Click for next page ( 20


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 19
Chapter 1 Introduction

OCR for page 19

OCR for page 19
1 INTRODUCTION HISTORICAL BACKGROUND AND PURPOSE OF THE REPORT Official registration of deaths began in England in 1837 under the supervision of William Farr, marking what many believe to be the birth of the modern era in epidemiology. The new era was only 8 years old in 1845, when analysis of the mortality data raised questions about the limitations of epidemiology in studying the problem of air pollution. The Registrar General's Fifth Report, in 1845, contained the following statement: That smoke is irritating to the air-passages, injurious to health, and one of the causes of death, to which the inhabitants of towns are more exposed than the inhabitants of the country, is probable; but if the effect were very considerable it would be most evident in the dense fogs, when the atmosphere is loaded with smoke. . . . Now we have never observed any connection between the increase of the mortality and the London fogs. Despite the absence of positive findings, Farr and his associates were inclined to believe that smoke in the ambient air was injurious to health. Current concerns about the role of epidemiology in the study of air pollution resemble theirs to a striking degree. Can epidemiologic studies detect the feared consequences in populations? Are these studies sensitive enough? How do we interpret apparently negative results? These questions continue to arise because the ability of epidemiologic studies to elucidate the relationship 21

OCR for page 19
between air pollution and health at any particular time depends on a dynamic balance among three elements: the general magnitude of exposure to air pollution, the magnitude of the effects of exposure relative to back- ground disease rates, and the strength of the epidemi- ologic tools available. When infectious disease rates began to decrease in England around the middle of the nineteenth century and pollution from rapid industrial- ization began to increase, epidemiology became effective in demonstrating that bad air affected the public's health. In fact, its impact could be estimated reasonably accurately by counting foggy days in London and counting the excess deaths that occurred on such days. The first two historical elements have changed enough to bring new demands on existing epidemiologic techniques, as well as a challenge to reassess their limits. This report provides such an assessment and indicates some opportunities for expanding the limits and for using I.. epidemiologic studies effectively in an overall program of research on air pollution. Until 1960, most epidemiologic studies of air pollution examined the impact of short-term episodes of large pollutant exposures. Studies of the air pollution "disasters" in the Meuse Valley during December 1930, Donor a in November 1948, and London in December 1952 established that sudden peaks of exposure to coal combus- tion products and industrial pollution, superimposed on persistently high relative levels of exposure, could cause excess deaths among persons in the community who suffered from chronic cardiorespiratory diseases.7 These epidemiologic demonstrations that air pollution can, at its worst, increase mortality due to respiratory diseases led to major clean air legislation in the United States and abroad. In the 1960s, new problems for study began to emerge as the concerns of researchers expanded from the gross effects of short-term peaks toward effects associated with long-term, low-dose exposures. At the same time, hypotheses that called for separation of the effects of individual pollutants began to receive greater attention. 22 -

OCR for page 19
CURRENT MAGNITUDE OF OUTDOOR AIR POLLUTION IN THE UNITED STATES The magnitude of ambient air pollution exposure might be decreasing, at least in the developed West. Data recently published by the Environmental Protection Agency (EPA) from a national monitoring survey show substantial decreases in average urban concentrations of sulfur dioxide, SO2, carbon monoxide, CO, and lead, Pb, in the United States as a whole between 1975 and 1983.~5 Furthermore, since passage of the Clean Air Act in 1970, concentrations of total suspended particles (TSP) have also decreased substantially in many areas. The current picture is not entirely bright, however. Nitrogen oxides and ozone, O3, have proved to be relatively harder to control. Although slight decreases in average urban concentrations of these air contaminants did occur during 1975-1983, some areas show no improvement, and O3 concentrations actually increased by 12% from 1982 to 1983, owing to the generation of increased amounts of O3 precursors by economic recovery and weather conditions favoring O3 formation. Regardless of such short-term trends, exposure to some pollutants, such as O3, might now be spread over larger populations, because of long-range transport phenomena and rapid urban growth. Other problems of increasing concern involve pollutants for which there are no standards, such as fine aerosols, sulfates, and woodsmoke. Nevertheless, in most areas of the country, the data as a whole show much progress in reducing the air pollution of the 1950s and 1960s. Therefore, we must ask what epidemiology can contribute to the study of air pollution effects that are likely to be less common and less severe than they were just 25 years ago. MAGNITUDE OF DISEASE RELATED TO POLLUTION AND OTHER CAUSES The second element to be considered is the magnitude of pollution-related disease relative to background disease. In Farr's day, air pollution effects might have been masked by epidemics of infectious disease; today, the effects might be masked by the predominance of cigarette smoking as a cause of respiratory disease. The impact of ambient air pollutants on the total respiratory disease burden in the United States must be small relative to the 23

