed (for example, by moving to a cleaner city) their excess risk of dying returns to normal.

While these estimates of los life expectancy for a 50 year old are simplistic, they provide a basis for comparing risks. Thus living/working in a moderately polluted city has comparable effects on life expectancy as living with a smoker or working in an environment with substantial second hand smoke.

Can the individual take actions to protect themselves, other than leaving?

The most effective strategy is to reduce your own baseline risk of cardiorespiratory death.[12] Air pollution affects those with pre-existing chronic cardiovascular conditions. Long-term strategies to reduce risk of cardiorespiratory disease as discussed in the accompanying articles have the added benefit of reducing the likelihood of death due to air pollution.

Masks and other breathing protection are not very effective in preventing individuals from breathing in (exposure to) ambient particles.[12]  Office and home air conditioning has some limited benefit in reducing exposures to ambient outdoor air pollution. Indeed gaseous pollutants (sulfur dioxide, ozone, nitrogen oxides, and other water soluble gases) are readily removed by air conditioning. However, the normal filters in air conditioners are only modestly helpful in removing inhalable, airborne particles. Office and home filters can be helpful if specifically designed for removal of small particles, for example, HEPA filters. However targeted filtering of inhalable particles is difficult, expensive, and requires regular cleaning and maintenance. Thus the preferred approach is cleaning up the ambient outdoor air. 

Summary and conclusions

From these calculations it becomes apparent that outdoor particulate air pollution is having a substantial effect on life-expectancy in much of the developing world.

As George Box reminded us “All models are wrong, but some are useful.”[13] This model of life expectancy valuing the effects of air pollution and cigarette smoking in the currency of days of lost life expectancy is simplistic, ignores many nuances in the actuarial data, and in that sense is clearly wrong. However, this approach helps us understand the comparative impact of air pollution relative to other known risk factors.

We are currently unable to measure all health costs or full losses of life due to air pollution in China, or elsewhere with absolute accuracy or certainty. However, when the Prime Minister quipped that living in Beijing would shorten his life by 5 years, he succinctly captured the reality that air pollution imposes a substantial health burden on the population in China.

Most countries have seen a dramatic improvement in life expectancy over the past 50 years that appears to be at least correlated with measures of economic development. Failure to address air pollution and other environmental concerns, however, is now being recognised as a significant public health burden and a potential impediment to effective economic growth.



1. Pope, C.A., 3rd, M. Ezzati, and D.W. Dockery, Fine-particulate air pollution and life expectancy in the United States. N Engl J Med, 2009. 360(4): p. 376-86.

2. China's Environment: A Great Wall of waste—China’s environment, in Economist. 2004, The Economist Newspaper Limited: London. p. 55–57.

3. Bell, M.L. and D.L. Davis, Reassessment of the lethal London fog of 1952: novel indicators of acute and chronic consequences of acute exposure to air pollution. Environ Health Perspect, 2001. 109 Suppl 3: p. 389-94.

4. Chen, Y., et al., Evidence on the impact of sustained exposure to air pollution on life expectancy from China's Huai River policy. Proc Natl Acad Sci U S A, 2013. 110(32): p. 12936-41.

5. Lim, S.S., et al., 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, 2012. 380(9859): p. 2224-60.

6. Brauer, M., et al., Exposure assessment for estimation of the global burden of disease attributable to outdoor air pollution. Environ Sci Technol, 2012. 46(2): p. 652-60.

7. AMBIENT AIR POLLUTION AMONG TOP GLOBAL HEALTH RISKS IN 2010: Risks Especially High in China and Other Developing Countries of Asia, in HEI International. 2013, Health Effects Institute: Boston, MA.

8. Arias, E., United States life tables, 2008. Natl Vital Stat Rep, 2012. 61(3): p. 1-63.

9. Pope, C.A., 3rd and D.W. Dockery, Health effects of fine particulate air pollution: lines that connect. J Air Waste Manag Assoc, 2006. 56(6): p. 709-42.

10. Brook, R.D., et al., Particulate matter air pollution and cardiovascular disease: An update to the scientific statement from the American Heart Association. Circulation, 2010. 121(21): p. 2331-78.

11. Pope, C.A., 3rd, et al., Cardiovascular mortality and exposure to airborne fine particulate matter and cigarette smoke: shape of the exposure-response relationship. Circulation, 2009. 120(11): p. 941-8.

12. Gold, D.R. and J.M. Samet, Air pollution, climate, and heart disease. Circulation, 2013. 128(21): p. e411-4.

13. Box, G.E.P. and N.R. Draper, Empirical Model Building and Response Surfaces. 1987, New York, NY: John Wiley & Sons.

Download full article (PDF 548 KB)


Douglas Dockery

Professor of Environmental Epidemiology; Chair, Department of Environmental Health, Harvard School of Public Health

Doug Dockery, MS, ScD, is internationally known for his innovative work in environmental epidemiology, particularly in understanding the relationship between air pollution and respiratory and cardiovascular mortality and morbidity. He was one of the principal investigators of the landmark Six Cities Study of Air Pollution and Health, which showed that people living in communities with higher fine particulate air pollution had shorter life expectancies.

