Blackwell Science, LtdOxford, UK

RESRespirology1323-77992003 Blackwell Science Asia Pty Ltd

82June 2003
460
COPD in Asia–Pacific
WC Tan
et al.
10.1046/j.1323-7799.2003.00460.x
Original Article192198BEES SGML

Respirology (2003) 8, 192–198

ORIGINAL ARTICLE

COPD prevalence in 12 Asia–Pacific countries and regions:

Projections based on the COPD prevalence estimation model
REGIONAL COPD WORKING GROUP*

COPD prevalence in 12 Asia–Pacific countries and regions: Projections based on the COPD
prevalence estimation model
REGIONAL COPD WORKING GROUP. Respirology 2003; 8: 192–198
Objective: COPD is a leading cause of mortality and morbidity worldwide. Despite the high rates
of cigarette smoking, and the wide use of biomass fuels, there is very little objective data on the
prevalence of COPD in Asia.
Methodology: We used a COPD prevalence model to estimate the prevalence of COPD in 12 Asian
countries. This model is a validated, computerized tool that uses epidemiological relationships and
risk factor prevalence to project the prevalence of COPD within a given population aged 30 years
and older.
Results: The total number of moderate to severe COPD cases in the 12 countries of this region, as
projected by the model, is 56.6 million with an overall prevalence rate of 6.3%. The COPD prevalence
rates for the individual countries range from 3.5% (Hong Kong and Singapore) to 6.7% (Vietnam).
Conclusions: The COPD prevalence rates projected by the model reflect the high prevalence of the
risk factors for the disease in Asia. The combined prevalence of 6.3% for these countries is consid-
erably higher than the overall rate of 3.8% as extrapolated from WHO data for this region. These
estimates highlight the need for further epidemiological studies to support appropriate allocation
of resources for the prevention and management of COPD.

Key words: Asia, chronic bronchitis, cigarette smoking, COPD, emphysema, environmental pollu-
tion, epidemiology, indoor air pollution predictive model, obstructive airways disease, prevalence.

INTRODUCTION dicted that it would rise in rank to be the third leading

cause of death after ischaemic heart disease and cere-
The World Health Organization (WHO) global burden brovascular disease by the year 2020.1,2
of disease study reported that COPD was the sixth COPD is formally defined by spirometric criteria
leading cause of death worldwide in 1990 and pre- as a slowly progressive disorder characterized by
airways obstruction (FEV1 < 80% predicted and
FEV1/FVC ratio < 70% predicted) that is not revers-
ible.3,4 This broad definition may encompass several
Correspondence: Wan C. Tan, Department of Medi- different clinical and pathological entities, includ-
cine, National University of Singapore, 5 Lower Kent ing peripheral airways obstruction from chronic
Ridge Road, Singapore 119074, Republic of Singapore. bronchitis, emphysema, and bronchiectasis and
Email: [email protected] tuberculosis.
Professor Wan C. Tan took no part in the decision- Cigarette smoking is a major risk factor worldwide
making process with respect to this manuscript. for COPD, although it is not known why only about
* Members of Regional COPD Working Group include: 15% of smokers develop clinically significant COPD.
Professor Wan C Tan, Singapore; Professor J. Paul Seale, Cigarette smoking has been estimated to account for
Australia; Professor Suchai Charaoenratanakul, Thai- 80–90% of all COPD in the population.5 Smoking
land; Dr Teresita de Guia, Philippines; Professor Mary prevalence rates for Asian countries are extremely
Ip, Hong Kong (S.A.R., China); Dr Aziah Mahayiddin, high and continue to increase. According to the WHO
Malaysia; Dr Hadiarto Mangunnegoro, Indonesia; Pro- estimates, the number of COPD cases in Asia exceeds
fessor Kwen-Tay Luh, Taiwan; Professor Young-Soo by three times the total number of COPD cases for the
Shim, Korea; Professor Nan-Shan Zhong, China and Dr rest of the world.1,5
Brigitte Schau, Germany. Both indoor and outdoor air pollution are recog-
Received 10 May 2002; revised 23 October 2002; nized risk factors for COPD. The use of solid fuels for
accepted for publication 24 October 2002. indoor cooking or heating with poor ventilation may
COPD in Asia–Pacific 193

