9.4 Respiratory disease
Oxford Textbook of Public Health
T. H. Lam and A. J. Hedley
Factors in respiratory disease
Global respiratory mortality
World burden of respiratory diseases
Smoking and respiratory deaths in developed countries
Deaths from smoking in developing countries
Recent international and national actions
Occupational factors and respiratory diseases
Respiratory health and passive smoking
Passive smoking at work
Passive smoking in Asia
Smoking, passive smoking, and occupational health
Chronic obstructive pulmonary disease
Adverse health effects of air pollution
Air quality and sustainable development
Factors affecting pollutant levels and exposures
Assessment of health effects
Health effects of pollutants
Mortality from air pollution
Air pollution, morbidity, and health-care utilization
Asthma and wheezing
Air pollution and lung function
Policy-making and prevention
The changing global situation
Identifying problems in global control of tuberculosis
Surveillance of tuberculosis
Factors influencing global trends in tuberculosis
Provision of care for tuberculosis and compliance with treatment
The need for a public health approach
Factors in respiratory disease
Factors that are associated with lung disease but are not readily avoidable include genetic constitution, age, gender, and socio-economic status. Tobacco smoking and environmental air pollution are the principal avoidable causes of morbidity and mortality from the most common and serious chronic lung diseases, lung cancer, and obstructive lung disease. High dietary intake of fresh fruit, including consumption during winter, which raises serum levels of vitamin C, is inversely related to respiratory symptoms and physician-diagnosed bronchitis (Schwartz and Weiss 1990), ventilatory function, and chronic non-specific lung disease (Strachan et al. 1991; Meidema et al. 1993; Schwartz and Weiss 1994), but the relationship between dietary factors and non-malignant respiratory problems is under-researched and has not attracted adequate attention in public health actions against respiratory disease.
A comprehensive review by the World Cancer Research Fund (1997) concluded that there is convincing evidence that consumption of vegetables and fruits protects against lung cancer and that carotenoids in foods of plant origin are probably protective. Regular physical activity, diets high in vitamins C and E and selenium possibly reduce the risk of lung cancer, and diets high in total fat, saturated fat and cholesterol, and in alcohol, possibly increase risk. Dose–response relationships were observed that show that the relative risk decreases by about 50 per cent as vegetable intake increases from 150 to 400 g/day. A United Kingdom Department of Health Working Group on Diet and Cancer (Department of Health 1998) concluded that the evidence was not strong enough to support causation, but that there is moderately consistent evidence that higher fruit consumption, and weakly consistent evidence that higher vegetable consumption, are associated with lower risks of lung cancer. A plausible mechanism is the antioxidant capacity in protecting against free-radical-induced DNA damage; however, b-carotene and a-tocopherol appear unlikely to be the mediators and the strongly consistent negative association between serum b-carotene and lung cancer has not been confirmed as causal by intervention studies (Omenn et al. 1996). The Working Group also warned that there is no evidence that higher fruit and vegetable consumption would mitigate the overwhelming effect of smoking on lung cancer and that smokers with a high consumption of vegetables and fruits are still at high risk of lung cancer.
Derby and Samet (1994) estimated that radon causes 10 per cent of all lung cancer in the United States and 6 per cent in the United Kingdom, but there is still much uncertainty about the role of non-occupational or domestic exposure especially in developing countries. Whereas reduction of radon exposure will certainly lower the risk of lung cancer, it should be noted that the issue of indoor air pollution, including radon emission, is often used by the tobacco industry to confuse or to divert attention away from the issue of passive smoking, which is more readily preventable than radon exposure.
Global respiratory mortality
The World Development Report 1993 provides a global overview of health status and health systems with special emphasis on developing countries (World Bank 1993). It describes the world pattern of mortality and burden of disease, including respiratory health problems. These data are also presented separately for developed and developing countries. The World Bank (1993) data have been revised by Murray and Lopez (1996, 1997a), and hence the results and Table 1 and Table 2 supersede the corresponding data in the previous edition of this book (Hedley and Lam 1997).
Table 1 Deaths caused by respiratory diseases in the world, 1990 (thousands)
Table 2 Burden of respiratory diseases in the world (1990) (hundreds of thousands of DALYs lost)
In 1990, 19 per cent of the 50 million annual deaths in the world were caused by respiratory diseases, including lower respiratory infections (8.5 per cent), chronic obstructive pulmonary disease (COPD) (4 per cent), tuberculosis (4 per cent), malignant neoplasms of trachea, bronchus, and lung (2 per cent), all broadly labelled as lung cancer, and asthma (0.3 per cent) and upper respiratory infections (0.1 per cent). After cardiovascular diseases, which accounted for 28 per cent of all deaths, respiratory diseases were the second most common cause of death. If tuberculosis and lung cancer are excluded and reclassified into other categories, respiratory diseases (mainly lower respiratory infections and COPD) would rank third after cardiovascular diseases, infections (including tuberculosis), and parasitic diseases, followed by malignant neoplasms, including lung cancer (Table 1). In 1998, of the 54 million deaths in the World Health Organization (WHO) member states, these respiratory diseases accounted for 16 per cent (tuberculosis, 2.8 per cent; lower respiratory infections, 6.4 per cent; upper respiratory infections, 0.1 per cent; lung cancer, 2.3 per cent; COPD, 4.2 per cent; asthma, 0.3 per cent), and still ranked second to cardiovascular diseases (31 per cent) (WHO 1999a).
While respiratory diseases in general were the second most common cause of deaths in the developed as well as in the developing countries, the relative importance of the specific diseases in each was different. In developed countries, 12 per cent of the 11 million deaths were the result of respiratory diseases: lung cancer (5 per cent), lower respiratory infections (4 per cent), and COPD (3 per cent). Two of the three leading respiratory diseases, lung cancer and COPD, are attributable to smoking. In developing countries, 21 per cent of the 40 million deaths were the result of respiratory diseases, including lower respiratory infections (10 per cent), tuberculosis (5 per cent), COPD (5 per cent), and lung cancer (1 per cent). The two leading causes are infections and the next two are smoking related. In developing countries respiratory disease caused a similar proportion of deaths to cardiovascular diseases (23 per cent), whereas cardiovascular diseases caused 48 per cent of all deaths in developed countries.
There were few gender differences in the mortality pattern in those aged up to 4 years in the developed and developing countries. This was true for respiratory infections and asthma from age 5 onwards. In the older age groups, the excess mortality in males caused by lung cancer and COPD in both developed and developing countries can mostly be attributed to smoking. Whereas the total deaths as a result of these two smoking-related diseases were similar to those caused by tuberculosis in developing countries, they were more than 20 times that caused by tuberculosis in developed countries. Therefore, in developing countries, tuberculosis and smoking-related diseases are of similar importance in causing respiratory deaths, but smoking-related diseases are the predominant causes in developed countries.
World burden of respiratory diseases
The World Bank and the WHO developed the disability-adjusted life year (DALY) to measure the burden of disease (World Bank 1993). The DALY combines healthy life years lost because of premature mortality with those lost as a result of disability. It can be used as a measure of benefits (e.g. the ratio of cost to DALYs gained or the amount needed to gain 1 DALY). The calculation of the DALY is based on the potential years of life lost as a result of a death at a given age, the relative value of a year of healthy life lived at different ages, the discount rate, and the disability weights used to convert life lived with a disability to a common measure with premature death. The updated data (Murray and Lopez 1996, 1997a) are presented in Table 2. Throughout the world in 1990, in males and females combined, 201.4 million DALYs (15 per cent) were the result of respiratory diseases. This was less than all injuries (208.6 million DALYs, 15 per cent), and was followed by neuropsychiatric conditions (11 per cent) and cardiovascular diseases (10 per cent).
In developing countries, respiratory diseases were the most important causes of DALYs lost in females, and the second most important in males (second to all injuries), mainly because of lower respiratory infections and tuberculosis. In developed countries, in each gender, respiratory diseases were less important than neuropsychiatric conditions, cardiovascular diseases, and injuries. Lung cancer, lower respiratory infections, and COPD were the main respiratory causes.
Smoking and respiratory deaths in developed countries
Peto et al. (1994) have provided a comprehensive estimate of mortality from smoking in developed countries from 1950 to 2000. Details of the problem of respiratory mortality attributed to smoking can be found in Hedley and Lam (1997). In 1990, in developed countries, smoking killed more people than all other causes apart from respiratory diseases; of all deaths caused by smoking, about 43 per cent were respiratory diseases. More than half (54 per cent) of all deaths resulting from respiratory diseases in males and females of all ages were attributed to smoking; the proportions in males and females aged 35 to 69 were 83 per cent and 53 per cent respectively.
From 1975 to 1995, in males, over 90 per cent of the deaths caused by lung cancer, about 70 per cent of those caused by COPD and about 14 per cent of those resulting from other respiratory diseases are attributed to smoking. Both the proportion and number of deaths caused by lung cancer are still increasing slowly because the smoking epidemic has already reached its peak in some countries.
The most striking pattern is seen in females. In 1975, about half of lung cancer, 20 per cent of COPD and 2 per cent of other respiratory disease deaths were attributed to smoking, rising to 72 per cent, 53 per cent, and 7 per cent respectively in 1995. Both the proportion and number of deaths from these causes is increasing sharply because of the increasing trend of smoking in women. There is no sign that the smoking epidemic in women has reached its peak.
The lung cancer death rate is the best indicator of the evolving tobacco epidemic in the developed world. In the European Union, there was a small fall in lung cancer deaths in men from 52.4 per 100 000 in 1985 to 1989 to 49.8 per 100 000 in 1990 to 1994; however, but there is a persistent rise in women during the same period, from 8.9 to 9.6 per 100 000. For truncated rates for men aged 35 to 64 years, the peak was 74.3 per 100 000 in 1980 to 1984, followed by a levelling off and a decline to 68.3 per 100 000 in 1990–94; however, in women, the rate increased more rapidly from 7.7 per 100 000 in 1955 to 1959 to 14.3 in 1990 to 1994 (Levi et al. 1999). Female lung cancer rates are still lower in the European Union than in the United States. Breast cancer is still the leading cause of cancer death in women in Europe, but the rate at 35 to 64 years has decreased from 43.0 per 100 000 in 1985 to 1989 to 41.9/100 000 in 1990 to 1994. A more sudden decrease was observed in the United Kingdom (Peto 1998). With the rising trend of lung cancer deaths and decreasing trend of breast cancer deaths, the former will inevitably overtake the position of breast cancer as the top cancer killer in European women early in the twenty-first century. In the United States, from the mid-1980s, lung cancer has killed more women than breast cancer and in Canada this has occurred in the early 1990s with the gap widening.
Both the United Kingdom and United States data show the long delay of several decades, which occurs between the peak of the population smoking prevalence ratios and the peak of lung cancer death rates in men and the benefits that result from the fall in smoking prevalence. Women took up smoking on a large scale much later than men and, therefore, the lung cancer death rates are still rising and the peak of the epidemic will arrive during the next few decades. These patterns of smoking and mortality are now being repeated in developing countries such as China.
Deaths from smoking in developing countries
Smoking killed more people in developed compared with developing countries in the twentieth century but the opposite will be true in the twenty-first century (Peto I>et al. 1994). Estimation of tobacco-attributable mortality in developing countries is more difficult because of the lack of reliable epidemiological data, particularly from large prospective studies. The interaction of smoking and other risk factors for cardiorespiratory diseases, such as cholesterol and air pollution, in developing countries is not clear and should be an important area for further studies. Peto et al. (1994) estimated that smoking caused 2 million deaths per year in developed countries and 1 million deaths per year in developing countries in the 1990s, but the deaths in developing countries cannot be reliably broken down by specific diseases as in developed countries. However, smoking causes many more deaths from other diseases than from lung cancer in both developed and developing countries. In the latter, COPD could be more important than coronary heart disease because of the lower cholesterol levels, but cholesterol levels are likely to increase with Westernization.
Apart from China, data from other developing countries are scanty. A simple, cheap, and quick method is needed to estimate the number of deaths due to tobacco. The case–control study method of collecting a history of smoking in a large number of deceased persons from their surviving relatives has great potential. In South Africa, smoking history has to be recorded on the death certificate (Bradshaw et al. 1998). In Hong Kong, relatives of deceased persons were successfully interviewed in death registries and the response rate was 94 per cent (Lam et al. 1998).
The pattern of smoking and smoking-related mortality, particularly deaths from chronic lung diseases, including tuberculosis, in China is an important warning signal for other developing countries where smoking prevalence is increasing rapidly together with economic developments during the past one to two decades. China has the highest tobacco production and consumption in the world with over 300 million smokers. The main increase in cigarette consumption in China took place only recently and Chinese tobacco mortality will increase substantially if the current smoking pattern persists. Two-thirds of men become smokers before the age of 25 and about half will be killed by tobacco. Of all male deaths at ages 35 to 69, the proportion attributed to tobacco will rise from 12 per cent in 1990 to about 33 per cent in 2030 (Liu et al. 1998; Niu et al. 1998). Worldwide, tobacco will cause about 10 million deaths a year by 2030, 70 per cent of them in the developing world (Lopez 1998). Unless urgent measures are implemented and the increasing trend of smoking in the developing world is rapidly reversed, the predictions for tobacco-induced health problems are gloomy. Tobacco control measures within developing countries are often weak and ineffective against the aggressive tobacco promotion strategies of the multinational tobacco companies. Successful tobacco control measures in the United States have resulted in faster market expansion of American tobacco companies into Asia and other developing countries to maintain their markets. Tobacco control to prevent this huge impending epidemic in developing countries urgently requires the support of an international public health movement.
Recent international and national actions
Internationally, the WHO (1999a) estimated that about 4 million deaths (3.24 million males and 782 000 females) and 49.3 million DALYs (40 million in males and 9.3 in females) in all its member states were attributed to tobacco. The WHO (1998) has published Guidelines for Controlling and Monitoring the Tobacco Epidemic and has established the Tobacco Free Initiative, which is a new WHO cabinet project with the express aim of focusing international attention and resources on the global tobacco epidemic. The WHO has also proposed the Framework Convention on Tobacco Control (WHO 1999b). This is the first time that the WHO has used its treaty-making powers to promote a global public health movement. A recent World Bank (1999) report also marks the first time an international finance institution, after examining the health and economic questions of tobacco use, has supported measures to reduce the demand for tobacco (see Chapter 10.1).
Nationally, the patterns of tobacco use and control measures of the WHO member states were described in a WHO (1997a) global status report. The 1998 United Kingdom White Paper on Tobacco, the first-ever in the country, is a new example of that government’s commitment, with expressed budget allocation and targets, to prevent its annual 120 000 tobacco deaths (SCOTH 1998).
Occupational factors and respiratory diseases
Following smoking, occupational factors are the next most preventable causes of respiratory diseases. Occupational factors consist of many different causal agents, which should be considered individually in different work environments. The causal agents or substances are grouped under ‘occupational factors’ because exposures to them occur predominantly in the work environment. The agents that can cause harm to the respiratory system are classified into dusts (including mineral and organic dusts and particulates), chemicals, biological substances (such as microbiological agents), and physical agents (such as ionizing radiation).
The adverse health effects from occupational factors on the respiratory system range from minor and non-specific symptoms such as cough and phlegm, to fatal and well-defined occupational diseases such as pleural mesothelioma following exposure to asbestos. The effects can also be asymptomatic and recognized only by radiography (such as pleural plaques and fibrosis), lung function tests (such as reduction in forced expiratory volume in 1 s (FEV1), the histamine challenge test (to demonstrate increased bronchial hyper-reactivity) and other laboratory tests such as gas transfer. Occupational lung disorders can present as acute responses (such as acute pneumonitis and pulmonary oedema), as a result of acute gassing (poisoning by inhaled toxic gases), acute allergic inflammation (such as acute allergic aveolitis), chronic diseases with a long latency period (such as lung cancer or silicosis), or acute exacerbation of a chronic condition such as asthma. Asthma can also be caused de novo by an occupational allergen.