OCR for page 19
impact of cigarette smoking, and occupational exposures might also have greater effects than pollution of ambient air. The dominance of these causal factors tends to complicate the task of air pollution studies, but does not in itself eliminate the need for them. The public health importance of air pollution can be underestimated unless three points are considered: . Air pollution might cause considerable burdens of respiratory disease, even though cigarette smoking has a greater impact. The diseases of concern are quite common. For instance, the National Center for Health Statistics estimates that in 1979-1981 there were approximately 10 million persons with chronic obstructive pulmonary disease (COLD). (For more detailed information on the magnitude of respiratory diseases in the United States during recent years, refer to the tables in Appendix A.) Even if the proportion of COPD cases attributable to air pollution were low, the absolute amount of attributable disease would be high, in proportion to other public health problems. Chapter 2 presents some of the evidence that current air pollution can cause substantial health effects. . Air pollution and cigarette smoking might have synergistic effects on the incidence of respiratory disease. Animal and epidemiologic studies suggest that air pollution exposure can multiply the risk of developing respiratory disease in smokers.~ 17 It is also plausible that air pollution aggravates conditions caused by smoking and thus leads to increases in symptoms and other expressions of morbidity. In either case, actions taken to reduce the impact of air pollution would have the concomitant effect of preventing some smoking-related illness. (This does not suggest that efforts to reduce cigarette smoking itself should be diminished.) Disease related to air pollution is relatively amenable to primary Prevention. Whatever the attributable disease burden, centralized intervention at the national or community level (e.g., catalytic converters, stack controls, and fuel substitution), although not simple or inexpensive, is likely to be more efficient than intervention at the level of individual behavior. 24

OCR for page 19
STRENGTH OF EPIDEMIOLOGIC TOOLS The third historical element, the discriminatory potential of available epidemiologic tools, has also changed. Today, "epidemiology is no more restricted to the study of striking outbreaks of disease than meteor- ology is to the study of hurricanes or astronomy to eclipses of the sun. n8 New techniques and strategies have made possible more effective approaches in air pollution epidemiology, leading to studies that are substantially different from those performed in the 1960s and 1970s. Much of this report deals specifically with the capabilities and limitations of methods and strategies of the 1980s and, as well as we can foresee them, later decades. Clean air legislation itself has begun to influence the definition of the new problems or questions for epidemiologic study. In the United States, clean air legislation is based on the view that safe or no-effect concentrations for particular categories of major pol- lutants can be estimated and incorporated into standards. (In contrast, such legislation in Great Britain is based on control of specific pollution sources.) Appendix B lists the current National Ambient Air Quality Standards. Epidemiologic methods are being invoked to substantiate these standards. This report discusses how this applica- tion further complicates the demands on epidemiology. In summary, the technical demands on epidemiologic studies of air pollution are greater than ever before. Meanwhile, the cost of obtaining each additional bit of information on the new kinds of problems has increased and will continue to increase. The questions now asked can sometimes be answered only by a coalition of engineer- ing, mathematical, and biologic scientists. The advanced technologies of these sciences, and especially their integrated use, are expensive. One of the purposes of this report is to provide decision-makers with information to assist them in planning and setting priorities for the investment of research resources. EPIDEMIOLOGY VIS-A-VIS OTHER FORMS OF RESEARCH Toxicology, clinical research, and epidemiology are the three disciplines that can contribute most to our 25

OCR for page 19
understanding of the human health effects of air pol- lution. Each has inherent advantages and limitations in studying air pollution. Each is essential, and their complementary and interactive nature must be appreciated. Animal experiments allow relatively precise setting of exposure conditions, control over factors that might confound observed dose-response relations, and study of a wide variety of specific end points, including effects at the biochemical or cellular level. For example, these studies have been particularly helpful in examining ques- tions concerning the chronic effects of O3 or nitrogen dioxide on lung structure and function. Clinical studies, which involve controlled exposure of human subjects, also offer an opportunity for precise quantification of dose and response and elimination of interfering conditions. Response to inhaled pollutants can be assessed with sophisticated physiologic techniques. These studies have provided the best means for investigation of the transient effects of pollutants in susceptible persons, such as asthmatics. The fundamental limitation of animal experiments is the difficulty of extrapolating results from animals to humans. Furthermore, the intentional simplicity of experimental conditions, which can include exposure to only one or two pollutants, rarely simulates conditions in the natural human environment. Information obtained from clinical studies is also constrained by the need for experimental simplicity. The use of relatively few subjects inhibits generalization of results to large populations. And ethical considerations generally require that controlled-exposure studies in humans be limited to investigation of acute, fully reversible effects. As most commonly defined, epidemiology is the study of the distribution and determinants of disease in popula- tions. It affords a direct means of obtaining scientific information on the impact of actual air pollution on free-living human populations. Epidemiology can be distinguished from other forms of air pollution research by the special role it plays in the formulation and evaluation of preventive strategies. The discipline may be seen as, among other things, a tool for diagnosing disease in a community, rather than an individual. Furthermore, the epidemiologic perspective starts, as 26