produce significant air pollution, which may account Model inputs

for the development of COPD. This causal factor may
be particularly relevant in several Asia–Pacific coun- The model is based on a systematic review of the
tries, where biomass fuel use is common. It has been published literature for the identification and quanti-
reported that up to 77% of the rural population in the fication of major risk factors for COPD. Model inputs
Asia–Pacific region may be using wood and dung or were derived from studies that provided data quanti-
crop residue for fuel.6 fying the relationship between COPD, its risk factors
Epidemiological studies of COPD in the Asia– and its prevalence.10–20 Choice of input categories was
Pacific region are scant. A study of elderly Chinese also based on the use of data that would be readily
living in Hong Kong found a COPD prevalence of available to the end user, from relevant clinical and
6.8%.7 The crude prevalence of chronic bronchitis epidemiological sources found in most countries.
has been reported to be over 18% in a rural commu- Although there exist additional risk factors beyond
nity in Nepal.8 There is a need for more reliable data those utilized in the model, it was felt that four fac-
on the prevalence of COPD in individual countries of tors were key and that the effects of these factors
the Asia–Pacific region. Such information is essential could be assumed to equally impact on the preva-
for the development of national health policies and lence across a population and to account for the
strategies for the prevention and control of the majority of risk. These factors included smoking,
disease. biomass fuel exposure, air pollution, and high-risk
Community-based survey data would provide the occupations.
most reliable estimates of COPD prevalence but it The populations exposed to these factors are first
is unlikely that most countries in Asia will have determined in the model algorithm. This determina-
the resources to conduct such detailed surveys in the tion is derived from user input of the following data:
immediate future. An alternative approach is to use smoking prevalence by age and sex, per cent of pop-
an epidemiological model to estimate the prevalence ulation living in urban areas (as a surrogate for
of COPD.9 This approach has the advantages of being moderate to severe air pollution), per cent of popu-
less costly, and being able to promptly provide an lation in high-risk occupations, and population
interim estimate of the size of the problem using stan- exposed to biomass fuel. If the user is unable to
dardized methodology and hence facilitate cross-sec- determine the above inputs, the model assigns
tional and serial comparisons between and within default values for these based on evidence from the
countries. We report the prevalence of COPD for 12 literature.
Asian countries using such a model.

Model outputs
METHODS
The outputs or projections of the model are based on
Description of model the relationships of the major risk factors for COPD to
the prevalence of COPD. Once the user has entered
The COPD Prevalence Estimation Model is an algo- the relevant data, the populations exposed to each of
rithm embedded in a software program.9 It estimates the risk factors are calculated by the model algorithm.
the current COPD population for a specific region or Based on the figures generated from the literature
country, based on the local prevalence of risk factors for cigarette smokers,10,11 users of biomass fuel,6,12–15
for COPD. Figure 1 shows the high level model pro- those in high-risk occupations,16,17 those living in
cess flow. The model estimates the number of COPD areas with moderate/severe air pollution,18–20 and
patients in the 30 years and older population. those unexposed to any of the above,16 the model then
determines the prevalence of COPD for each of these
population categories.
Additional features of the model include projec-
tions of the distribution of COPD cases into moder-
ate/severe or mild disease, and future projections for
the prevalence of COPD based on changing demo-
graphics. (Details of the model development process
and its features are available in reference or by writing
to WCT).