In 1999, the International Labour Office estimated that work-related injuries and diseases kill 1.1 million people annually worldwide, 34 per cent due to cancer, 25 per cent injuries, 21 per cent chronic respiratory diseases, 15 per cent cardiovascular diseases, and 5 per cent others (including pneumoconiosis, nervous and renal disorders). Annually, an estimated 160 million new cases of work-related diseases occur worldwide and respiratory diseases are one of the most common. In 1997, the overall economic losses resulting from work-related diseases and injuries were about 4 per cent of the world’s gross national product. Respiratory diseases were responsible for 9 per cent of the costs for work-related injuries and diseases (WHO 1999c). It should be noted that the global burden of occupational disease and injuries is certainly underestimated. Data from developed countries are fragmented, and underdiagnosis and under-reporting is much more serious in developing countries. For example, in Latin America, only 1 to 4 per cent of all occupational diseases are reported.
Non-occupational exposures or agents can cause the same respiratory diseases and therefore the question commonly asked is how much of a respiratory health problem can be attributed to occupational exposure. For compensation purposes, occupational lung diseases as the result of exposure to certain agents or associated with a history of working in certain work processes are usually defined or specified in the regulations, so that workers suffering from such diseases are assumed to have derived their diseases solely (or mainly) from their work and are eligible for compensation. There is no requirement to prove that their diseases could not have derived from exposures outside work, even though non-occupational exposure, such as smoking, could be a more important or plausible cause.
Many occupational agents and processes are recognized as occupational causes of lung cancer. Smoking is an independent cause of lung cancer in workers and it can interact with other occupational carcinogens. When there is an interaction of two causal agents, removal of one agent will take away the effect of that agent and the interaction effect of both factors. For example, for a smoker exposed to asbestos, cessation of smoking alone will result in an elimination of 92 per cent of the excess rate, and removal of asbestos alone will result in an elimination of 81 per cent of the excess risk (Harrington and Saracci 1994).
Occupational health practices tend to emphasize occupational carcinogens and ignore smoking, as the latter is not traditionally considered as an occupational factor. The best preventive approach should aim at the removal of both smoking and the known occupational carcinogens. However, measures to remove a carcinogen in the workplace can be difficult, expensive, or even impracticable, for example in developing countries where the worksite situations are the worst. Prevention of smoking, for example by banning smoking and by smoking cessation services, can be more readily implemented and can prevent a large proportion of lung cancer in the workforce, regardless of whether there are exposures to known or unknown occupational lung carcinogens.
Other important respiratory diseases attributable to occupational factors include pneumoconiosis, such as silicosis and asbestosis, occupational asthma, and diseases or syndromes from exposure to organic dusts, such as byssinosis. Workers can present with a whole spectrum of non-specific respiratory symptoms. Whereas chronic non-specific respiratory diseases, including chronic bronchitis, emphysema, and asthma may be common in certain occupational groups exposed to dust or agents inhaled at work, among smokers smoking can be a more important cause than the specific dust or chemicals. Few studies can clarify the independent roles of smoking and a specific dust or chemical and the interaction between the two in cases of pneumoconiosis because most workers affected are smokers. For example, most patients with silicosis are chronic heavy smokers and their lung function shows an obstructive pattern of chronic bronchitis as well as the restrictive pattern due to silica deposition. When the workers claim for compensation the regulations ignore the role of smoking and both the affected workers and their physicians are reluctant to highlight smoking as a probable cause of the health problems for fear of jeopardizing compensation and other claims. Many such patients continue to smoke after receiving compensation and their condition can deteriorate rapidly. Obviously compensation does not improve health per se, and pneumoconiosis patients should receive active pulmonary rehabilitation, even if the disease is not curable. Apart from removal from further exposure to the putative occupational agent, smoking cessation should be considered the most important method to improve health and prevent further disability and premature death.
The extent of occupational respiratory disorders is often underestimated by official statistics based on statutory notification or compensation. Special surveys are needed and in the United Kingdom, the Surveillance of Work Related and Occupational Respiratory Diseases project is one good example (Sallie et al. 1994).
A new approach is to measure the prevalence of self-reported work-related illness by interviewing a representative sample of the working population. In 1995 in the United Kingdom about 40 000 persons who had ever worked were interviewed in a Labour Force Survey and 7 per cent reported an illness in the last year that was caused or made worse by work (Jones et al. 1998). Of the estimated prevalence of 2 million illnesses, lower respiratory disease (mainly asthmatic symptoms) constituted about 10 per cent and ranked third among the most common self-reported work-related disease categories, following musculoskeletal conditions (60 per cent) and stress (25 per cent), whereas pneumoconiosis constituted about 1 per cent. Such an approach based on self-reporting has many limitations but comparison of the results in 1995 with those in a similar survey in 1990 suggests good reliability. It is the only cost-effective method to give a comprehensive estimate of the extent of work-related respiratory problems in the working population as a whole. For an individual worker, an occupational cause claimed by the worker is a good start as this indicates awareness and concern and should prompt the health professionals concerned to discuss the problem and to examine carefully the potential or claimed health risks at work. From a public health perspective, such estimates of occupational health problems in the total population are useful for monitoring the problems and for guiding occupational health promotion strategies.
The principles of primary prevention of occupational respiratory disorders have been discussed by Hedley and Lam (1997) and are not repeated here. Although it is still not clear what would be the interaction effect between smoking and occupational exposure to pneumoconiosis and other specific and non-specific occupational respiratory disorders, the principles and benefits described above for the prevention of lung cancer should apply. Prevention of smoking, including provision of smoking cessation programmes, should be an important part of an occupational health programme to prevent occupational respiratory disorders, including cancer and other health problems.
Since 1997, the International Labour Office and the WHO have jointly started an initiative for the global elimination of silicosis. Elimination is not eradication (as in the case of smallpox) and it implies the possibility of recurrence (Wagner 1998). However, this is the first time that an occupational disease has been targeted for elimination. In more developed countries such as the United States, elimination is a possibility, but in developing countries the situation is likely to get worse before it gets better. The initiative, however, can remind governments, employers, employees, and occupational health professionals that silicosis is preventable and the ultimate target of elimination should be achievable in the not too distant future.
Respiratory health and passive smoking
Passive smoking is the inhalation by non-smoking subjects of the cigarette smoke emitted from the burning end of the cigarette (side-stream smoke) and the smoke inhaled (mainstream smoke) and then exhaled by the active smoker. This mix of tobacco smoke is called environmental tobacco smoke (ETS) or second-hand smoke (SHS). Concentrations of some carcinogenic and toxic substances are higher in side-stream than mainstream smoke, although ETS is diluted by ambient air.
In the late 1980s, two major reviews in the United States (National Research Council 1986; US Department of Health, Education and Welfare 1986) and one from the United Kingdom (Froggatt 1988) all concluded that ETS causes lung cancer and childhood respiratory illnesses. The conclusions of a comprehensive review from the United States Environmental Protection Agency (USEPA 1992) were the same. The relative risk for lung cancer due to ETS, based on 11 United States studies, was 1.19 (USEPA 1992). The excess risk estimated by the Froggatt Committee (Froggatt 1988) was about 10 to 30 per cent. In a meta-analysis of 37 studies Hackshaw et al. (1997) produced a relative risk of 1.24 (95 per cent confidence intervals 1.13 to 1.36). The excess risks were not affected after adjusting for bias and confounding, including confounding due to diet. Their conclusion on lung cancer was accepted by the United Kingdom Scientific Committee on Tobacco and Health (SCOTH 1998), which also reached similar conclusions about the other adverse health effects of ETS, including heart disease and childhood respiratory ill health.
Despite the overwhelming amount of evidence and the unanimous conclusions of many reviews by committees and individual authors that the association between ETS and lung cancer is causal, objections from the tobacco industry and its scientists are still prevalent. A tobacco industry sponsored European Working Group (1996) found a significantly increased risk in its meta-analysis on ETS and lung cancer but then discredited this finding. Their arguments, including misclassification bias, dietary confounding, publication bias, and thresholds were criticized by Smith and Phillips (1996). A recent example is the tobacco industry’s misinterpretation of the evidence from a WHO study on the links between passive smoking and lung cancer (Boffetta et al. 1998) and this led the WHO to denounce newspaper reports based on the industry’s analysis as ‘false and misleading’ (Bates and Brookes 1999). Barnes and Bero (1998) examined 106 review papers on ETS and concluded that the affiliation with the tobacco industry was the only factor associated with authors concluding that ETS is not harmful.
In children, many reviews by government agencies or committees, non-governmental organizations and individual scientists have concluded that ETS causes respiratory ill health (Royal College of Physicians of London 1992; USEPA 1992; Australian National Health and Medical Research Council 1997; SCOTH 1998). Compared with the issue of ETS and lung cancer, the conclusions about ETS and child health have been subject to much less critique from the tobacco industry and its scientists. On an international level, the 1997 Declaration of the Environmental Leaders of the Eight (G8) on Children’s Environmental Health stated that children exposed to ETS are more likely to suffer from reduced lung function, lower respiratory tract infections, and respiratory irritations, that asthmatic children are especially at risk and that many of those symptoms lead to increased hospital admission (WHO 1999d). This is the first time that ministers of the environment were involved as a group in the passive smoking issue. In response to this declaration, the WHO convened an International Consultation on Environmental Tobacco Smoke and Child Health in January 1999.
The WHO consultation has concluded that ETS is a real and substantial threat to child health and that ETS causes lower respiratory tract infections such as pneumonia and bronchitis, coughing and wheezing, worsening of asthma, and middle ear disease (WHO 1999d). About 700 million, or almost half of the world’s children are exposed. For lower respiratory illness, during the first year of life, in children whose mothers smoke, the pooled relative risk from over 40 studies was 1.7 (95 per cent confidence intervals 1.6 to 1.9); for paternal smoking alone, the relative risk was 1.3 (95 per cent confidence intervals 1.2 to 1.4). Both asthma and respiratory symptoms are increased in children whose parents smoke, and the pooled relative risks from over 60 studies, for either parent smoking, range from 1.2 to 1.4. The California Environmental Protection Agency (1997) estimated that, each year in the United States, ETS exposure causes: 3000 deaths from lung cancer; 35 000 to 62 000 deaths from ischaemic heart disease; 1900 to 2700 cases of sudden infant death syndrome; 9700 to 18 600 cases of low-birth weight infants; 8000 to 26 000 new cases of asthma in children; exacerbation of asthma in 400 000 to 1 million children; and 150 000 to 300 000 cases of bronchitis or pneumonia children aged 18 months and younger, of which 7500 to 15 000 require hospital admission (Davis 1998).
Passive smoking at work
In the United States EPA (1992) review, no clear conclusion was reached about the adverse health effects of ETS exposure at work. The conclusions about ETS and coronary heart disease from most reviews were based on results predominantly from exposures at home. As for exposure at work, He et al. (1994) first reported an association between coronary heart disease and passive smoking at work in Chinese never-smoking women, with a dose–response relationship after adjusting for major risk factors and exposure to passive smoking from the husband. A review of seven studies on passive smoking in the workplace by Wells (1998) concluded that the relative risks for heart disease from passive smoking at work are roughly equal to those from home-based exposures. A meta-analysis by He et al. (1999) calculated a relative risk of 1.11 (95 per cent confidence intervals 1.00 to 1.23) for exposure in the workplace.
Although many reviews have concluded that ETS is causally related to lung cancer, few have made specific conclusions about risk estimates for passive smoking at work. The United States Occupational Safety and Health Administration (OSHA 1994) assumed that risk estimates based on residential exposures should accurately reflect occupational risks for most workplaces. Siegel et al. (1995) reviewed 10 studies on lung cancer and non-smokers exposed to ETS in the workplace and concluded that these studies support the assumption of a small elevated risk. Using modelling based on nicotine from ETS in office air and salivary cotinine in non-smoking workers, Repace et al. (1998) estimated that at the current 28 per cent prevalence of unrestricted smoking in the office workplace, 4000 heart disease deaths and 400 lung cancer deaths occur annually among office workers from passive smoking in the workplace in the United States.
Apart from lung cancer, evidence on the relationship between ETS and respiratory ill health in adults is scarce, in contrast with the huge amount of such evidence in children. In the risk assessment of passive smoking at work, the United Kingdom Health and Safety Commission (HSC 1999) Proposal for an Approved Code of Practice is only based on the two main effects of passive smoking exposure in the workplace: (a) it exacerbates certain diseases, such as asthma and chronic bronchitis; (b) it causes discomfort as it irritates the eyes, nose, throat, and chest, and tobacco smoke has an unpleasant smell. The HSC was not certain about the size and extent of the risk for either cancer or heart disease.
Only a few studies have shown an association between respiratory ill health (other than lung cancer) and ETS exposure in adults. Two of them observed an association for ETS exposure at home in women (Ng et al. 1993; Pope and Xu 1993). Wong et al. (1999) found a dose–response relationship between respiratory symptoms in non-smoking women and ETS exposure at home after adjusting for outdoor air pollution. White et al. (1991) reported that those with ETS exposure at work had more respiratory symptoms but no dose–response relationship was found. Leuenberger et al. (1994) reported a significant trend of increasing risk with total ETS exposure at home and at work, but there were no separate data for exposure at work. Although Jaakkola et al. (1996) measured ETS at home and at work separately, only the relationship between total exposure and respiratory symptoms was reported.
Smokers are obviously also exposed to passive smoking. Mannino et al. (1997) reported that in the United States, 87 per cent of current smokers reported exposure to ETS at home or work, whereas only 20 per cent never-smokers and 23 per cent former smokers were exposed. There are limited data for ETS exposure at home and lung cancer in smokers. Siegel et al. (1995) conducted a meta-analysis of six studies that examined lung cancer among smokers (men and women) exposed to ETS at home and the pooled relative risk was 1.3 (95 per cent confidence intervals 1.1 to 1.5). This was consistent with the value of 1.25 in the United States EPA (1992) meta-analysis of seven studies among female smokers exposed to ETS. There is a lack of data on exposure to ETS at work in smokers and no definitive conclusion on the risk of cancer in this group.
In a study on the Hong Kong Police, Lam et al. (1999) compared the prevalence of having frequent cough or phlegm among four groups of male officers and the odds ratio adjusted for age, education, marital status, police rank and type of police work, and past occupational dust exposure was 1.00 for never-smokers not exposed to ETS, 1.91 (1.58 to 2.30) for never-smokers exposed to ETS, 2.44 (1.93 to 3.07) for current smokers not exposed to ETS and 4.30 (3.57 to 5.18) for current smokers exposed to ETS (p < 0.001). The results were similar after further adjustment for ETS exposure at home.
Only one study has examined the effect on respiratory health before and after legislative prohibition of smoking in the workplace. Eisner et al. (1998) found that bar tenders’ respiratory symptoms, forced vital capacity (FVC) and FEV1 improved after a decline of self-reported ETS exposure as a result of the smoking ban in California. Stratification by smoking status showed similar results in non-smokers and smokers. The results were also controlled for recent upper respiratory infections. This study has provided important evidence that passive smoking is harmful and that banning smoking can produce immediate benefits on the respiratory health of workers exposed at work.
More research on the effects of ETS exposure, particularly at work in smokers will certainly help to clarify the issue and to support legislative control of smoking in the workplace. The results so far are sufficient to conclude that ETS at work causes ill health in both non-smokers and smokers. Whereas smokers may be convinced that breathing other smokers’ smoke is harmful to themselves, the tobacco industry is certain to challenge any legislative restriction of smoking at work and in public places. However, if both non-smoking and smoking employees support restriction or total bans on smoking in their own workplaces, they and their employers can make their own decision to implement control measures to eliminate ETS exposure. Such measures have been adopted in many government and private workplaces in developed countries but progress in developing countries is slow. The only satisfactory smoking areas designed to protect non-smokers are separately ventilated and exhausted areas, but it is likely these areas will be a hazard to smokers. As a comprehensive smoke-free policy always produces better outcomes (Siegel 1995), this is the preferred approach. Strategies to protect non-smokers include education, regulation, legislation, and litigation, but a large number of people continue to be exposed involuntarily to ETS. Education can increase public support for legislation to ban or restrict smoking; however, education alone is usually not effective and legislation is needed. Litigation is being used increasingly in the United States to compensate non-smokers whose health has been adversely affected (Davis 1998) and there have been some well publicized successful cases in the United States, United Kingdom, and Australia. However, such moves are almost unheard of in developing countries.