Mr Dockery has studied the health effects of air pollution in studies of people who have been followed for a few months up to 25 years. His research has shown that combustion particles in the air are linked to increased morbidity and mortality even at the relatively low concentrations observed in developed countries today. Specifically, his work has shown that episodes of particulate air pollution are associated with increased numbers of deaths, increased hospital admissions and emergency room visits, respiratory conditions including asthma attacks, increased respiratory symptoms and lower lung function and cardiovascular conditions including heart attacks and cardiac arrhythmias. Long-term follow-up studies have shown particulate air pollution is associated with shortened life expectancy in adults and increased chronic respiratory illness and lower lung function in children. This research has led to the current standards for particulate air pollution both nationally and internationally. He was first author of the most cited air pollution paper in the peer-reviewed literature.

Mr Dockery is currently evaluating the benefits of improved air quality on people’s health. He has been mentor to some of the outstanding investigators in environmental epidemiology including Bert Brunekreef, Annette Peters, Arden Pope, and Joel Schwartz. The International Society for Environmental Epidemiology honored him with its first award for Outstanding Contributions to Environmental Epidemiology in 1999 and the first Best Paper in Environmental Epidemiology Award in 2010.

C. Arden Pope III

Mary Lou Fulton Professor of Economics, Department of Economics, Brigham Young University, Provo, Utah

C. Arden Pope III, PhD has conducted or collaborated on various key studies of human health effects of short- and long-term air pollution exposure.He has played prominent roles in reviewing and interpreting this literature and is a widely cited and recognized expert on the health effects of air pollution.

He has been the recipient of various honors and awards including the Thomas T. Mercer Joint Prize from the American Association for Aerosol Research and the International Society for Aerosols in Medicine (2001) and is an Honorary Fellow of the American College of Chest Physicians (FCCP Hon, 2008).

Related articles

Overview of health risk factors in Brazil

Marcia C. Castro, 16 Jan 2014

Brazil has made significant progress in reducing infectious diseases, although problems still persist. Currently non-communicable diseases (NCDs) represent the largest mortality and morbidity burden in the country. This situation may become more critical as the population ages and the prevalence for being overweight and obese increases. Brazil has a historically good record of facing health challenges through large scale campaigns. It most recently became the global leader in tobacco control, despite being among the top 5 producers of tobacco in the world. This article discusses historical and current pattern of mortality, morbidity and associated risk factors in Brazil, particularly when faced with structural demographic change. It also discusses government programmes, such as the Family Health Strategy and the Strategic Action Plan to tackle NCDs.

Risk factors for cardiovascular disease in Brazil: Time trends and current status

Marcia C. de Oliveira Otto, 16 Jan 2014

As in other emerging economies, chronic conditions such cardiovascular disease (CVD) and diabetes have emerged as the major causes of mortality in Brazil, accounting for 37% of deaths in the country. Although there has been a considerable reduction of smoking and physical inactivity, the epidemiological transition in Brazil is marked by trends toward unhealthy diet patterns and a poor cardio-metabolic profile. This is characterised by increased consumption of processed foods, as well as a higher prevalence of overweight individuals, obesity, impaired fasting glucose and diabetes, potentially reversing the decline in CVD mortality in the medium- to long-term.

Time trends of cardiovascular disease risk factors in China

Yangpin Li, MD, PhD, Frank Hu, 16 Jan 2014

Fueled by rapid urbanization and changes in dietary and lifestyle choices, cardiovascular and other chronic diseases have emerged as a critical public health issue in China. The prevalence of hypertension in 2010 reached 33.5% (an estimated 330 million hypertensive patients), and the awareness and control rates are extremely low. Type 2 diabetes is an increasing epidemic in China with more than 100 million people affected. Although the Chinese population has a lower BMI than the global average level, abdominal obesity has become especially common in Chinese adults. Despite tobacco control efforts, the prevalence of smoking in China remains at a high level and domestic production of cigarette continue to rise. With unprecedented growth in urbanization, work and transportation-related physical activity levels have declined sharply, accelerating the epidemics of obesity and chronic diseases, which not only affect health and quality of life, but also have economical and social consequences.

Disease Risk Factor Matrix: improving risk assessment in populations with changing disease risk factors

16 Jan 2014

Risk Dialogue Series: Understanding the Drivers of Longevity - The global burden of chronic disease is expected to exceed USD47 trillion in the next 20 years. Preventing and predicting disease risk in various populations is critical for adapting to and curtailing future disease burden and economic cost burdens. However, understanding future disease risk and economic burden is not easy, and poses difficulties for public health planning as well as actuarial financial models.