Data collection for application model

A questionnaire was developed to assist in collect-

ing the data required to populate the model. This
questionnaire contained fields for all the manda-
tory as well as optional data elements required for
Figure 1 Model scheme and process flow. input into the model. Mandatory data consisted of
194 WC Tan et al.

status of country as developed or developing, pop- Data management and analysis

ulation size by age groups, smoking prevalence and
percentage of population living in urban or rural All data were then returned to the data coordinator
areas. Optional data consisted of percentage of for input into the computer model. The mandatory
population exposed to non-smoking risk factors for data elements were available for all the countries, but
COPD and percentage of smokers who develop the optional data elements were not available for
COPD. The questionnaires were sent to representa- most countries. To ensure consistency, only the man-
tive COPD experts from each participating country datory data elements were ‘input’ into the model for
along with standardized guidelines for completing all the countries. Default values derived from pub-
the questionnaires. The experts were requested to lished studies are provided by the model algorithm for
provide the data for each country and to specify use in case the optional data elements are not avail-
the source from which each data element was able. These default values were then used as the
obtained. For smoking prevalence rates, the experts optional data elements for each country in order to
were asked to use the most recent figures available generate the model projections. Figures 2 and 3 illus-
from each country. Official population figures for trate the main mandatory data elements collected
the year 2000 provided by the Census Bureau from from each country. The model projections for moder-
each country were used to populate the model. If ate to severe COPD were then generated for each
these figures were not available, the year 2000 pop- country for its population in the year 2000 (Table 1).
ulation was estimated based on the last census, and
the annual change in the population observed over
the last 5 years. The percentages of the population RESULTS
living in urban and rural areas were similarly based
on the figures from the last census performed in As projected by the model, the total number of mod-
the country. erate to severe COPD cases among individuals 30

Figure 2 Smoking prevalence

by country and gender.

Figure 3 Urban population as

per cent of total population by
country.
COPD in Asia–Pacific 195

years and older within all the 12 identified countries, report prevalence as a percentage of the entire adult
is 56.6 million. This translates to a mean COPD prev- population or by specific ages within the population.
alence rate of 6.3% for the region. Table 1 shows the The overall smoking prevalence rates, which have
data for each country. The COPD prevalence rates the greatest impact on the model projections, also
vary twofold between the 12 Asian countries and vary considerably between these countries (14% in
range from a minimum of 3.5% (Hong Kong and Hong Kong and 36% in Japan and Vietnam). Differ-
Singapore) to a maximum of 6.7% (Vietnam). ences in smoking prevalence rates between the
Table 2 compared the overall result from the female population (range 2–21%) and male popula-
present study with similar estimates made for other tion (range 27–73%) may contribute to these varia-
countries using the same model. The rate for the Asia– tions between the countries of the region (Fig. 2).
Pacific region was intermediate between those of the The proportion of the population living in urban
USA and Nepal, but was closer to those of Denmark areas, another major factor in the model projections,
and Norway. Table 2 also shows that the estimates also differs widely between the countries. The highest
vary depending on the criteria used to define COPD urban populations are in Singapore and Hong Kong
in the individual studies. For instance, some studies (100%), while the lowest urbanization is reported
used spirometry alone, while others used symptoms, from Thailand (12%) (Fig. 3, Table 3).
with or without spirometry. In addition, countries Table 3 shows the estimated number of cases
attributed to smoking (tobacco exposure) versus non-
smoking causes (occupational exposure, biomass
fuel exposure and outdoor air pollution). The ratio
Table 1 Model projections of the prevalence of moderate of COPD cases from smoking versus non-smoking
to severe COPD in those 30 years and older for 12 countries causes ranged from 1.52 for Thailand to 5.81 for
in the Asia–Pacific region Japan.

Moderate/severe
Country COPD cases Prevalence DISCUSSION

1. Australia 558 000 4.7% In the implementation phase of global practice guide-
2. China 38 160 000 6.5% lines (GOLD), information on the burden of disease is
3. Hong Kong 139 000 3.5% essential for the planning and allocation of healthcare
4. Indonesia 4 806 000 5.6% resources.21 Epidemiological data on the prevalence
5. Japan 5 014 000 6.1% of COPD are available from very few countries, mak-
6. South Korea 1 467 000 5.9% ing it necessary to generate estimates for understand-
7. Malaysia 448 000 4.7% ing the global size of the problem. In the WHO Global
8. Philippines 1 691 000 6.3% Burden of Disease study, data for the burden of COPD
9. Singapore 64 000 3.5% were presented according to three categories: estab-
10. Taiwan 636 000 5.4% lished market economies, former socialist economies
11. Thailand 1 502 000 5.0% of Europe and a third category called demographi-
12. Vietnam 2 068 000 6.7% cally developing regions which included China, India
Total 56 553 000 6.3% and ‘other Asian islands’.3 There is, therefore, insuffi-
cient focus and detailing of the countries in the Asia–