The recognition of passive smoking as a health hazard is an important milestone in tobacco control in the past two decades in the West, particularly in the United States. The United States OSHA was the first government occupational health agency to propose regulations to ban smoking in work sites except in separately ventilated areas.
A quasi-legal approach using an Approved Code of Practice was proposed in 1999 by the United Kingdom Health and Safety Commission (UKHSC 1999) on passive smoking at work. The code will clarify how the Health and Safety At Work Act of 1974 should be applied to passive smoking as the Act requires ‘every employer to ensure, so far as is reasonably practicable, the health, safety and welfare of all his employees’ (Bates and Brookes 1999). The Code of Practice does not require a ban on all smoking in all workplaces—only where it is reasonably practicable. It is not aimed to protect customers of restaurants and pubs (or bars), but employers have to do what is reasonably practicable to reduce employees’ exposure if they do not ban smoking. If the proposal is approved, the United Kingdom approach would be a useful example for other countries to follow, if legislation for a total ban is not feasible.
Banning smoking in the workplace will encourage smokers to quit smoking. Chapman et al. (1999) reviewed 19 studies of the impact of smoke-free workplaces from 1986 to 1996 and showed that 18 studies reported reduction of daily smoking rates and 17 reported reduction of smoking prevalence. The tobacco industry also estimated a big reduction of consumption and a big increase in quitting rate as a result of workplace smoking restriction (Bates and Brookes 1999). Prevention of smoking and passive smoking at work should be an urgent priority in occupational health and public health.
Passive smoking in Asia
The ETS issue is a major problem for the tobacco industry and its efforts to confuse the evidence are well documented (Glantz et al. 1996) and will continue to be uncovered. Action against ETS is probably the greatest threat to the expansion of the multinational tobacco companies in Asia and they have gone to great efforts to block it. Whereas the recognition of ETS as a health hazard has prompted strong public health interventions in the West, particularly in the United States, great inertia and barriers to the prevention of passive smoking are common in developing countries. Although scientists in Asia have made significant contributions to the global evidence to support the causal conclusions about ETS, their local evidence has not yet made significant impacts on governments. Most governments, including their environmental and occupational health departments do not consider ETS a high priority (Lam and Hedley 1999). There is slow progress against great barriers and no sign of a major breakthrough in the developing world. Public health workers in developing countries should use the ETS issue, in high profile campaigns, to promote campaigns to protect the health of non-smokers, particularly women and children, and experience and support from the West would be beneficial.
Smoking, passive smoking, and occupational health
Workers who are most exposed to occupational hazards usually come from the lowest socio-economic strata and they have the highest smoking prevalence. This is true in developed countries such as the United States (Stellman et al. 1988) and in developed countries such as China (Lam et al. 1996). Such workers’ excess morbidity and mortality from both exposures is reflected in the highest standardized mortality ratios in the lowest social class in developed countries, such as the United Kingdom, and is the major contributor of poor health in a much larger sector of the working population in developing countries. Occupational health professionals should work closely with tobacco control advocates, target the most vulnerable groups at risk and make the prevention of smoking a top priority. On the other hand many occupational health professionals have been targeted for recruitment by the tobacco industry and become the industry’s consultants to speak against the evidence on passive smoking and against tobacco control measures. Such recruitments in Asia have been revealed in internal documents of the industry (Rupp and Billings 1990). Occupational health professionals should be aware of such strategies of the tobacco industry and avoid being inadvertently recruited as the industry’s consultants.
The International Commission on Occupational Health (1999) is the most important international organization for occupational health professionals. It has 30 scientific committees and four scientific working groups as of 1 March 1999, but there is none on smoking or passive smoking at work. Under the auspices of the Commission, the International Congress on Occupational Health is the most important international occupational health conference, but smoking and passive smoking have never been given any prime status as topics in plenary sessions, mini-symposia, or other scientific sessions. Unless occupational health professionals change their attitudes about smoking and passive smoking at work and start to tackle these problems seriously, many workers in the new millennium will continue to work in unhealthy smoky workplaces and their health will be seriously affected by both active and passive smoking.
Chronic obstructive pulmonary disease
COPD (ICD9 490–496, excluding 493) is a group of related or overlapping conditions that comprises chronic bronchitis, emphysema, bronchiectasis, bronchial catarrh, and other non-specific obstructive airway diseases. Under different definitions it may be referred to as chronic obstructive lung disease. The prevalence of chronic bronchitis is increasing as the populations of postindustrialized countries age. It may be considerably underdiagnosed in some populations (den Otter et al. 1998). Most (75 per cent) mortality from respiratory disease is attributable to COPD, which is the fifth most common cause of death after infectious disease (including respiratory infections), cardiovascular, and malignant and traumatic conditions (WHO 1999a).
In the United States, where it is the fourth most common cause of death, COPD is uncommon in those under 40 years of age but the prevalence reaches 10 per cent in 60- to 85-year-olds (Anthonisen 1997). In 1996 it was estimated to be the principal cause of mortality in 100 000 deaths and contributory in many more (Petty and Weinmann 1997).
The strongest risk factor for COPD is smoking (Fletcher et al. 1976). It increases the rate of decline of lung function (Camilli et al. 1987) and increases the risk of symptoms and premature death (Kuller et al. 1989). The global patterns of COPD strongly reflect geographic and demographic patterns of smoking. Future patterns of COPD will be determined by smoking trends and gender differences in rates of increase can be expected in view of the increased prevalence of smoking in women. Among WHO member states the global burden of COPD is estimated at 28.6 million cases, about half of all respiratory diseases, with more than 90 per cent occurring in lower and middle income groups and nearly 60 per cent in China alone (WHO 1999a).
In the United Kingdom the pattern of increase and decline in COPD is associated with increases and later declines in smoking in men in the twentieth century (ONS 1997). Other potential causes of chronic respiratory illness and decreased lung function in adult life include birth weight, serious childhood infections, occupational exposures, and ambient air pollution.
Only a minority of all smokers develop COPD (Fletcher et al. 1976) and family studies suggest that genetic factors also contribute to its development (Sandford et al. 1997). The only established genetic risk factor is homozygosity for the 2 allele of the a1-antitrypsin gene. Heterozygotes may also be at increased risk. Other genes, including those for a1-antichymotrypsin and cystic fibrosis membrane regulator and blood groups, and several others have shown a significant association between polymorphisms and COPD (Sandford et al. 1997).
The epidemiological evidence for a causal relationship between nutrition and either adverse effects for or protection from asthma and COPD has focused on intake of sodium, n-3 fatty acids, antioxidant vitamins, and fresh fruit and vegetables (Smit et al. 1999). Although the evidence for a beneficial effect of consuming fish, fruit, and vegetables on both asthma and COPD is increasing the effect of supplementation in subjects with existing asthma remains uncertain. High sodium diets may increase airways reactivity in asthmatics.
The health-care costs of COPD are high with several cost drivers, including frequent hospital admission and long-term oxygen therapy. Other economic costs include years of life lost, disability, and reduced quality of life. In primary care settings in the United Kingdom consultations for COPD are up to four times as frequent as for ischaemic heart disease. In the United States COPD is the third most common cause of hospital admission.
Treatment and rehabilitation may improve both clinical aspects of the disease and quality of life and reduce the overall costs of care. Long-term oxygen therapy is an intervention that improves survival and quality of life in COPD (British Thoracic Society 1997). In patients with acute exacerbations of chronic bronchitis in COPD the use of third line antibiotics increased pharmacy costs but showed a trend towards lower mean total costs (Destache et al. 1999). The improved outcomes included reduced failure rate, lower need for hospital admission, and increased remission times. The distress caused by dyspnoea may be associated with a higher prevalence of depression (van Ede et al. 1999). Pulmonary rehabilitation may confer substantial benefits in COPD, including improved walking distance and maximal exercise capacity. However, all interventions are likely to be ineffective without smoking cessation and the provision of quitting support services is of paramount public health importance.
Adverse health effects of air pollution
Air quality and sustainable development
The global effects of environmental degradation are now a major public health concern, including the adverse effects of poor-quality indoor and outdoor air. A recent report on health and environment in sustainable development concludes that when assessed in terms of DALYs, environmental factors are associated with almost a quarter of the global burden of disease (WHO 1997b). A large proportion of this is attributable to different forms of air pollution. Air pollution is now a problem that affects most human settlements across the planet. In the nineteenth century it typically affected expanding urban areas that arose from the industrial revolution. In the West the principal cause of air pollution in the past has usually been coal, wherever it has been used. Legislation and controls led to cleaner air in many Western industrialized countries, but twentieth-century urban development and the internal combustion engine have been associated with deterioration in air quality. This is now recognized as a hazard for both acute and chronic adverse health effects worldwide.
In many wealthy countries levels of sulphur dioxide and particulates have fallen following clean air legislation and changes in the types of fuels used. On the other hand, concentrations of nitrogen oxides, ozone, ultrafine particulates, and carbon monoxide have risen along with a massive increase in road traffic, particularly diesel-powered vehicles. In developing countries, such as China, both sources of pollution are now present and largely uncontrolled. The growth in vehicular road traffic in Asia must be regarded as still in its early stages so the contribution to pollution from mobile sources will predictably continue to increase for many years.
In addition to escalating numbers of badly adjusted diesel engines the burning of unwashed coal and other high sulphur content fuels has placed many cities, for example those in the Asia Pacific region, among the most polluted in the world. Daily concentrations of pollutants in some cities are similar to those observed in the pollution episodes of London and New York 50 years ago (McMichael and Smith 1999). The London smog was largely generated by domestic coal fires, but in recent years evidence has emerged on the importance in developing countries of indoor pollution from cooking and heating fuels. The combined effects of poor air quality will continue to be a major cause of premature death in developing countries.
In the 1990s the problem of haze in Asia Pacific countries has received international attention. In Malaysia, in the Kelang Valley, satellite imaging reveals haze in association with highways, factories, new housing developments, and agricultural burning. Analysis also suggests that photochemical reactions contribute to the build up of haze in some areas (Samah 1992). Particulate concentrations of over 400 µg/m3 are typical in haze episodes. In Sumatra, Indonesia, wilful burning of rain forest and scrub in the late 1990s led to disastrous haze episodes lasting for months during the dry season and permeating over large areas of the region, including Singapore and Malaysia. Thousands of additional cases of respiratory and eye problems have been reported, but the overall impact on health, particularly of infants and children, is largely undocumented.
A wide spectrum of air pollutants are implicated in adverse health effects, although these effects may vary and there is still a lack of evidence and consensus on causal relationships. The principal pollutants of interest can be classified as particulates (either finely divided solids or finely dispersed liquid aerosols), gases, and toxic chemicals such as hydrocarbons.
Comprehensive reviews of particulate and other forms of pollution are given by Holland et al. (1979), the American Thoracic Society (1996), and Holgate et al. (1999). The characteristics of the four criteria pollutants, respirable suspended particulates, sulphur dioxide, nitrogen dioxide, and ozone together with carbon monoxide, polycyclic hydrocarbons, and volatile organic compounds are summarized by Hedley and Lam (1997).
Factors affecting pollutant levels and exposures
Ambient pollutant levels and the exposure of individuals are affected by rainfall, wind speeds and currents, atmospheric inversion, and human activities. Low wind velocities hinder horizontal dispersion and high pressure and temperature inversions limit vertical dispersion. Local geographical features are also a contributor. Weather patterns result in elevated levels of pollutants because of reduced rain and wind and increased build up of inversion layers in the atmosphere. However, rainfall may increase the exposure of some individuals, because of pollutant laden aerosols resulting from the scouring of the air by raindrops. There are marked seasonal variations in ambient pollutant levels that result from climatic changes. In many Western countries ozone levels are higher in the warm season when mean sunlight levels are highest; however, in tropical or subtropical climates such as Hong Kong ozone levels are highest in the cool dry season when cloud cover is lower (Fig. 1).
Fig. 1 Hong Kong: smoothed monthly average of air pollutant concentrations.
Assessment of health effects
The current problems in assessment of outdoor air pollution effects on health, and therefore the extent to which epidemiology can support policy-making, can be summarized as follows.
Patterns of pollution are heterogeneous between different sites and they cannot be assumed to have the same effects. Examples include the ‘reducing’ forms of industrial winter smog comprising particulates and sulphur dioxide and ‘oxidizing’ forms with nitrogen oxides and hydrocarbons and products of photochemical decomposition.
Exposure prevalence probably varies markedly within and between different communities and countries.
Data from monitoring are strongly influenced by meteorological conditions, including dispersion of pollutants, their concentrations and, therefore, exposures. The interpretation of relationships between pollution and health is confounded because both pollutants and health indices (morbidity, mortality, and health-care utilization) show marked seasonal patterns.
The causal relationships between adverse health effects and exposure to specific pollutants or mixtures remain uncertain.
Socio-economic factors may confound the interpretation of studies, such as those based on mixed residential and industrial communities, educational attainment, smoking, environmental tobacco smoke, and other indoor pollutants. Housing and nutrition may influence health and confound the interpretation of exposure studies.
If poor health or health hazards, or conversely higher income and socio-economic status, are factors in the movement of people away from adverse environments then this will lead to a change in the population of residents and their estimated risks. Such patterns may be difficult to recognize or take account of in ecological studies that involve two or more population groups, but are not based on individual exposures and health records.
Health effects of pollutants
The identification of environmental hazards and their subsequent elimination through social and economic policy, to protect the health of the community, is dependent on the establishment of valid evidence. The environmental effects of air pollution have been studied in plants, animals, and humans, and through laboratory-based in vitro methods. The aims of epidemiological research are as follows (Department of Health Committee on Medical Effects of Air Pollution 1998):
to identify the types and concentrations of pollutants
to assess the probable exposures of the general population to these levels of pollution
to define the adverse health effects
to quantify the risk of the health effect if applied to the whole population.
The possible adverse health effects of air pollution on a population can be represented in the form of a pyramid (Fig. 2).
Fig. 2 The population pyramid of air pollution health effects.
There are three general groups of evidence from scientific studies.
Time series: acute effects
Cohort studies: chronic efects
Surveys of high-risk occupational groups
For example, controlled laboratory experimental work has demonstrated that specific pollutants (i.e. single pollutant exposures) may be associated with inflammatory reactions in tissues; in human volunteers changes in lung function and the development of symptoms such as cough and chest pain have been observed. In animals, genetic factors, genotoxicity, and carcinogenesis can be examined. In humans, controlled studies in volunteers have examined the effects of age, gender, ethnicity, and predisposing existing health problems such as asthma. Special susceptibility to pollutant exposures may be found both in well individuals and those with health problems such as asthma.
The evidence to support public health advocacy for clean air can be drawn from panel studies of patients and epidemiological cross-sectional surveys, and case–control and cohort studies, including cohort-based nested case–control studies. Panel studies are prospective cohort studies but the subjects are not necessarily free of respiratory problems at the start. Usually enquiries form a series of cross-sectional arrays from multiple assessments of the subjects at different points in time. There are relatively few case–control and long-term cohort studies on the health effects of air pollution. Many studies are based on ecological designs in which the health experience of whole populations is examined using routinely collected morbidity and mortality data together with pollutant concentrations. The two types of ecological studies commonly reported are cross-sectional surveys of two or more groups and time-series studies, which use data on daily, weekly, or monthly events such as deaths or hospital admissions. The limitations of these and other epidemiological approaches are summarized by Samet and Jaakkola (1999) and include the so-called ecological fallacy (Last 1995).