Table 2 Comparison of model estimates of COPD prevalence in various countries with the overall model estimate of COPD
prevalence in the Asia–Pacific region

Country No. COPD cases COPD prevalence by model* COPD prevalence in literature

Asia–Pacific* 56 553 000 6.3% 3.8%a

UK 1 817 000 5.0% 2.1–3.9%b
US 7 477 000 4.8% 6.8%c
Denmark 214 000 6.4% 3.7%d
Norway 167 000 6.3% 5.4%e
Nepal 897 000 11.1% 18.3%f
Russia 6 526 000 7.6% N/A

*Sum of the figures for 12 countries in the Asia–Pacific region in this study in those 30 years of age and older
a
WHO by extrapolation.
b
40–74 year olds by symptoms.
c
≥ 17 year olds by spirometry.
d
20–90 year olds by spirometry.
e
18–70 year olds by symptoms and spirometry.
f
≥ 20 years age group by symptoms.
N/A, not applicable.
196 WC Tan et al.

Table 3 Projected number of cases of COPD due to smoking and non-smoking causes in those 30 years and older in 12
Asia–Pacific countries

Non-smoking details
COPD cases Population Urban (exposure)*
Non- Rural Urban Air Rural (exposure)**
Country Smoking Smoking % % pollution Occupational Occupational Biomass Unexposed

Australia 436 000 122 000 20% 80% 96% 4% 5% 0% 95%

China 29 394 000 8 767 000 69% 31% 96% 4% 1% 87% 12%
Hong Kong 91 000 48 000 0% 100% 96% 4% 0% 0% 0%
Korea 1 241 000 226 000 22% 78% 96% 4% 5% 0% 95%
Malaysia 354 000 94 000 44% 57% 96% 4% 5% 0% 95%
Philippines 1 269 000 422 000 70% 30% 96% 4% 1% 87% 12%
Singapore 42 000 22 000 0% 100% 96% 4% 0% 0% 0%
Taiwan 524 000 113 000 20% 80% 96% 4% 5% 0% 95%
Vietnam 1 643 000 424 000 60% 40% 96% 4% 1% 87% 12%
Thailand 906 000 595 000 88% 12% 96% 4% 1% 87% 12%
Indonesia 3 410 000 1 396 000 64% 36% 96% 4% 1% 87% 12%
Japan 4 278 000 736 000 10% 90% 96% 4% 5% 0% 95%

*Urban exposure, outdoor air pollution and occupational; **rural exposure, biomass and occupational.