Time-series analyses are usually based on large pre-existing databases of, respectively, pollutant concentrations, health-care utilization (such as hospital admissions), and mortality derived from registrations of deaths. These databases are used to estimate variations in daily events such as admissions and deaths within a single population. The advantages of this approach are that it is feasible and relatively inexpensive and time-series analysis can show how short-term distributions of deaths are related to short-term variations in concentrations of pollutants. Therefore it can provide a rapid method of generating evidence on a possible causal relationship between short-term health effects and pollution. Longitudinal time-series analyses, which follow single or multiple populations over time, have several advantages over other types in that they avoid the need to control for common individual-level confounding factors, such as educational attainment, socio-economic status, and smoking prevalence, because these factors are more or less constant from day to day and the population acts as its own control. For example, the frequency of health events in the population being studied on high pollution days is compared with that on lower pollution days (Thurston and Ito 1999). Because they include all members of a population, including those most susceptible to pollution, the results may be the most relevant and generalizable. Their limitations (Thurston and Kinney 1995) include the previously mentioned problem of shared long-term cycles, such as the seasonal variation in both health events and pollutant concentrations.
Daily time-series mortality studies in single populations are now frequently employed by researchers in many countries. McMichael et al. (1998), however, have pointed out that time-series analyses, as they are usually performed, cannot distinguish between life shortening due to deaths brought forward (e.g. by days or weeks) by acute effects of pollution in individuals with serious advanced chronic disease and deaths from chronic disease induced by long-term exposure to air pollution. The phenomenon of deaths brought forward, for example due to an episode of pollution, is referred to as harvesting. If this occurs then fewer people would die over the period following the episode. The possible effect of harvesting on mortality estimates remains uncertain. Further studies on variation in daily deaths may help to separate out the acute and chronic effects of pollution. In contrast these conceptual problems may not apply to other non-fatal, outcomes such as hospital admissions and time-series data can contribute to the assessment of health, quality of life, and utilization of services. A schema to represent the possible distribution of deaths from respiratory and cardiovascular disorders due to air pollution has been suggested by the United Kingdom Committee on Medical Effects of Air Pollution (Fig. 3).
Fig. 3 Relationship between deaths advanced by exposure to particles and deaths from illness induced by exposure to particles.
Chronic health effects, in terms of incidence and long-term death rates, can best be estimated from long-term cohort studies, in which a large defined group of subjects is followed forward in time with the collection of the best possible information on local pollutant concentrations and individual exposures. Such studies obviously need to be managed over many years; they are inevitably costly and demanding in terms of personnel, expertise, and other facilities. There may also be questions about the generalizability of the findings to other settings because of differences in both prevalent pollutant types and individual exposures. In public health terms the effects of long-term exposure to air pollution may be more important than that due to brief episodes, for example in terms of person-years of life lost.
In general, although laboratory studies and evidence from occupational exposures in certain types of process workers do provide indications of the toxicity of specific pollutants they do not necessarily allow us to extrapolate the results to estimate the impact of breathing polluted air by the general public. This is at least partly because the different components of general air pollution are more numerous and complex than the pollutants in experimental or occupational exposure. On the other hand even in studies of human populations this complexity of the urban pollutant mixture means that we may not be able to ascribe confidently causality to the observed association between health events, such as admissions and deaths, and variations in concentrations of single specific pollutants such as sulphur dioxide, nitrogen dioxide, respirable suspended particulates (RSPs), and ozone. All we may be able to conclude safely is that certain general levels of pollutants of all types and, therefore, their sources are likely to be injurious to health.
Air pollution may cause concern because it degrades the environment and is perceived to impair quality of life, not least because visibility, between ourselves and the horizon, is impaired and the general nature of the environment deteriorates. The visibility index in many Asian cities has been deteriorating markedly during the past 5 years. As judged by letters to newspapers, media reports, and other surveys this undoubtedly has an adverse psychological effect on the community as a whole. Pollutants may cause relatively minor health problems, such as eye or skin irritation, whereas the serious effects of air pollution are associated with lung inflammation and reduced lung function, critical cardiovascular events, and premature death.
The American Thoracic Society (1985) put forward guidelines, on the definition of adverse effects, covering a wide spectrum of measurable or potential adverse health outcomes. They can be summarized on a five-point scale.
Irritation and infections that do not interfere with the normal activity of the affected person.
Episodes of respiratory illness that require medical attention.
Impairment of respiratory function and failure to thrive in children.
Increased exacerbations of illness in people with chronic cardiopulmonary disease.
Increased incidence of cancer and premature death.
The American Thoracic Society guidelines attach a value judgement to the observed effect so that some may be classified as ‘trivial’ or ‘not medically important’. However, from both a medical and ethical viewpoint it could be argued that it is not justifiable to attach arbitrary values to any adverse effects. This is especially the case if the exposures are involuntary and the impact on quality of life and activities of daily living have not been measured. In terms of health, the wider use of methodology capable of detecting subclinical physiological changes, such as bronchial hyper-reactivity or blood viscosity will continue to redefine levels of impairment. Interventions and long-term epidemiological studies will be needed to determine whether they are associated with permanent harm to the respiratory and cardiovascular systems and premature death.
The measures used to estimate the biological damage of pollution and development of clinical health problems are summarized in Table 3. Laboratory and occupational studies show that air pollution could be harmful, whereas population epidemiological studies are needed to show that pollution is harmful to the health of the public.
Table 3 Health effects and biological markers of response associated with air pollutiona
Mortality from air pollution
The public health effects of environmental air pollution and its changing patterns have been observed worldwide. In the London pollution episode of 1952 the association of high levels of particulates (British Smoke) and sulphur dioxide was associated with 4000 excess deaths (Logan 1953). In a similar episode in 1962 there were only 350 excess deaths (Scott 1963). The difference in pollutants between the two episodes was mainly an 80 per cent reduction in the particulates; however, since then further work has implicated the acid component of particulates.
In addition to the observations on the pattern of mortality in the London smog, there is also the evidence from the winter hazes of New York (1960–1964), total particulate levels in Steubenville, Ohio (1974–1984) and Pennsylvania (1973–1980), and haze in California (1980–1986). Dockery et al. (1993) demonstrated the importance of long-term follow-up studies to identify health effects. After adjustment for smoking, occupational exposure, and other factors, the excess risk derived from a mortality rate ratio between cities with the highest and lowest pollution levels in 8000 adults followed for 14 to 16 years in six cities was 26 per cent. Mortality was most strongly associated with levels of fine particulates, including sulphates.
In a pooled analysis (meta-analysis) of nine United States cities and London an increase in PM10 of 10 µg/m3 was associated with a 3.4 per cent increase in respiratory mortality and 1.8 per cent for cardiovascular deaths (American Thoracic Society 1996). In Birmingham, Alabama, Schwartz (1993) estimated that population exposures to a rise of 100 µg/m3 are associated with an 11 per cent excess risk of overall mortality (relative risk is equal to 1.11; 95 per cent confidence intervals 1.02 to 1.20). Because cardiovascular deaths are very common they make the biggest contribution to the overall association between pollution and total mortality.
Positive associations between individual air pollutants and mortality have been demonstrated in time-series analyses in many American, European, and other country studies. In 12 European cities, pooled estimates of the increase in daily mortality per 50 µg/m3 increases in concentrations of sulphur dioxide and particulates were 3 per cent and 2 per cent respectively (Katsouannyi et al. 1997). Ozone was significantly associated with total mortality in Mexico City (Borja-Aburto et al. 1997) (relative risk equals to 1.024 per 200 µg/m3 increase in the hourly maximum in ozone). The Air Pollution and Health: a European Approach (APHEA) study (Touloumi et al. 1996) of four western European cities found a relative risk equal to 1.02 per 50 µg/m3 increase in both the hourly and the 8-h ozone concentrations. However, in this APHEA meta-analysis nitrogen dioxide showed no association with mortality. In a London study of air pollution and mortality between 1987 and 1992 (Anderson et al. 1996), same day ozone levels were associated with a significant increase in all cause, cardiovascular and respiratory mortality with the effects greater in the warm season and apparently independent of the effects of other pollutants.
In China, in the Dongcheng and Xicheng districts of Beijing, there was a highly significant association between log sulphur dioxide levels and daily deaths, adjusted for temperature and humidity. The strongest effects were for cardiorespiratory disease; chronic bronchitis, COPD (29 per cent), and cor pulmonale (19 per cent). Total mortality risk increased by 11 per cent with doubling of sulphur dioxide concentration (Xu et al. 1994).
Cohort studies from the United States have examined, in defined population groups, the relationship between health outcomes and long-term exposures to ambient concentrations of particulates and other pollutants. Two earlier reports (Dockery et al. 1993; Pope et al. 1995) showed positive associations between long-term concentrations of particles and mortality. The most recent report from the Adventists Health Study of Smog, describes the follow up of a cohort of 6338 non-smoking Californian Seventh Day Adventists from 1977 (Abbey et al. 1999). Past smoking and living or working with smokers was controlled for in the analysis. In 1977 to 1992 long-term concentrations of inhalable particles (PM10) and other pollutants, sulphur dioxide, sulphates, nitrogen dioxide, and ozone were related to mortality in males. The statistically significant findings were for all (non-accidental) causes in males for PM10 and also for non-malignant respiratory disease and lung cancer in males. In addition, ozone showed an association with lung cancer in males and nitrogen dioxide showed a strong association with lung cancer in both sexes. The investigators adjusted the data for both occupational and indoor sources of air pollutants. The conclusions drawn were that long periods of residence and work in areas of high ambient pollution are associated with increased mortality. However, there are some differences in outcomes between the three cohort studies, and so uncertainties remain about the long-term effects of specific pollutants and their impact on specific health outcomes.
Air pollution, morbidity, and health-care utilization
Since 1980 in North America, Europe, and Asia more than 18 cohort type studies have reported increased risks of adverse health outcomes such as cough, dyspnoea, exacerbation of asthma, upper and lower respiratory infections, and increased use of medication.
Whittemore and Korn (1980) used diaries for asthmatic subjects to record respiratory symptoms on a daily basis. Similar studies have demonstrated associations between air pollution and respiratory symptoms in The Netherlands (Hoek and Brunekreef 1993) and Switzerland (Braun-Fahrlander et al. 1992). Another measure employed in the United States is the restricted activity day resulting from respiratory problems associated with particulate concentrations (Ostro and Rothschild 1989). Particulates are associated with local upper respiratory irritation, cough, upper respiratory infections and bronchitis, and lower respiratory tract illness in children in the United States (Ware et al. 1986; Dockery et al. 1989), Finland (Jaakkola et al. 1991), and Hong Kong (Peters et al. 1996). In Hong Kong studies of primary school children aged 8 to10 years, living in areas of high and low pollution, have defined the importance of the air pollution hazards together with both active and passive smoking by the children. After adjustment for socio-economic factors, the risks for any cough, phlegm and wheeze symptoms were highest in children (10 per cent) who had ever experimented with cigarettes (odds ratio 1.85; 95 per cent confidence intervals 1.55 to 2.21), followed by those (46 per cent) exposed to smoking in the home (odds ratio 1.15; 95 per cent confidence intervals 1.01–1.30 for exposure to one smoker and odds ratio 1.45; 95 per cent confidence intervals 1.21–1.75 for two or more smokers) followed by the risks attached to living in the most polluted district (odds ratio 1.14; 95 per cent confidence intervals 1.00 –1.29) (Hedley et al. 1993). These findings are likely to reflect the situation in many developing urban communities in Southeast Asia.
The urban nature of the hazard is also demonstrated by comparisons of communities living in cities and rural areas. In German studies (Wjst et al. 1993; Weiland et al. 1994) the 9- to 15-year-olds living close to urban traffic had higher frequencies of asthma-like symptoms, wheezing, and allergic rhinitis. A two- to threefold excess of wheeze, dyspnoea, and rhinitis was found in those living in Pisa compared with the unpolluted Po Delta Valley (Viegi et al. 1991). The mean annual total suspended particulate concentrations were three to five times higher in the Pisa urban area.
Emergency department attendances and hospital admissions have been used to estimate the effects of seasonal pollution, such as ‘summer haze’ and other pollution episodes, and make inferences about the possible importance of specific pollutants. Pope (1989) demonstrated a marked effect on respiratory hospital admissions in Utah Valley following the closure and reopening of a steel mill, which was the principal source of PM10. When 24-h PM10 levels exceeded 150 µg/m3 average admissions for children nearly tripled.
Outdoor exposure to increased levels of nitrogen dioxide was associated with respiratory symptoms and their duration in Switzerland (Braun-Fahrlander et al. 1992) and hospital admission for respiratory disease in Birmingham, England (Walters et al. 1995). In Germany a time-series study showed a 28 per cent increase in admissions for croup with a daily mean nitrogen dioxide increase of 10 to 70 µg/m3 (Schwartz et al. 1991).
Particulates and sulphur dioxide usually occur together, but there are relatively few data on possible specific effects of sulphur dioxide on health care utilization. In the 1980s, in Vancouver, Bates et al. (1990) found emergency room visits were correlated with sulphur dioxide during both winter and summer. Studies in Barcelona and Santander found associations between COPD and relatively low levels of sulphur dioxide. In Hong Kong following the introduction of low sulphur fuels (Peters et al. 1996) there was a sharp fall in sulphur dioxide and sulphate in RSP (Fig. 4). Monthly averages for ambient sulphur dioxide in the polluted district ranged from 100 to 140 µg/m3 for the 6 months before the intervention, falling to less than 20 µg/m3 following the imposition of fuel regulations. In the same period, total and respirable particulates showed an initial but unsustained decline of 23 per cent and 18 per cent respectively. Sulphate concentrations in respirable suspended particulates showed more stable declines of 47 per cent and 35 per cent in the most and least polluted districts respectively. This was followed by a marked change in the risk of respiratory symptoms between a pair of high and low pollution districts. The between district differences in the prevalence of sore throat or cough, phlegm and wheeze, and doctor consultations for respiratory complaints were reduced, but there was no change in risk attributable to exposure to second-hand smoke for those children who lived in smoking homes. In Paris, Dab et al. (1996) found sulphur dioxide to be associated with respiratory admissions but in Amsterdam, Schouten et al. (1996) observed a significant negative effect.
Fig. 4 Trends in air pollution as shown by levels of sulphur dioxide (SO2) and sulphates in respirable particulates (SO4 RSP) at monitoring stations in Kwai Tsing (solid line) and in Southern (dotted line) districts measured by the Hong Kong Environmental Protection Department 1988 to 1992, before and after the introduction of regulations restricting the sulphur content of fuels.
Health outcome studies for the effects of pollution on subjects with asthma show marked variation. Bates and Sizto (1987) in Ontario found no association in the 0 to 14 year age group between asthma admissions and ozone or other single pollutants, but a positive relationship for all ages with ozone and sulphates. In southern Ontario air pollution has declined as indicated by the measurement of sulphur dioxide but not ozone, nitrogen dioxide, or coefficient of haze. During the period 1974 to 1983 total respiratory admissions declined 15 per cent, but admissions for asthma rose. However, a consistent summer relationship between sulphates, ozone, temperature, and respiratory admissions was identified independently of asthma. The authors pointed to aerosol sulphates, which explained the highest percentage of the variance in respiratory admissions in summer, but not in winter. They postulated that ozone and sulphates are only indicators of an ‘acid summer haze’.
Ito and Thurston (1999) estimated a combined relative risk of 1.18 (95 per cent confidence intervals 1.07 to 1.30) from North American and European studies for asthma admissions associated with an increase in the daily/hourly maximum concentration of ozone of 100 ppb For respiratory admissions in children there is apparently heterogeneity in estimated outcomes across age groups within the same geographical area and in the risks associated with different pollutants. Admission data may vary for many reasons, including local cultural factors and health-care seeking behaviour, the role and efficiency of primary medical care, the nature of the referral system, and the costs of care at all levels. The completeness and quality of the data may also influence findings.
Anderson (1999) has reviewed the health effects of pollution episodes. Major pollution episodes in urban settings have not produced any convincing evidence of exacerbations of asthma in children. These include the 1930 Meuse Valley episode, when those with serious asthmatic problems were mainly older adults. In the 1952 London fog, the general practitioner John Fry noted that asthmatic children in his practice did not become ill (Fry 1953). On the other hand forest fires and volcanic eruptions have caused an increase in deaths and health-care episodes. In a 1987 Californian fire asthma visits increased by 40 per cent and after the Mount St Helens eruption by a factor of 4, but unlike the Californian episode no effect on asthma was observed in the 1994 western Sydney bushfire, which was of similar intensity. In 1991, London experienced a unique high pollution episode with very high levels of nitrogen dioxide reaching an hourly average of 423 ppb All causes of mortality increased for all age groups (relative risk equals to 1.10), together with cardiovascular disease. Admissions of the elderly (65 and above) increased for respiratory disease. However, in children there were no increases in admissions for respiratory diseases overall and only a small non-significant increase for asthma (Anderson et al. 1995).
Asthma and wheezing
Burney (1988) has emphasized the need for epidemiological rather than solely clinical and experimental pathophysiological studies as the basis for establishing causation and explaining the current global distribution of asthma. Problems with the definition of asthma complicate the interpretation of prevalence estimates and the apparent differences between surveys.
Asthma is an important cause of ill health in Western developed countries and increasingly in new industrial countries in the developing world but a low prevalence of the disease is found in poor rural communities (Burney 1992). It has been increasing during the last 30 years in terms of estimated incidence, prevalence, and utilization of health care. In the United Kingdom the 30-year increase has been about 50 per cent. These data are drawn from records of national follow-up studies, hospital admissions, use of medication, and registered causes of death.
During the period when asthma has been increasing, industrial and domestic emission of smoke and sulphur dioxide have been declining; particles from diesel exhaust, together with nitrogen oxides and volatile organic compounds have increased during this time. On the other hand, urban concentrations of ozone and nitrogen dioxide have remained relatively stable.
Urban areas have a higher prevalence of asthma and there has been a strong popular view, among the public and media, that this is due to air pollution mainly from mobile sources. However, as Burney (1992) points out air pollution is not the only defining characteristic of urban areas and atopy is generally more common among young people living in more developed environments. This would include those in modern settings in the rural areas of post-industrialized countries. In contrast, there was no association between extremely high domestic particulate concentrations and the prevalence of asthma in children in Papua New Guinea (Anderson 1978).
The United Kingdom Department of Health, Committee on the Medical Effects of Air Pollutants (Department of Health 1995) has reported its conclusions on the association between asthma and air pollution. It stated that:
there is little or no association between the regional distribution of asthma and air pollution
comparisons of prevalence between high and low pollution areas have produced inconsistent results
in the United Kingdom there is no convincing evidence of an urban–rural gradient
in the United Kingdom and other countries there is a modest relationship between asthma prevalence and local traffic density
most asthma patients should not be affected by usual ambient levels of pollutants in the United Kingdom
the observed asthma trends over 30 years are unlikely to be related to changes in non-biological air pollution, excluding spores, fungi, and other allergens.
It has been postulated that population exposure to increasing levels of pollutants such as nitrogen dioxide decrease the threshold of allergen exposure needed to trigger allergic asthma and increase the morbidity associated with existing asthma (Anto and Sunyer 1995). In Barcelona the inhalation of soybean dust originating from the port area, enhanced by temperature inversions, which were also associated with increased concentrations of other pollutants, was associated with outbreaks of asthma. The presence of the inversions initially led to the conclusion that oxides of nitrogen were the principal single cause (Anto et al. 1989).
Boezen et al. (1999) demonstrated that children who have both bronchial hyper-responsiveness and high concentrations of immunoglobulin E are susceptible to air pollution. In this group the prevalence of lower respiratory symptoms increased significantly by between 32 and 139 per cent for every 100 µg/m3 increase in particulate matter. The estimated excess risks are quite low but when applied to a population the public health impact would be substantial. The findings provide a possible unifying explanation for individual susceptibility to air pollution but leave unanswered questions about the precise role of allergens in the air during pollution episodes. Furthermore, they do not explain the lack of effect during some severe pollution episodes.
Air pollution and lung function
The impact of air pollution on the general health of the population may be indicated by acute and chronic effects on lung function in both children and adults. All standard measures of lung function, including FEV1, FVC, and peak expiratory flow rate have been shown to be affected by ambient air pollution exposures. In children reductions in lung function have been observed in both well populations and those with a either a past history of wheezing or a medical diagnosis of asthma.
In the United States (Steubenville, Ohio (Dockery et al. 1982; Brunekreef et al. 1991) and Utah (Pope et al. 1992)) and The Netherlands (Hoek et al. 1993) exposure to particulates measured as PM10 or total suspended particulates (TSP) was associated with delayed or lagged reduction in lung function. In children, Dockery et al. (1982) found a 2.7 per cent decrease in FEV0.75 for a 35-µg/m3 increase in TSP and in 48 communities across the United States, Schwartz (1989) demonstrated a 6 per cent decrease in FVC with a 100 µg/m3 increase in TSP. It is postulated that the inhalation of fine particles may lead to an inflammatory response that is followed by later decline in lung function; however, a different mechanism appears to underlie the immediate reduction in lung function seen following ozone exposures (Gold et al. 1999).
Age and lifestyle, particularly exercise, which increases ventilation, are associated with variations in exposure and response to pollutants. Young children are particularly active and have high ventilation rates, and in Munich reduced peak expiratory flow rates were observed in 9- to 11-year-old children living in close proximity to dense urban traffic (Wjst et al. 1993).
Bronchial hyper-responsiveness, a bronchial constriction response to standardized inhalations of histamine, carbachol, or methacholine, can be measured as a decrease in FEV1. In Norway, exposure to low ambient sulphur dioxide levels in infancy has been related to bronchial hyper-responsiveness in 7- to 13-year-olds (Soyseth et al. 1995). In Hong Kong (Tam et al. 1994) bronchial responsiveness to a histamine challenge in 8- to10-year-olds varied between districts with contrasting levels of pollution, as did the prevalence of respiratory symptoms (Peters et al. 1996). The significance for future health of changes in lung function in children has remained uncertain. Following the introduction of low sulphur fuel regulations in Hong Kong, in the year following the intervention, the prevalence of bronchial hyper-responsiveness declined in both districts, but continued to decline in the second year in the most polluted district (Wong et al. 1998).
The observations from investigations such as time-series studies are generally accepted as inputs to environmental risk assessment and management, but major questions about the causal mechanisms remain. Evans and Wolff (1996) explored quantitatively the plausibility of one mechanism that could explain the outcome of cross-sectional studies. Air pollution can accelerate decline in lung function (Xu et al. 1991) and lung function (FEV1) is a good predictor of mortality. Evans and Wolff constructed a mathematical model of the chronic effects of continuous exposure to air pollution (1 µg/m3 per year) for smokers and non-smokers, to estimate the age-specific decline in lung function and loss of life expectancy (Fig. 5). There is an important three-way interaction between smoking and pollution exposure, together with the expected age-related loss of lung function, with about 25 years loss of life expectancy from the combined effect of smoking and pollution. In the model, below the age of 40 the effects of pollution are more readily detected in non-smokers than in smokers. On the basis of their work Evans and Wolff suggest that a figure of 50 000 premature deaths a year is possible due to relatively low levels of ambient particulates.
Fig. 5 Schematic diagram of lung function versus age showing loss of life expectancy (LOLE).
Policy-making and prevention
Air-quality guidelines are under revision by the WHO, and many countries are trying to address the problem of both setting and maintaining air quality objectives in their own locality. Local data on health effects are the most suitable, but in the developing world few countries have the resources to generate this information. The current world literature on the health effects of pollution clearly indicates that there are no firm grounds on which to characterize specific levels of individual pollutants as safe. What can be inferred is that both short- and long-term health effects are associated with pollutant mixtures and at levels of individual pollutants that are below current standards. These health effects impact on the current and future health of well populations, including children. Although research designed to explore the possible causal relationship between different pollutant species and health outcomes will continue there is ample evidence to support stringent environmental controls on industrial and mobile sources of pollution. The United Kingdom has set out objectives for further reductions in air pollutants by 2005. A Department of Health Ad Hoc Group on the Economic Appraisal of the Health Effects of Air Pollution (1999) has taken the United Kingdom Committee on Medical Effects of Air Pollution findings on deaths and admissions as the basis for estimating the health-care costs of changes in levels of air pollutants.
The United Kingdom group suggested that the loss of life expectancy incurred when deaths are advanced or ‘brought forward’ is likely to be not more than a month to a year. Factors used to adjust the valuation of the life foregone include the age of the individuals who die (mostly elderly); the pre-existing quality of life (generally lower because of chronic disease such as COPD). These led inevitably to marked downward adjustments of the estimated willingness to pay to reduce air pollution mortality risks, and much lower than the value attached to road traffic accidents.
Much of the current literature focuses on health events that are represented in the upper part of the population pyramid representing pollution effects such as deaths and hospital admissions. However, surveys of children, a sensitive sentinel group for pollution health effects, show that in many situations pollution is the cause of very high endemic levels of respiratory complaints and use of health services. Cost–benefit analysis based only on more severe outcomes will inevitably understate the overall benefits of pollution controls to a community.
Furthermore, there is an uncertain progression between epidemiological research, development of guidelines, and policy decisions from regulatory agencies and government departments, and the decisions of legislators and the judiciary. In 1997, a United States federal appeals court rejected several USEPA clean air standards. The appeal was brought to court by the United States Chamber of Commerce on grounds that the new air pollution rules were too expensive, without scientific basis, and that the EPA had exceeded its authority. This court decision caused the EPA to issue the following statement on 14 May 1999 (EPA 1999):
EPA stands by the need for the health protections embodied by the clean air standards and the science behind them. The soot and smog standards put in place almost two years ago will protect the health of 125 million Americans including 35 million children…. If the courts fail to uphold these protective standards, congress must ensure that these protections are preserved for the American people and EPA stands ready to work with them.
The way ahead for effective local legislation will be supported by better evidence on health effects of pollution within countries and regions. But the introduction of environmental controls will require political will, which may only stem from analyses showing that pollution not only harms the health of populations, but the economy as a whole.
The changing global situation
Tuberculosis is the most frequent cause of death from any infectious agent. About one-third of the world’s population is infected, with about 8 million new cases a year occurring in the mid-1990s, associated with 3 million deaths. Tuberculosis is a leading cause of death in the age group 15 to 44, especially in women, where it is responsible for 10 per cent of deaths (Murray and Lopez 1997b). Tuberculosis is the cause of about one-third of all AIDS-related deaths in Africa. In 1993, the WHO declared tuberculosis to be a global emergency.
The estimates for deaths from tuberculosis in Table 4 (WHO 1999a) are for individuals who were HIV negative. The estimated mortality in 1998 was 1.49 million with a range of 1.1 to 2.2 million deaths. Table 5 (also modified from WHO 1999a) includes gender-specific estimates for incidence, deaths, and DALYs for HIV-positive and HIV-negative individuals with tuberculosis. An additional 365 000 HIV-positive individuals are estimated to have died from tuberculosis in 1998. Eighty-three per cent of these deaths occurred in Africa, with approximately equal numbers of males and females. Low and middle income groups in the Americas and Southeast Asia, including India contributed 49 000 deaths compared with 9000 from Europe and the whole Western Pacific.
Table 4 Mortality and DALYs by gender and WHO member states: estimates for 1998
Table 5 Tuberculosis: magnitude of the problem by sex and WHO region: estimates for 1998
The largest numbers of new cases of tuberculosis arise in Southeast Asia (2.9 million), China (1.4 million), and Africa (1.05 million). Forty-nine per cent of African cases are HIV positive. The estimated burden of the disease, expressed as disability adjusted days, is greatest in Southeast Asia, Africa, China, and eastern Mediterranean regions. However, in the Asia Pacific region there is marked variation in reported notifications between both high-income and low-income countries (Table 6) (SEAMIC 1998). Some of these data appear implausible. For example, it seems highly unlikely that a country such as Thailand, with a very high prevalence of HIV infection, would have a tuberculosis notification rate that is less than half of that in Hong Kong where the risk of HIV is still relatively low. On the other hand in three high-income postindustrialized countries or regions, Japan, Singapore, and Hong Kong there is a threefold variation between the highest and lowest notification rates. Given that the information systems supporting notification are relatively well developed these differences are unlikely to be due to under-reporting in Japan and Singapore (Table 6). Many data sources are, however, incomplete (as in, for example, Indonesia and Malaysia for the certification of tuberculosis in deaths), and this raises questions about the validity of the notification data.
Table 6 Tuberculosis notifications and mortality in Asia Pacific countries
Identifying problems in global control of tuberculosis
In the 1980s and 1990s the reverse in the previously established decline of tuberculosis led the WHO to declare the global emergency. The earlier fall in tuberculosis incidence in the nineteenth and twentieth centuries was attributed to improvements in socio-economic conditions and the isolation of infected cases. However, the pattern of disease control this century now describes a U-shaped curve in which the seriousness and complexity of disease is much greater in the second limb of the U (Reichman 1991).
In 1991, The World Health Assembly proposed an effective tuberculosis control strategy with two goals, to be achieved by the year 2000: successful treatment of 85 per cent of cases and detection of 70 per cent. The features of the WHO Directly Observed Therapy Short-course (DOTS) strategy were:
government commitment to tuberculosis control
case detection focusing on patients with symptoms who self-report to health services
use of sputum smear microscopy
standardized administration of the short-course treatment
direct observation of the chemotherapy for at least 2 months
adequate supplies of treatments
good records and information systems to allow evaluation of treatment results.
The WHO set up a surveillance and monitoring project in 1995 to assess national tuberculosis programmes and to compare regions that had adopted the WHO strategy (DOTS) and those that had not. Raviglione et al. (1997) estimated that within those countries that adopted the strategy only 23 per cent of the world population was covered.
Improved delivery of care must focus on early diagnosis and reliable delivery of effective treatments, but in 1998 the WHO identified 22 countries, which accounted for 80 per cent of tuberculosis cases worldwide, of which 16 countries were not making satisfactory progress in the control of the disease (Table 7). Half of those with a poor performance were middle income countries that had the resources but not the political will to tackle the problem (Wise 1998). There is also a wide disparity in funding, in the form of external aid, between different infectious diseases for each patient over the age of 5 years who died from the disease. Tuberculosis is clearly the most underfunded health-care problem in this group (Table 8) (Zumla and Grange 1999).
Table 7 Progress in the control of tuberculosis in 22 countries
Table 8 Resources available in the form of external aid, between different infectious disease groups, for each patient over the age of 5 years who died from the disease
The failure of tuberculosis controls is exemplified by the problem in the Philippines, which has been estimated to have the highest numbers of cases of tuberculosis per capita in Southeast Asia and is ranked fourth in the world. No progress has been made in two decades, and the situation is complicated by low utilization of services by people with symptoms, high levels of drug resistance (60 per cent for isoniazid), the social stigma attached to the disease, and inability to pay for a full course of treatment (US$75) (Easton 1998).
Tuberculosis is associated with war, famine, social disruption, and poverty, including homelessness, unemployment, imprisonment, and alcoholism. In Russia the incidence and mortality has increased steadily since 1990 with rates of 67.5 per 100 000 and 17.0 per 100 000 respectively since 1996 (Wares and Clowes 1997).
Prospects for the future, in projections of mortality and disability by cause for 1990 to 2020 as part of the Global Burden of Disease Study (Murray and Lopez 1997c), indicate that tuberculosis ranked seventh as a cause of death in 1990 and its rank is not predicted to change throughout the 30-year period (Table 9). The projected total number of DALYs attributed to tuberculosis worldwide in 2020 is 42.5 million, of which 42.4 million arise in developing regions (Table 10).