Pacific region which is distinctive in several ways. It is proportion of the population living in rural areas. The
a heterogeneous region that is geographically and presumed risk for COPD in rural dwellers is exposure
economically diverse, and includes countries with to biomass fuels, which are commonly used for cook-
developed established market economies such as ing and heating in poorly ventilated dwellings, lead-
Japan and Australia, and countries from the develop- ing to high levels of particulate matter in indoor
ing regions such as China and Southeast and East air.6,12–15 The contributions to the burden of COPD
Asia. The region is the most populous and arguably from occupational exposure and from outdoor air
the most economically dynamic with rapidly chang- pollution are small compared with that of cigarette
ing demographics, lifestyle and disease patterns. smoking. The countries with the lowest prevalence
The present study was initiated with the aim of pro- have the lowest rates of smoking and the lowest pro-
ducing estimates of the prevalence of COPD for this portion of rural dwellers. Although these figures for
vast region using a validated algorithm model which different risk factors are estimates, nevertheless they
utilized easily obtainable and reasonably reliable are helpful for healthcare planning.
national statistics and standardized epidemiological The overall prevalence of 6.3% for COPD in the 12
default data as inputs in order to generate preva- countries surveyed is higher than the figure of 3.8%
lence estimates for intercountry and interregion that has been extrapolated from data in the WHO
comparisons. report.1,2 This rate has been derived to provide a com-
The model is especially applicable to Asia because parison of the prevalence rate for a similar region to
it also incorporates COPD cases caused by non- the 12 countries in the present study.
smoking factors, which are increasingly being There are several reasons why the estimate in the
recognized as having a significant impact on COPD present study varies from the WHO data of Murray
prevalence in developing countries. and Lopez.2 First, the demographic regions reported
The results from the model provided an estimate of by Murray and Lopez (1997) to have a prevalence of
the size of the problem for individual countries where 3.8% were designated ‘China and other Asian islands’.
no previous data existed. The extent of the variation The 12 countries reported in the present study to have
in smoking prevalence between the countries in the a prevalence of 6.3% included China and nine other
Asia Pacific region is unique compared with that in countries in common with those reported by Murray
the Western countries as the high rates are found not and Lopez.2 However, the present study also included
only in countries with established economies such as Japan and Australia, both of which have a prevalence
Australia and Japan but also in developing countries greater than 3.8% (4.7 and 6.2%, respectively) but the
such as Vietnam and China. The size of the tobacco present study did not include India.
problem highlights the urgent need for effective Second, the data of Murray and Lopez were
national tobacco control programs as the key to the recorded as the total prevalence aggregated across
primary prevention of COPD and the consequent several age bands, one of which was 15–44 years. To
reduction of the burden of disease. determine the prevalence in the population aged 30
The study clearly shows that there are considerable years and older (to compare with the data in the
variations in the prevalence of COPD (range 3.5– present study), it was necessary to make an estimate
6.7%) in the countries of the Asia–Pacific region. The of what proportion of the age band between 15 and
major determinants of this variation are differences 44 years was over the age of 30. Hence, there is some
between countries in smoking prevalence and in the imprecision in this estimate.
COPD in Asia–Pacific 197