Table 9 Changes in ranking for most important causes of death from 1990 to 2020 in baseline scenario
Table 10 Ten projected leading causes of DALYs in 2020 according to baseline projection
These current trends are complicated throughout the world by the growing problem of double infection by tuberculosis and HIV. Those developing countries in warm climates will experience the sharpest rises in incidence and this is directly related to the incidence of tuberculosis in the emerging HIV epidemics in Africa, India, and Asian Pacific countries. It will dictate the need for a marked change in priorities and a reallocation of resources from both within and outside of existing services, but the public health priorities are the alleviation of poverty and provision of good quality services to which all those in need have unhindered access.
Many predictions for tuberculosis are pessimistic, but others argue that the tools to implement cost-effective programmes are available. One year of healthy life can be gained for no more than US$3, making tuberculosis intervention one of the most cost-effective interventions in health care (Kochi et al. 1997).
Surveillance of tuberculosis
Poor information quality
In many countries, especially those with the highest risks, surveillance and information systems cannot provide information of adequate quality for community diagnoses and monitoring. Surveys on case detection and notification are notoriously unreliable and estimations are largely based on the prevalence of tuberculin reactivity as an indicator of infection by tuberculosis and, by extrapolation, the expected numbers of infected persons who will develop tuberculosis disease (Zumla and Grange 1999).
Incidence and prevalence
The current estimates of the global problem, in terms of both morbidity and mortality of tuberculosis, have been based on a range of clinical and microbiological measures of tuberculosis infection. The principal epidemiological variables used to describe the magnitude, trends, and impact of tuberculosis are:
incidence (including incidence of smear positive cases)
prevalence (including prevalence of smear positivity in all cases)
predicted incidence (including use of skin tests)
mortality (including case fatality in smear positive and other cases).
Assessment of the emergent problem depends on the availability of comprehensive, accurate, and continuous information on incidence and mortality. The development and maintenance of surveillance systems is a vital component of any strategy for tuberculosis control and treatment, which in itself is a cost-effective approach to population health. Tuberculosis is a notifiable disease in most countries, but routine audit of the notification of cases is usually lacking and both accuracy and reliability are uncertain. When audit is carried out under-notification is usually identified and improvement occurs when notification procedures are revised and reinforced (Brown et al. 1995). Effective notification systems require record linkage between clinical units, pathology laboratories, and pharmacies.
The problem of linking information management with clinical practice requires systems analysis and continuing operations research directed at the chosen solutions. The WHO global tuberculosis programme has started the Global Tuberculosis Research Initiative (McConnell 1998). The aim is to promote operations research and develop the skills and funding to support it. Operations research is need to achieve improvements in surveillance and epidemiology. New approaches are required worldwide for the achievement of these goals in mixed medical economies and between different levels and sectors of health-care systems, which are often complicated by the politics of relationships between government, non-governmental organizations, and health professionals. The first problem to be addressed in tuberculosis control is the quality of information.
Factors influencing global trends in tuberculosis
Several authors have reported on the rising age of tuberculosis patients in developed countries and a slowing down in the decline of notification rates overall.
To examine long-term trends in tuberculosis, Tocque et al. (1998) used a birth cohort analysis approach. They calculated age-specific rates of disease, by different age groups for different birth cohorts, for England and Wales and Hong Kong. In Hong Kong each birth cohort showed a similar pattern of disease by age with rates peaking in the 29- to 39-year age groups and then gradually declining (Fig. 6). Since 1978, regardless of age at that time, all age cohorts showed an increase in tuberculosis rates with age particularly in females. A similar pattern was seen in England and Wales but the peak occurred at an earlier age (less than 25 years) and the pattern of decline with age did not cease until 1984. Life expectancy is increasing steadily, particularly in countries with high or increasing per capita incomes. In the Asia Pacific region this includes the post-industrialized countries or regions of Japan, Taiwan, Singapore, and Hong Kong, but the numbers of elderly are also increasing across many other developing nations. The ageing of communities, therefore, is becoming an important potential factor in the deterioration in control of tuberculosis.
Fig. 6 Age incidence curves for male tuberculosis cases in Hong Kong from 1958 to 1993. The arrows indicate how rates were calculated for two individual birth cohorts as examples of how data were derived to compare age-cohort data between Hong Kong and England and Wales.
In addition to its contribution to the burden of cancer and heart and respiratory disease, smoking is an importance influence on the risk of tuberculosis mortality.
In China the smoker to non-smoker tuberculosis mortality ratio for men is 1.17 (rural) and 1.42 (urban) and in women 1.25 and 1.56 respectively (Liu et al. 1998), and smoking was estimated to account for 12 per cent of all tuberculosis deaths. A lung that is damaged by smoking may offer a propitious environment for the tuberculosis bacillus (WHO 1999a). Because of the high prevalence of smoking in men, in excess of 60 per cent in many Asian countries, tobacco will continue to make an important contribution to the burden of tuberculosis in the twenty-first century.
A new epidemic wave of tuberculosis is following the global increase in HIV with 6 million cases of both HIV and tuberculosis. Seventy-five per cent are estimated to be in Africa (WHO 1994).
In the mid-1990s the prevalence of combined infection with HIV and tuberculosis was about 6 million and HIV was estimated to cause up to half a million new cases of tuberculosis per year (WHO 1996). Information from African countries with reliable information systems (Burundi, Malawi, Tanzania, and Zambia) indicates the impact on their health-care systems from the HIV-related rise in incidence of tuberculosis (WHO 1996). The harm to maternal and child health is particularly serious (Chintu and Zumla 1995); in Zambia, where 25 per cent or more pregnant women are HIV positive, tuberculosis is the principal cause of death in pregnancy (Fylkesnes et al. 1997).
Tuberculosis is now a major problem in patients with HIV infection in the United States with an incidence of up to 9000 new cases per year (Markowitz et al. 1997). In New York the rate of tuberculosis is four times the national average and associated with nosocomial spread and multidrug-resistant tuberculosis. In London, tuberculosis rates have increased by 35 per cent compared with 15 per cent in England and Wales, but the extent of HIV-associated tuberculosis is uncertain because notification of tuberculosis in HIV is unreliable (Pym et al. 1995).
Cost, access to antiretroviral drugs, and adherence to treatment are major obstacles that should be priorities in aid programmes to Africa and in health care for HIV-positive immigrants to Western countries. HIV-related tuberculosis is a barometer for tuberculosis control and highlights weaknesses in its prevention and treatment (Coker and Miller 1997).
Prisons have been identified as important reservoirs of tuberculosis, particularly drug-resistant strains, in both Western industrialized and developing countries. Surveys of prisons in New York City, Russia, Azerbaijan, Malawi, and Ethiopa have demonstrated a strong association between tuberculosis and HIV infection in prisoners and both the opportunities for and problems of achieving solutions in these neglected settings (Nyangulu et al. 1997; Reyes and Coninx 1997; Coninx et al. 1998).
Poor clinical practice, inadequate supervision and resources, and unethical commercial sales of inappropriate antituberculosis drugs and combinations have all combined to create the problem of drug resistance. The risk of multidrug-resistant tuberculosis, described as the ‘Third Epidemic’, may become the dominant pattern of tuberculosis spread on several continents, fanning out from countries such as Peru, northern India, Sierra Leone, the Baltic States, and Russia. Outbreaks of drug-resistant tuberculosis, once rare, are now common place and many involve HIV-positive patients. Early outbreaks were associated with high mortality because of failure of recognition and appropriate action. Several outbreaks of multidrug-resistant tuberculosis have occurred in Europe, including nosocomial infections in London and Madrid.
The WHO International Union Against Lung and Tuberculosis Disease global project on Anti-Tuberculosis Drug Resistance and Surveillance 1994 to 1997 in 35 countries found evidence of resistance in all of them, in patients with no previous treatment the prevalence was 9.9 per cent (range 2 to 41 per cent). Mycobacterium tuberculosis strains were resistant to one or more drug. High prevalences were found in the former Soviet Union, Asia, Dominican Republic, and Argentina (Pablos-Mendez et al. 1998).
Provision of care for tuberculosis and compliance with treatment
It is estimated that less than half of those with tuberculosis who need medical care are in contact with treatment services. Continuity of care for tuberculosis often breaks down and this represents an additional hazard for community spread of the disease. Diagnosis has always presented problems (Zumla and Grange 1999). The protean manifestations, mimicry of other conditions, validity of tests, including lack of sensitivity (e.g. sputum microscopy) or specificity (chest radiology), and slow procedures (culture) all contribute to the problem.
Nucleic acid technology (polymerase chain reaction and ligase chain reaction) are evolving as potentially useful diagnostic tools. Sequencing of the M. tuberculosis genome should also allow development of rapid tests; however, the availability of new technology to those countries that need it most urgently will be limited for many years (Zumla and Grange 1999). Near-patient (as opposed to central laboratory) testing could yield important public health benefits, such as rapid initiation of contact tracing. However, the validity and reliability of such tools in the field will determine their cost-effectiveness. There may also be negative effects on the capture of epidemiological information from central diagnostic facilities, compounding the existing shortfalls in record keeping and notification. Therefore, implementation of these techniques will require their integration with new information technology, for example to enable automatic storage and transmission of results from multiple geographically separate sites (Borriello 1999).
Short-course multiple drug regimens
A wide range of short-course regimens are employed in different countries. Choices are based on past experience and established practice, cost and availability of drugs and evidence for effectiveness in trials. In Hong Kong (Hong Kong Chest Service 1984a,b) randomized trials of different drug combinations and durations of treatment showed that four-drug regimens were more effective than a three-drug regimen without streptomycin, at preventing relapse at 30 months. Four-drug regimens will contribute to population control of the disease by preventing the development of drug resistance and primary resistance in contacts.
Directly observed therapy
Directly observed therapy has been strongly promoted in the 1990s as the key to the control of tuberculosis in both developing and developed countries. Iseman et al. (1993) argued that we cannot afford not to do it and in 1997 the World Bank and the WHO contended that directly observed therapy is the most cost-effective of all health interventions. This revelation, implying a breakthrough in therapeutic management, caused concern and strong arguments that good evidence on the best approaches to implementation of supervised therapy and rigorous evaluation were lacking (Grange and Zumla 1997).
There are clearly mixed views on many aspects of this strategy. The idea of directly observed therapy was prompted by the need to tackle non-adherence to treatment and prevent the disastrous consequences of it. It was developed in settings characterized by poverty (Bayer and Wilkinson 1995). Now, because of enormous escalation of the scale of tuberculosis in populations, for example from 5000 to 19 000 in Malawi between 1985 and 1995 (Malawi National Tuberculosis Programme), high costs and unavailability of hospital care, directly observed therapy is seen to be a potentially important contribution to new global tuberculosis control strategies such as short-course drug regimens (Squire and Wilkinson 1997).
In the Central African Republic (where directly observed therapy has not been implemented), 24 months after the start of treatment, the cumulative death rate was 20 per cent in HIV-1 negative patients and 58 per cent in HIV-positive patients (Fig. 8). Age and HIV were the strongest risk factors. Multidrug-resistant tuberculosis was not associated with increased mortality, probably due to the small numbers in the study and high overall mortality, which is the highest in Africa. Lower mortality was associated with completion of an 8-month treatment plan (Garin et al. 1997).
In China the adoption of the WHO DOTS strategy was associated with a cure rate of nearly 90 per cent in new cases and a fall in the treatment failure rate in previously treated cases from 17.6 to 6.2 per cent. Supervision of treatment was provided by village doctors with financial incentives and careful supervision (China Tuberculosis Control Collaboration 1996).
However, a randomized controlled trial of directly observed therapy compared with self-supervised patients showed the equivalence of the two approaches with a trend for better outcomes in the self-supervised group (Zwarenstein et al. 1998). The investigators concluded that directly observed therapy is authoritarian, reduces self-reliance, and alienates patients. Flexibility of approach taking into account local conditions is necessary as demonstrated by the success of other approaches in rural Nepal (Jochem et al. 1997). The trial carried out by Zwarenstein and colleagues, which included a relatively small number of eligible patients with high treatment interruption rates, was prompted by the high workload imposed by a directly observed therapy programme and its report led to many cautionary statements about advocacy of any one method without adequate testing of its individual components.
Volmink and Garner (1997a) argued that it is necessary to develop more reliable reviews of directly observed therapy to support organizations such as the WHO and the World Bank. Without this, rhetoric about directly observed therapy may lead to currently effective programmes being discarded without a thorough evaluation of all the options available. A systematic review of five randomized or pseudo-randomized controlled trials of other preventive or curative strategies for tuberculosis showed that all six interventions tested were effective (Volmink and Garner 1997b). The interventions tested that did not include the ‘supervised swallowing’ procedures of directly observed therapy, were component reminder letters, monetary incentives to patients, health education, intensive supervision of staff and two combinations of these strategies. As all the interventions were effective the findings may indicate a form of Hawthorne effect in which increased overall commitment and intensity and comprehensiveness of interventions yield the desired clinical and public health results. Such an effect can obviously be exploited for benefit, although the precise elements of any one strategy that is responsible for its success may remain unidentified.
The need for a public health approach
Few tuberculosis treatment programmes achieve the aim of 85 per cent adherence and completion of the drug regimen. Inadequate adherence to treatment is very common. This is one of the most serious problems in tuberculosis management and has profound public health implications. Factors that are associated with, or predict, non-adherence have been identified from case studies rather than randomized trials (Table 11). In addition to directly observed therapy and other approaches mentioned previously it is essential that we improve the approach of services to delivering patient communication and education (Table 12). This approach is low cost and extremely effective and in any case should be a mandatory feature of good-quality (i.e. ethical) medical care. It is estimated that 15 to 20 min of patient education yields 2 months of regular attendance without default.
Table 11 Factors associated with non-adherence in the treatment of tuberculosis
Table 12 Steps to improve adherence and overall effectiveness of programmes
The lessons learned from the New York tuberculosis epidemic demonstrate that even in the West tuberculosis is as much a political and fiscal issue as a medical management problem, and therefore so are the solutions (Coker 1998). In New York, funding for tuberculosis was severely cut in the 1970s. To this was added the internal management problems in tuberculosis services and the growth of overcrowding, inequalities, and HIV in high-risk neighbourhoods. Treatment rates dropped to an average of 60 per cent. The response to the crisis cost US$1 billion, but through effective public health action led to a halving of case loads and an 85 per cent reduction in multidrug resistance. Coker recalls the New York City Board of Health slogan of 1915:
The city can have as much reduction of preventable disease as it wishes to pay for. Public health is purchasable; within natural limitations a city can determine its own death rate.
Abbey, D.E., et al. (1999). Long term inhalable particles and other air pollutants related to mortality in non-smokers. American Journal of Respiratory and Critical Care Medicine, 159, 373–82.
American Thoracic Society (1985). Guidelines as to what constitutes an adverse respiratory health effect, with special reference to epidemiologic studies of air pollution. American Review of Respiratory Disease, 131, 666–8.
American Thoracic Society/Committee of the Environmental and Occupational Health Assembly of the American Thoracic Society (1996). Health effects of outdoor pollution. American Journal of Respiratory and Critical Care Medicine, 153, 3–5, 477–98.
Anderson, H.R. (1978). Respiratory abnormalities in Papua New Guinea children: the effects of locality and domestic wood-smoke pollution. International Journal of Epidemiology, 7, 63–71.
Anderson, H.R. (1999). Health effects of air pollution episodes. In Air pollution and health (ed. S.T. Holgate, J.M. Samet, H.S. Koren, and R.L. Maynard), pp. 461–82. Academic Press, London.
Anderson, H.R., Limb, E.S., Bland, J.M., de Leon, A.P., Strachan, D.P., and Bower, J.S. (1995). Health effects of an air pollution episode in London December 1991. Thorax, 50, 1188–93.
Anderson, H.R., de Leon, A.P., Bland, J.M., Bower, J.S., and Strachan, D.P. (1996). Air pollution and daily mortality in London: 1987–92. British Medical Journal, 312, 665–9.
Anthonisen, N. (1997). Epidemiology and the lung health study. European Respiratory Review, 7, 202–5.
Anto, J.M. and Sunyer, J. (1995). Nitrogen dioxide and allergic asthma: starting to clarify an obscure association. Lancet, 345, 402–3.
Anto, J.M., Sunyer, J., Rodriguez-Roisin, R., Suarez-Cervera, M., and Vazquez, L. (1989). Community outbreaks of asthma associated with inhalation of soybean dust. New England Journal of Medicine, 320, 1097–102.
Australian National Health and Medical Research Council (1997). The health effects of passive smoking. Report of the NHMRC Passive Smoking Working Party. Australia Government Publishing Service, Canberra.
Barnes, D.E. and Bero, L.A. (1998). Why review articles on the health effects of passive smoking reach different conclusions. Journal of the American Medical Association, 279, 1566–70.
Bates, C. and Brookes, K. (1999). New measures to tackle passive smoking in the workplace: question and answers. UK Action on Smoking and Health. http://www.ash.org.uk/papers/acop.html.
Bates, D.V. and Sizto, R. (1987). Air pollution and hospital admissions in southern Ontario: the acid summer haze effect. Environmental Research, 43, 317–31.
Bates, D.V., Baker-Anderson, M., and Sizto, R. (1990). Asthma attack periodicity: a study of hospital emergency visits in Vancouver. Environmental Research, 51, 51–70.
Bayer, R. and Wilkinson, D. (1995). Directly observed therapy for tuberculosis: history of an idea. Lancet, 345, 1545–8.
Boezen, H.M., et al. (1999). Effects of ambient air pollution on upper and lower respiratory symptoms and peak expiratory flow in children. Lancet, 353, 874–8.
Boffetta, P., et al. (1998). European multicentre case-control study of lung cancer in non-smokers. Technical Report No. 33, International Agency for Research on Cancer, Lyon.
Borja-Aburto, V.H., Loomis, D.P., Bangdiwala, S.L., Shy, C.M., and Rascon-Pacheco, R.A. (1997). Ozone, suspended particulates, and daily mortality in Mexico City. American Journal of Epidemiology, 145, 258–68.
Borriello, S.P. (1999). Near patient microbiological tests. British Medical Journal, 319, 298–301.
Bradshaw, D., Kielkowski, D., and Sitas, F. (1998). New birth and death registration forms—a foundation for the future, a challenge for health workers. South African Medical Journal, 88, 971–4.
Braun-Fahrlander, C., et al. (1992). Air pollution and respiratory symptoms in preschool children. American Review of Respiratory Disease, 145, 42–7.
British Thoracic Society/The COPD Guidelines Group of the Standards of Care Committee of British Thoracic Society (1997). BTS guidelines for the management of chronic obstructive pulmonary disease. Thorax, 52 (Supplement 5), 1–28.
Brown, J.S., Wells, F., Duckworth, G., Paul, E.A., and Barnes, N.C. (1995). Improving notification rates for tuberculosis. British Medical Journal, 310, 974.
Brunekreef, B., et al. (1991). Sensitive subgroups and normal variation in pulmonary function response to air pollution episodes. Environmental Health Perspectives, 90, 189–93.
Burney, P.G. (1992). Asthma. Epidemiology. British Medical Bulletin, 48, 10–22.
Burney, P.G.J. (1988). Why study the epidemiology of asthma? Thorax, 43, 425–8.
Camilli, A.E., Burrows, B., Knudson B., Lyle, S.K., and Lebowitz, M.D. (1987). Longitudinal changes in forced expiratory volume in one second in adults. Effects of smoking and smoking cessation. American Review of Respiratory Disease, 135, 794–9.
Chapman, S., et al. (1999). The impact of smoke-free workplaces on declining cigarette consumption in Australia and the United States. American Journal of Public Health, 89, 1018–23.
Chief Medical Officer’s Committee on Medical Aspects of Food (1998). Nutritional aspects of the development of cancer. HMSO, London.
China Tuberculosis Control Collaboration (1996). Results of directly observed short course chemotherapy in 112 842 Chinese patients with smear-positive tuberculosis. Lancet, 347, 358–62.
Chintu, C. and Zumal, A. (1995). Childhood tuberculosis and infection with the human immunodeficiency virus. Journal of Royal College of Physicians of London, 29, 92–5.
Coker, R. (1998). Lessons from New York’s tuberculosis epidemic. British Medical Journal, 317, 616.
Coker, R. and Miller, R. (1997). HIV associated tuberculosis. A barometer for wider tuberculosis control and prevention. British Medical Journal, 314, 1847.
Coninx, R., et al. (1998). Drug resistant tuberculosis in prisons in Azerbaijan: case study. British Medical Journal, 316, 1423–5.
Cuneo, W.D. and Snider, D.E., Jr (1989). Enhancing patient compliance with tuberculosis therapy. Clinics in Chest Medicine, 10, 375–80.
Dab, W., et al. (1996). Short term respiratory health effects of ambient air pollution: results of the APHEA project in Paris. Journal of Epidemiology and Community Health, 50, S42–6.
Darby, S.C. and Samet, J.M. (1994). Radon. In Epidemiology of lung cancer (ed. J.M. Samet), p. 219. Marcel Dekker, New York.
Davis, R.M. (1998). Exposure to environmental tobacco smoke: identifying and protecting those at risk. Journal of the American Medical Association, 280, 1947–9.
den Otter, J.J., van Dijk, B., van Schayck, C.P., Molema, J., and van Weel, C. (1998). How to avoid underdiagnosed asthma/chronic obstructive pulmonary disease? Journal of Asthma, 35, 381–7.
Department of Health (1998). Nutritional aspects of the development of cancer. Working Group on Diet and Cancer of the Committee on Medical Aspects of Food and Nutrition Policy. Report on the Health and Social Subjects, No. 48. Department of Health, HMSO, London.
Department of Health/Ad Hoc Group on the Economic Appraisal of the Health Effects of Air Pollution (1999). Economic appraisal of the health effects of air pollution. HMSO, London.
Department of Health/Committee on the Medical Effects of Air Pollutants (COMEAP) (1995). Asthma and outdoor pollution (Chairman S.T. Holgate). HMSO, London.
Department of Health/Committee on the Medical Effects of Air Pollutants (COMEAP) (1998). Quantification of the effects of air pollution on health in the United Kingdom. HMSO, London.
Destache, C.J., Dewan, N., O’Donohue, W.J., Campbell, J.C., and Angelillo, V.A. (1999). Clinical and economic considerations in the treatment of acute exacerbations of chronic bronchitis. Journal of Antimicrobial Chemotherapy, 43 (Supplement A), 107–13.
Dockery, D.W., et al. (1982). Change in pulmonary function in children associated with air pollution episodes. Journal of the Air Pollution Control Association, 32, 937–42.
Dockery, D.W., et al. (1989). Effects of inhalable particles on respiratory health of children. American Review of Respiratory Disease, 139, 587–94.
Dockery, D.W., et al. (1993). An association between air pollution and mortality in six US cities. New England Journal of Medicine, 329, 1753–9.
Easton, A. (1998). Tuberculosis controls in Philippines have failed so far. British Medical Journal, 317, 557.
Eisner, M.D., Smith, A.K., and Blanc, P.D. (1998). Bartenders’ respiratory health after establishment of smoke-free bars and taverns. Journal of the American Medical Association, 280, 1909–14.
EPA (Environmental Protection Agency) (1999). Statement of the US EPA on court’s decision on clean air rules, 14 May. http://www.epa.gov/ttn/oarpg/gen/epastat.pdf.
European Working Group (1996). Environmental tobacco smoke and lung cancer: an evaluation of the risk. European Working Group, Tronheim.
Evans, J. and Wolff, S. (1996). Modeling of air pollution impacts: one possible explanation of the observed chronic mortality. In Particle in our air: concentrations and health effects (ed. R. Wilson and J.D. Spengler). Harvard University Press, Boston, MA.
Fletcher, C., Peto, R., Tinker, C., and Speizer, F.E. (1976). The natural history of chronic bronchitis and emphysema. Oxford University Press.
Froggatt, P. (1988). Fourth report of the independent scientific committee on smoking and health. HMSO, London.
Fry, J. (1953). Effects of a severe fog on a general practice. Lancet, i, 235–3.
Fylkesnes, K., et al. (1997). The HIV epidemic in Zambia: socio-demographic prevalence patterns and indications of trends among childbearing women. AIDS, 11, 339–45.
Garin, B., et al. (1997). High mortality rates among patients with tuberculosis in Bangui, Central African Republic. Lancet, 350, 1298.
Glantz, S., Slade, J., Bero, L.A., Hanauer, P., and Barnes, D.E. (1996). The cigarette papers. University of California Press, Berkeley, CA.
Gold, D.R., et al. (1999). Particulate and ozone pollutant effects on the respiratory function of children in southwest Mexico City. Epidemiology, 10, 8–16.
Grange, J.M. and Zumla, A. (1997). Making DOTS succeed. Lancet, 350, 157.
Hackshaw, A.K., Law, M.R., and Wald, N.J. (1997). The accumulated evidence on lung cancer and environmental tobacco smoke. British Medical Journal, 315, 980–8.
Harrington, J.M. and Saracci, R. (1994). Occupational ancer: clinical and epidemiological aspects. In Hunter’s diseases of occupations (8th edn) (ed. P.A.B. Raffle, P.H. Adams, P.J. Baxter and W.R. Lee), pp. 654–88. Edward Arnold, London.
He, Y., et al. (1994). Passive smoking at work as a risk factor for coronary heart disease in Chinese women who have never smoked. British Medical Journal, 308, 380–4.
He, J., et al. (1999). Passive smoking and the risk of coronary heart disease—a meta-analysis of epidemologic studies. New England Journal of Medicine, 340, 958–9.
Hedley, A.J. and Lam, T.H. (1997). Respiratory disease. In Oxford textbook of public health, Vol. 3 (3rd edn) (ed. R. Detels, W.W. Holland, J. McEwen, and G.S. Omenn), pp. 1081–11. Oxford University Press.
Hedley, A.J., et al. (1993). Air pollution and respiratory health in primary school children in Hong Kong 1989–1992. Report to Environmental Protection Department, Hong Kong Government. Department of Community Medicine, the University of Hong Kong.
Hoek, G. and Brunekreef, B. (1993). Acute effects of a winter air pollution episode on pulmonary function and respiratory symptoms of children. Archives of Environmental Health, 48, 328–35.
Hoek, G., Brunekreef, B., Kosterink, P., Van den Berg, R., and Hofschreuder, P. (1993). Effect of ambient ozone on peak expiratory flow of exercising children in The Netherlands. Archives of Environmental Health, 48, 27–32.
Holgate, S., Koren, H.S., Samet, J.M., and Maynard, R.L. (ed.) (1999). Air pollution and health. Academic Press, London.
Holland, W.W., et al. (1979). Health effects of particulate pollution: reappraising the evidence. American Journal of Epidemiology, 110, 525–659.
Hong Kong Chest Service/Tuberculosis Research Centre, Madras/British Medical Research Council (1984a). A controlled trial of 2-month, 3-month and 12-month regimens of chemotherapy for sputum smear negative pulmonary tuberculosis: results at 60 months. American Review of Respiratory Disease, 130, 23–8.
Hong Kong Chest Service/Tuberculosis Research Centre, Madras/British Medical Research Council (1984b). A controlled trial of 3-month, 4-month, and 6-month regimens of chemotherapy for sputum-smear negative tuberculosis: results at 5 years. American Review of Respiratory Disease, 130, 871–6.
International Commission on Occupational Health (1999). Membership directory 1999.
Iseman, M.D., Cohen, D.L., and Sbarbaro, J.A. (1993). Directly observed treatment of tuberculosis. We can’t not afford to do it. New England Journal of Medicine, 328, 576–8.
Ito, K. and Thurston, G.D. (1999). Epidemiological studies of ozone exposure effects. In Air pollution and health (ed. S.T. Holgate, J.M. Samet, H.S. Koren, and R.L. Maynard), pp. 485–510. Academic Press, London.
Jaakkola, J.J.K., Paunio, M., Virtanen, M., and Heinonen, O.P. (1991). Low level air pollution and upper respiratory infections in children. American Journal of Public Health, 81, 1060–3.
Jaakkola, M.S., Jaakkola, J.J., Becklake, M.R., and Ernst, P. (1996). Effect of passive smoking on the development of respiratory symptoms in young adults: an 8-year longitudinal study. Journal of Clinical Epidemiology, 49, 581–6.
Jochem, K., et al. (1997). Tuberculosis control in remote districts of Nepal comparing patient-responsible short-course chemotherapy with long-course treatment. International Journal of Tuberculosis and Lung Disease, 1, 502–8.
Jones, J.R., Hodgson, J.T., Clegg, T.A., and Elliott, R.C. (1998). Self-reported work-related illness in 1995: results from a household survey. HSE Books, Sudbury.
Katsouyanni, K., et al. (1997). Short term effects of ambient sulphur dioxide and particulate matter on mortality in 12 European cities: results from time series data from the APHEA project. British Medical Journal, 314, 1658–63.
Kochi, A., Nunn, P., Dye, C., and Tayler, E. (1997). Global burden of disease. Lancet, 350, 142.
Kuller, L.H., et al. (1989). The epidemiology of pulmonary function and COPD mortality in the Multiple Risk Factor Intervention Trial. American Review of Respiratory Disease, 140, S76–81.
Lam, T.H. and Hedley, A.J. (1999). Environmental tobacco smoke in Asia: slow progress against great barriers. Journal of the American Medical Association, Southeast Asia, 15, 7–9.
Lam, T.H., et al. (1996). Smoking and exposure to occupational hazards in 8304 workers in Guangzhou, China. Occupational Medicine, 46, 351–5.
Lam, T.H., Ho, S.Y., Hedley, A.J., and Mak, K.H. (1998). Mentioning smoking as a cause of death on death certificates. British Medical Journal, 317, 1456.
Lam, T.H., et al. (1999). Respiratory symptoms and environmental tobacco smoke in police officers in Hong Kong. The 15th International Scientific Meeting of the International Epidemiological Association, Abstract Book, Vol. 1, p. 164.
Last, J.M. (1995). Dictionary of epidemiology (3rd edn). Oxford University Press.
Leuenberger, P., et al. (1994). Passive smoking exposure in adults and chronic respiratory symptoms (SAPALDIA Study). American Journal of Respiratory and Critical Care Medicine, 150, 1222–8.
Levi, F., Lucchini, F., Vecchia, C.L., and Negri, E. (1999). Trends in mortality from cancer in the European Union, 1955–94. Lancet, 354, 742–3.
Liu, B.Q., et al. (1998). Emerging tobacco hazards in China: 1. Retrospective proportional mortality study of one million deaths. British Medical Journal, 317, 1399–400.
Logan, W.P.D. (1953) Mortality in the London fog incident 1952. Lancet, i, 336–8.
Lopez, A.D. (1998). Counting the dead in China. Measuring tobacco’s impact in the developing world. British Medical Journal, 317, 1399–400.
McConnell, J. (1998). WHO’s tuberculosis research initiative. Lancet, 351, 852.
McMichael, A.J. and Smith, K.R. (1999). Seeking a global perspective on air pollution and health. Epidemiology, 10, 1–4.
McMichael, A.J., Anderson, H.R., Brunekreef, B., and Cohen, A.J. (1998). Inappropriate use of daily mortality analyses to estimate longer term mortality effects of air pollution. International Journal of Epidemiology, 27, 450–3.
Mannino, D.M., Siegel, M., Rose, D., Nkuchia, J., and Etzel, R. (1997). Environmental tobacco smoke exposure in the home and worksite and health effects in adults: results from the 1991 National Health Interview Survey. Tobacco Control, 6, 296–305.
Markowitz, N., et al. (1997). Incidence of tuberculosis in the United States among HIV-infected persons. The Pulmonary Complications of HIV Infection Study Group. Annals of Internal Medicine, 126, 123–32.