Third, our data on smoking prevalence, which is highly dependent upon the accuracy of smoking
the major determinant of COPD prevalence, is prob- rates. We believe that we have obtained accurate
ably accurate as it has been sourced from national information on smoking for the countries included in
databases in the respective countries, which are the present study. Another drawback of the model is
reported in the present study. In contrast, Murray and that it does not take into account the influence of the
Lopez had to resort to ‘encouraging experts to make interaction of several concurrent risk factors on
estimates in regions where no data were available’. the prevalence of COPD.
Compared with the literature8,23–25 cited for the Furthermore the model does not include all of the
countries in Table 2, the model appears to both over- potential risk factors for a population, particularly
estimate and underestimate depending on factors those genetic factors associated with ethnicity. How-
such as the criteria for definition of COPD and the ever, objective information of this type is currently
age group of the study population. One of the issues limited in the literature. In addition, the intent of the
is that the model predicts in the population aged model is to provide a reliable estimate of COPD in a
over 30 years, while the studies cited vary in their population to help guide policy makers and health-
populations. Also, for Nepal8 in particular, the diag- care advocates. We believe the current model ade-
nosis is by symptoms for two rural communities. It is quately fulfills this important role.
unclear whether this number truly reflects the total It must also be noted that it is possible that these
COPD population in the country. The UK study estimates of COPD prevalence may be underesti-
quoted is for chronic bronchitis symptoms only.23 mates because the model’s logic is based on the spiro-
The Norway study design is population-based and metric definition of obstruction based on FEV1/FVC
uses both symptoms and spirometry (5.4% is the which is widely used in published studies. The use of
overall prevalence).16 There is one prevalence study absolute FEV1/FVC ratio in defining COPD may result
for Estonia (a former Soviet Republic) where the in the potential underestimation of the COPD burden
English abstract stated merely that ‘5.5% to 9.3% of in the population due to a number of technical fac-
the population corresponded to the definition of tors. The variation of the lower normal limit of FEV1/
chronic airway disease’ (Table 2). To date, when FVC ratio with age may result in an underestimation
actual population-based studies of prevalence using of the total disease burden. Furthermore, the use of
symptoms and spirometry for definition16 and the FEV1/FVC ratio rather than FEV1/VC may also result
model have been compared, differences have been in an overestimation of the ratio (hence underestima-
small. For example, for Norway the model estimates tion of the burden) as the FVC is lower than the VC
a prevalence of 6.3% and an actual study reported a because of earlier airway closure due to a forced
prevalence of 5.4%.16 manoeuvre in patients with airway obstruction.22
The model also projected for COPD cases that were Finally, the model’s focus on moderate-to-severe
due to non-smoking factors which comprise exposure cases is likely to result in overall underestimation of
to high-risk occupations, biomass fuel and ambient total COPD burden.
air pollution. Although air pollution has decreased The present study may appear to have overesti-
significantly in developed countries, it is becoming a mated the prevalence to a small extent as has been
major concern in many cities in developing countries. shown in countries where direct comparison between
It is unclear which components of ambient air pollu- the prevalence determined by field studies and the
tion are the most harmful, but there is evidence that model has been made. However, for the reasons
particles in polluted air add to the individual’s inhaled stated above, it is probably a more accurate estimate
burden. There is also no information on the long- than the previous estimate of Murray and Lopez.2
term cumulative effects of repeated severe exacerba- Nevertheless, the COPD prevalence figures in this
tions of air pollution. report should be interpreted with caution because
The projection for COPD cases due to non- they are derived from a statistical model, which
smoking risk factors may have been overestimated requires further validation in other epidemiological
for some countries in the study. In both Hong Kong settings in order to confirm its accuracy. Despite
and Singapore, the proportion of non-smoking these limitations the main value of the model lies in
COPD cases is higher than that in other regions. its ease of application and ability to provide system-
Both regions were considered to be entirely urban atic estimates of COPD prevalence and to provide
areas. The model’s default calculations enhance the some insight into the relative contributions of
contribution of air pollution to the non-smoking tobacco and non-tobacco risk factors for planning
COPD cases for urban regions. This may result in an purposes. Because these figures are only statistical
overestimation of total COPD cases. Although air estimates there is still a need for properly conducted
pollution has been considered an important factor population-based studies of COPD in the Asia–
in respiratory disease, its exact role in causing Pacific region in order to confirm their accuracy.
COPD is not clear.18–20 Until further studies can While we await the results of these field studies, the
quantify the impact of air pollution, it appears rea- model has given us a very good idea of the extent of
sonable to include air pollution as a risk factor for the problem of COPD in the region. These figures
COPD development. from the model are important as they have generated
The main limitation of the model is that it is hypotheses that await further research and can be
extremely sensitive to estimates of smoking preva- used to assist policy makers as the first step in deter-
lence, and errors in smoking rates will greatly influ- mining COPD expenditures and establishing
ence the prevalence predictions. Hence, the model is resource allocation priorities.
198 WC Tan et al.