Meidema, I., Feskens, E.J.M., Heederik, D., and Kromohout, D. (1993). Dietary determinants of long term incidence of chronic non-specific lung diseases. The Zutphen Study. American Journal of Epidemiology, 138, 37–45.
Murray, C.J.L. and Lopez, A.D. (ed.) (1996). The global burden of disease: a comprehensive assessment of mortality and disability from diseases, injuries, and risk factors in 1990 and projected to 2020. Harvard University Press, Cambridge, MA.
Murray, C.J.L. and Lopez, A.D. (1997a). Mortality by cause for eight regions of the world: Global Burden of Disease Study. Lancet, 349, 1269–76.
Murray, C.J.L. and Lopez, A.D. (1997b). Global mortality, disability, and the contribution of risk factors. Lancet, 349, 1436–42.
Murray, C.J.L. and Lopez, A.D. (1997c). Alternative projections of mortality and disability by cause 1990–2020: Global Burden of Disease Study. Lancet, 349, 1498–504.
National Research Council (1986). Environmental tobacco smoke: measuring exposures and assessing health effects. National Academy Press, Washington, DC.
Ng, T.P., Hui, K.P., and Tan, W.C. (1993). Respiratory symptoms and lung function effects of domestic exposure to tobacco smoke and cooking by gas in non-smoking women in Singapore. Journal of Epidemiology and Community Health, 47, 454–8.
Niu, S.R., et al. (1998). Emerging tobacco hazards in China: 2. Early mortality results from a prospective study. British Medical Journal, 317, 1423–4.
Nyangulu, D.S., et al. (1997). Tuberculosis in a prison population in Malawi. Lancet, 350, 1284–7.
Omenn, G.S., et al. (1996). Effects of a combination of beta carotene and vitamin A on lung cancer and cardiovascular disease. New England Journal of Medicine, 334, 1150–5.
ONS (Office for National Statistics) (1997). The health of adult Britain 1841–1994, Vol. 2. Series DS No. 13. HMSO, London.
Ostro, B.D. and Rothschild, S. (1989). Air pollution and acute respiratory morbidity: an observational study of multiple pollutants. Environmental Research, 50, 238–47.
Pablos-Mendez, A., et al. (1998). Global surveillance for anti-tuberculosis-drug resistance. New England Journal of Medicine, 338, 1641–9.
Peters, J., et al. (1996). Effects of an ambient air pollution intervention and environmental tobacco smoke on children’s respiratory health in Hong Kong. International Journal of Epidemiology, 25, 821–8.
Peto, R. (1998). Mortality from breast cancer in UK has decreased suddenly. British Medical Journal, 317, 476–7.
Peto, R., Lopez, A.D., Boreham, J., Thun, M., and Heath, C. Jr (1994). Mortality from smoking in developed countries 1950–2000. Oxford University Press.
Petty, T.L. and Weinmann, G.G. (1997). Building a national strategy for the prevention and management of and research in chronic obstructive pulmonary disease: National Heart, Lung and Blood Institute workshop summary. American Journal of Medical Association, 277, 246–53.
Pope, C.A. III (1989). Respiratory disease associated with community air pollution and a steel mill, Utah Valley. American Journal of Public Health, 79, 623–8.
Pope, C.A. and Xu, X. (1993). Passive cigarette smoke, coal heating, and respiratory symptoms of nonsmoking women in China. Environmental Health Perspective, 101, 314–16.
Pope, C.A. III, Schwartz, J., and Ransom, M.R. (1992). Daily mortality and PM10 pollution in Utah Valley. Archives of Environmental Health, 47, 211–17.
Pope, C.A. III, et al. (1995) Particulate air pollution as a predictor of mortality in a prospective study of US adults. American Journal of Respiratory and Critical Care Medicine, 151, 669–74.
Pym, A.S., Churchill, D.R., Gleissberg, V., and Coker, R.J. (1995). Reasons for increased incidence of tuberculosis. Audit suggests that undernotification is common. British Medical Journal, 311, 570.
Raviglione, M.C., Dye, C., Schmidt, S., and Kochi, A. (1997). Assessment of worldwide tuberculosis control. WHO Global Surveillance and Monitoring Project. Lancet, 350, 1329–30.
Reichman, L.B. (1991). The U-shaped curve of concern. American Review of Respiratory Disease, 144, 741–2.
Repace, J.L., Jinot, J., Bayard, S., Emmons, K., and Hammond, S.K. (1998). Air nicotine and saliva cotinine as indicators of workplace passive smoking exposure and risk. Risk Analysis, 18, 71–83
Reyes, H. and Coninx, R. (1997). Pitfalls of tuberculosis programmes in prisons. British Medical Journal, 315, 1447–50.
Royal College of Physicians of London (1992). Smoking and the young: a report of a working party of the Royal College of Physician. Royal College of Physicians, London.
Rupp, J.P. and Billings, D.M. (1990). Tobacco industry documents. Privileged and confidential attorneys work product, February 14, 1990. Asia ETS consultant status report. http://www.smokescreen.org/documents.
Sallie, B.A., Ross, D.J., Meredith, S.K., and McDonald, J.C. (1994). SWORD ’93 Surveillance of work-related and occupational respiratory disease in the UK. Occupational Medicine, 44, 177–82.
Samah, A.A. (1992). Investigation into the haze episodes in the Kelang Valley, Malaysia. In Association of South East Asian institutions of higher learning seminar proceedings. The role of the ASAIHL in combating health hazards of environmental pollution (ed. A.J. Hedley et al.), pp. 221–7. University of Hong Kong.
Samet, J.M. and Jaakkola, J.J.K. (1999). The epidemiological approach to investigating outdoor air pollution. In Air pollution and health (ed. S.T. Holgate, J.M. Samet, H.S. Koren, and R.L. Maynard), pp.431–60. Academic Press, London.
Sandford, A.J., Weir, T.D., and Paré, P.D. (1997). Genetic risk factors for chronic obstructive pulmonary disease. European Respiratory Journal, 10, 1380–91.
Schouten, J.P., Vonk, J.M., and de Graaf, A. (1996). Short term effects of air pollution on emergency hospital admissions for respiratory disease: results of the APHEA project in two major cities in The Netherlands, 1977–89. Journal of Epidemiology and Community Health, 50, S22–9.
Schwartz, J. (1989). Lung function and chronic exposure to air pollution: a cross-sectional analysis of NHANES II. Environmental Research, 50, 309–21.
Schwartz, J. (1993). Air pollution and daily mortality in Birmingham, Alabama. American Journal of Epidemiology, 137, 1136–47.
Schwartz, J. and Weiss, S.T. (1990). Dietary factors and their relationship to respiratory symptoms: NHANES II. Environmental Research, 50, 309–21.
Schwartz, J. and Weiss, S.T. (1994). Relationship between dietary vitamin C intake and pulmonary function in the first national health and nutrition examination survey (NHANES I). American Journal of Clinical Nutrition, 59, 110–14.
Schwartz, J., Spix, C., Wichmann, H.E., and Malin, E. (1991). Air pollution and acute respiratory illness in five German communities. Environmental Research, 56, 1–14.
SCOTH (Scientific Committee on Tobacco and Health) (1998). Report of the Scientific Committee on Tobacco and Health (SCOTH). Department of Health. HMSO, London.
Scott, J.A. (1963). The London fog of 1962. Medical Officer, 109, 250–3.
Siegel, M., Husten, C., Merritt, R.K., Giovino, G.A., and Eriksen, M.P. (1995). Effects of separately ventilated smoking lounges on the health of smokers: is this an appropriate public health policy? Tobacco Control, 4, 22–9.
Smit, H.A., Grievink, L., and Tabak, C. (1999). Dietary influences on chronic obstructive lung disease and asthma: a review of the epidemiological evidence. Proceedings of the Nutrition Society, 58, 309–19.
Smith, G.D. and Phillips, A.N. (1996). Passive smoking and health: should we believe Philip Morris’s ‘experts’? British Medical Journal, 313, 929–33.
South East Asian Medical Information Center/International Medical Foundation of Japan (1998). SEAMIC Health Statistics. South East Asian Medical Information Center, Tokyo.
Soyseth, V., et al. (1995). Relation of exposure to airway irritants in infancy to prevalence of bronchial hyper-responsiveness in schoolchildren. Lancet, 345, 217–20.
Squire, S.B. and Wilkinson, D. (1997). Strengthening ‘DOTS’ through community care for tuberculosis. British Medical Journal, 315, 1395–6.
Stationery Office (1998). Smoking kills: a White Paper on tobacco. HMSO, London.
Stellman, S.D., Boffetta, P., and Garfinkel, L. (1988). Smoking habits of 800,000 American men and women in relation to their occupations. American Journal of Industrial Medicine, 13, 43–58.
Strachan, D.P., Cox, B.D., Erzinclioglu, S.W., Walters, E.D., and Whichelow, M.J. (1991). Ventilatory function and winter fresh fruit consumption in a random sample of British adults. Thorax, 46, 624–9.
Tam, A.Y.C., et al. (1994) Bronchial responsiveness in children exposed to atmospheric pollution in Hong Kong. Chest, 106, 1056–60.
Thurston, G.D. and Ito, K. (1999). Epidemiological studies of ozone exposure effects. In Air pollution and health (ed. S.T. Holgate, J.M. Samet, H.S. Koren, and R.L. Maynard), pp. 485–510. Academic Press, London.
Thurston, G.D. and Kinney, P.L. (1995). Air pollution epidemiology: considerations in time series analyses. Inhalation Toxicology, 7, 71–83.
Tocque, K., et al. (1998). Long-term trends in tuberculosis. Comparison of age-cohort data between Hong Kong and England and Wales. American Journal of Respiratory and Critical Care Medicine, 158, 484–8.
Touloumi, G., Samoli, E., and Katsouyanni, K. (1996). Daily mortality and ‘winter type’ air pollution in Athens, Greece—a time series analysis within the APHEA project. Journal of Epidemiology and Community Health, 50, S47–51.
UKHSC (UK Health and Safety Commission) (1999). Proposal for an approved code of practice on passive smoking at work. HSE Books, Sudbury.
USCEPA (US California Environmental Protection Agency) (1997). Health effects of exposure to environmental tobacco smoke. California Environmental Protection Agency, Office of Environmental Health Hazard Assessment, Sacramento, CA.
USDHEW (US Department of Health, Education and Welfare) (1986). The health consequences of involuntary smoking. A report of the Surgeon General. US Department of Health, Education and Welfare, Washington, DC.
USEPA (US Environmental Protection Agency) (1992). Respiratory health effects of passive smoking: lung cancer and other disorders. US Environmental Protection Agency, Washington, DC.
USOSHA (US Occupational Safety and Health Administration (1994). Notice of proposed rulemaking; notice of informal public hearing. Federal Register, 5 April 5 1994, 29 CFR Parts 1910, 1915, 1926, and 1928.
van Ede, L., Yzermans, C.J., and Brouwer, H.J. (1999). Prevalence of depression in patients with chronic obstructive pulmonary disease: a systematic review. Thorax, 54, 688–92.
Viegi, G., et al. (1991). Prevalence rates of respiratory symptoms in Italian general population samples exposed to different levels of air pollution. Environmental Health Perspective, 94, 95–9.
Volmink, J. and Garner, P. (1997a). Promoting adherence to tuberculosis treatment. In Infectious diseases module of the Cochrane database of systematic reviews (ed. P. Garner et al.). Cochrane Library, Update Software, Oxford.
Volmink, J. and Garner, P. (1997b). Systematic review of randomised controlled trials of strategies to promote adherence to tuberculosis treatment. British Medical Journal, 315, 1403–6.
Wagner, G.R. (1998). Preventing pneumoconioses and eliminating silicosis: opportunities and illusions. In Advances in the prevention of occupational respiratory diseases (ed. K. Chiyotani, Y. Hosoda, and Y. Aizawa), pp. 3–11. Elsevier, Amsterdam.
Walters, S., Phupinyokul, M., and Ayres, J. (1995). Hospital admission rates for asthma and respiratory disease in the West Midlands: their relationship to air pollution levels. Thorax, 50, 948–54.
Ware, J.H., et al. (1986). Effects of ambient sulphur oxides and suspended particles on respiratory health of preadolescent children. American Review of Respiratory Disease, 133, 834–42.
Wares, D.F. and Clowes, C.I. (1997). Tuberculosis in Russia. Lancet, 350, 957.
Weiland, S.K., Mundt, K.A., Ruckmann, A., and Keil, A. (1994). Self-reported wheezing and allergic rhinitis in children and traffic density on the street. Annals of Epidemiology, 4, 243–7.
Wells, A.J. (1998). Heart disease from passive smoking in the workplace. Journal of the American College of Cardiology, 31, 1–9.
White, J.R., Froeb, H.F., and Kulik, J.A. (1991). Respiratory illness in nonsmokers chronically exposed to tobacco smoke in the work place. Chest, 100, 39–43.
Whittemore, A.S. and Korn, E.L. (1980). Asthma and air pollution in the Los Angeles area. American Journal of Public Health, 70, 687–96.
WHO (World Health Organization) (1994). TB—a global emergency. WHO report on the TB epidemic. WHO, Geneva.
WHO (World Health Organization) (1996). Tuberculosis in the era of HIV. A deadly partnership. WHO/TB/96.204. WHO, Geneva.
WHO (World Health Organization) (1997a). Tobacco or health: a global status report. WHO, Geneva.
WHO (World Health Organization) (1997b). health and environmental in sustainable development. five years after the earth summit. WHO/EHG/97.8. WHO, Geneva.
WHO (World Health Organization) (1998). Guidelines for controlling and monitoring the tobacco epidemic. WHO, Geneva.
WHO (World Health Organization) (1999a). The world health report 1999. WHO, Geneva.
WHO (World Health Organization) (1999b). The framework convention on tobacco control—a primer. WHO, Geneva.
WHO (World Health Organization) (1999c). Occupational health: ethically correct, economically sound. Fact Sheet No. 84. http://www.who.int/inf-fs/en/fact084.html. WHO, Geneva.
WHO (World Health Organization) (1999d). International consultation on environmental tobacco smoke (ETS) and child health, 11–14 January 1999. NCD/TFI/ETS/99.2. WHO, Geneva.
Wise, J. (1998). WHO identifies 16 countries struggling to control tuberculosis. British Medical Journal, 316, 957.
Wjst, M., et al. (1993). Road traffic and adverse effects on respiratory health in children. British Medical Journal, 307, 596–600.
Wong, C.M., et al. (1998). Comparison between two districts of the effects of an air pollution intervention on bronchial responsiveness in primary school children in Hong Kong. Journal of Epidemiology and Community Health, 52, 571–8.
Wong, C.M., Hu, Z.G., Lam, T.H., Hedley, A.J., and Peters, J. (1999). Effects of ambient air pollution and environmental tobacco smoke on respiratory health of non-smoking women in Hong Kong. International Journal of Epidemiology, 28, 859–64.
World Bank (1993). World development report 1993. Oxford University Press.
World Bank (1999). Curbing the epidemic governments and the economics of tobacco control. World Bank, Washington, DC.
World Cancer Research Fund (1997). Food, nutrition and the prevention of cancer: a global perspective. World Cancer Research Fund and American Institute for Cancer Research, Washington, DC.
Xu, X.P., Dockery, D.W., and Wang, L.H. (1991). Effects of air pollution on adult pulmonary function. Archives of Environmental Health, 46, 198–206.
Xu, X., Gao, J., Dockery, D.W., and Chen, Y. (1994) Air pollution and daily mortality in residential areas of Beijing, China. Archives of Environmental Health, 49, 216–22.
Zumla, A. and Grange, J.M. (1999). The ‘global emergency’ of tuberculosis. Proceedings of the Royal College of Physicians of Edinburgh, 29, 104–15.
Zwarenstein, M., Schoeman, J.H., Vundule, C., Lombard, C.J., and Tatley, M. (1998). Randomised controlled trial of self-supervised and directly observed treatment of tuberculosis. Lancet, 352, 1340–3.