ACKNOWLEDGEMENTS women. A case-control study. Am. J. Respir. Crit. Care

Med. 1996; 154: 701–6.
We would like to thank Dr Carol Zaher, Dr Ahmar 13 Zhang J, Smith KR. Hydrocarbon emissions and health
Iqbal, Dr Manuel Distel and Ms Serene Hsu for their risks from cook stoves in developing countries. J. Expo
invaluable and indispensable help in coordinating Anal. Environ. Epidemiol. 1996; 6: 147–61.
the preparation of the manuscript. 14 Chen BH, Hong CJ, Pandey MR, Smith KR. Indoor air
pollution in developing countries. World Health Stat. Q.
1990; 43: 127–38.
15 Smith KR. National burden of disease in India from
REFERENCES indoor air pollution. Proc. Natl Acad. Sci. USA 2000; 97:
13 286–93.
1 Murray CJL, Lopez AD. Global Health Statistics. A Com- 16 Bakke PS, Baste V, Hanoa R, Gulsvik A. Prevalence of
pendium of Incidence, Prevalence, and Mortality Esti- obstructive lung disease in a general population: rela-
mates for Over 200 Conditions. Harvard University Press, tion to occupational title and exposure to some airborne
Cambridge, 1996. agents. Thorax 1991; 46: 863–70.
2 Murray CJL, Lopez AD. Mortality by cause for eight 17 Oxman AD, Muir DC, Shannon HS, Stock Sr, Huizdo E,
regions of the world; global burden of disease study. Lange HJ. Occupational dust exposure and chronic
Lancet 1997; 349: 1269–76. obstructive pulmonary disease. A systematic overview of
3 COPD Guidelines Group of the Standards of Care Com- the evidence. Am. Rev. Respir. Dis. 1993; 148: 38–48.
mittee of the British Thoracic Society. BTS guidelines for 18 Rokaw SN, Detels R, Coulson AH et al. The UCLA popu-
the management of chronic obstructive pulmonary dis- lation studies of chronic obstructive respiratory disease.
ease. Thorax 1997; 52 (Suppl. 5): S1–28. 3. Comparison of pulmonary function in three commu-
4 American Thoracic Society. Standards for the diagnosis nities exposed to photochemical oxidants, multiple pri-
and care of patients with chronic obstructive pulmo- mary pollutants, or minimal pollutants. Chest 1980; 78:
nary disease. Am. J. Respir. Crit. Care Med. 1995; 152: 252–62.
S77–120. 19 Pope AC, Bates DV, Raizenne ME. Health effects of par-
5 Collishaw NE. The Tobacco Epidemic: A Global Public ticulate air pollution: time for reassessment? Environ.
Health Emergency Tobacco. World Health Organization, Health Perspect. 1995; 103: 472–80.
Geneva, 1996. 20 Kunzli N, Kaiser R, Medina S et al. Public health impact
6 de Koning HW, Smith KR, Last JM. Biomass fuel combus- of outdoor and traffic-related air pollution: a European
tion and health. Bull. World Health Organ. 1985; 63: 11– assessment. Lancet 2000; 356: 795–801.
26. 21 Global Initiative for Chronic Obstructive Lung Disease.
7 Woo J, Pang J. Spirometry in healthy elderly Chinese. Global strategy for the diagnosis, management, and pre-
Thorax 1988; 43: 617–20. vention of chronic obstructive pulmonary disease.
8 Pandey MR. Prevalence of chronic bronchitis in a rural NHLBI/WHO, report. NIH, NHLBI publication Number
community of the hill region of Nepal. Thorax 1984; 39: 2701 April 2001.
331–6. 22 American Thoracic Society. Standardization of spirome-
9 Schau BS, Lopez-Vidriero M, Peabody J et al. Chronic try, 1994 Update. Am. J. Respir. Crit. Care Med. 1995; 152:
Obstructive Pulmonary Disease. A prevalence estima- 1107–36.
tion model. Am. J. Respir. Crit. Care Med. 2000; 161: 23 Littlejohns P, Ebrahim S, Anderson R. Treatment of adult
A799. asthma: is the diagnosis relevant? Thorax 1989; 44: 797–
10 Fletcher C, Peto R. The natural history chronic airflow 802.
obstruction. BMJ. 1977; 1: 1645–8. 24 Mannino DM, Gagnon RC, Petty TL, Lydick E. Obstruc-
11 Tashkin DP, Clark VA, Coulson AH et al. The UCLA pop- tive lung disease and low lung function in adults in the
ulation studies of chronic obstructive respiratory dis- United States: data from the National Health and Nutri-
ease VIII. Effects of smoking cessation on lung function: tion Examination Survey, 1988–1994. Arch. Intern. Med.
a prospective study of a free-living population. Am. Rev. 2000; 160: 1683–9.
Respir. Dis. 1984; 130: 707–15. 25 Lange P, Groth S, Nyboe J et al. Chronic obstructive lung
12 Perez-Padilla R, Regalado J, Vedal S et al. Exposure to disease in Copenhagen: cross-sectional epidemiological
biomass smoke and chronic airway disease in Mexican aspects. J. Intern. Med. 1989; 226: 25–32.