OCR for page 19
Doll has said, with the 10,000 cases of disease that are present, rather than with the 10,000 chemicals that might have caused them. 2 Epidemiologic data, like data from toxicology and clinical research, can sometimes elucidate disease etiology and mechanisms, but they can also provide a clear basis and guidance for public health intervention. Knowledge of specific etiologic factors is not necessarily required. Utilizing at least 80 years of observation on the effects of London's killer fogs, health authorities chose to emphasize control of visible smoke from coal combustion before the specific agents responsible for illness could be identified.) The choice was propi- tious, inasmuch as analyses have since shown that fine particles appear to have played a greater role in causing episodic excesses in mortality than SO2.9 This direct use of epidemiology to develop preventive measures, without reliance on knowledge of specific causal paths, must be kept in mind in evaluating the potential contribution of epidemiologic studies of air pollution. DIFFICULTIES IN EPIDEMIOLOGIC STUDIES OF AIR POLLUTION - Nearly all epidemiologic studies on air pollution are observational, rather than experimental. Observational studies of air pollution face several major difficulties, which are related to the three types of variables involved: . population, is difficult to assess. Air pollution is a highly complex phenomenon that involves many species of pollutants undergoing continuous chemical and physical transformation. Simultaneous exposure to several major pollutants is the rule. Determining who is exposed, to what, and to how much is not easy, particularly for chronic exposure to varying concentrations. - . Effects of air pollution are difficult to assess. Earlier research on air pollution attempted to identify a "smog lesion.--the descriptive pathology of an effect peculiar to air pollution. Those attempts were fruitless, and it is now recognized that air pollution is one of many factors involved in multicausal disease processes that ultimately converge to a few pathways of organ dysfunction. Air pollution can play either a causal or an aggravating role in these disease processes. 27

OCR for page 19
Discerning who has suffered an effect specifically due to air pollution is not easy, nor is estimation of the extent of the effect. These tasks are all the more complicated when a chronic disease with a vague, insidious onset is considered many years after first exposure to the presumed causal agent. Many factors confound the exposure-effect relationship. An epidemiologic study of air pollution might be rather straightforward if it had to account only for an exposure and an effect. However, the observation that a given exposure and a given pathologic condition are found together in populations can also be explained by confounders, factors that are unevenly distributed among exposure groups and that include occupational exposure, weather conditions, pre-existing health con- ditions, and smoking habits. These potential confounders are numerous. Identifying them, designing studies to reduce their effects, estimating any effects that remain, and adjusting the measurement of the exposure-response relation accordingly are among the most difficult tasks in any air pollution study. Those three fundamental problems lead to two subsidiary problems worthy of note. First, the combination of possible errors in the estimation of key variables and the need to detect a relatively small (but not necessarily unimportant) effect create a problem with respect to the sensitivity of studies. An analogy with a diagnostic tool, such as the microscope, applies here, if we think of limits to the resolving power of lenses. To discern the broad but relatively faint image of the air pollution health effect among surrounding images, epidemiologic studies must either reduce error in the estimation of variables used in the analysis (improve the optical quality of the lenses) or maintain a very large sample size (increase the magnification). When studies are insensitive, failure to see effects of air pollution cannot properly be interpreted to mean that they are not there. m e second subsidiary problem lies in the difficulty of generalizing epidemiologic results from one population to another. Because the distribution of numerous con- founding factors can vary widely among populations, epidemiologic results might have only local applicability, or at least must be extended cautiously. For instance, 28

OCR for page 19
failure to account for the existence of sensitive subgroups can distort the prediction of response in a community. This report emphasizes the importance of specifying the research questions for which epidemiology is to provide answers. No research tool or method, however precise its measurements or powerful its inferences, can achieve much if it is applied to inappropriate questions. Generalizations regarding "air pollutants, n Epidemiologic studies," Adverse health effects, n or "populations" have limited utility, given the complex and diverse categories implied by each of these terms. Furthermore, questions posed for epidemiologic study must be asked in the proper form; the right question asked in the wrong way might be unanswerable. SCOPE OF THE REPORT The subject of this report is epidemiology and its methods; other forms of research on air pollution are discussed only as they are related to epidemiologic endeavors. The report does not address the question, "How important is air pollutions Therefore, it does not include a complete critical review of the epidemiologic literature on air pollution. Several excellent and recent reviews are available, and the inclusion here of yet another would divert attention from issues more pertinent to the charge. 3 ~ 2 - ~ ~ ~ ~ The report is also not intended as a primer on epidemiologic principles, although it discusses some specific principles related to the problems of studying air pollution. The reader should consult current textbooks on epidemiologic methods for more detailed discussion.4~ 6 Because decisions about the future control of air _ pollutants must be based on more sensitive ena points than have generally been used in the past, we emphasize indexes of morbidity, rather than mortality. we stress effects on the respiratory system, because that system is the most vulnerable to the ubiquitous air pollutants and studies of respiratory morbidity encompass most of the serious questions regarding methods. Furthermore, The report focuses on problems associated with common area sources of pollution, rather than those associated 29

OCR for page 19
with point sources of contamination that occur in specific neighborhoods--hence the relative lack of emphasis on the so-called toxic air pollutants, which are currently of great concern because of, for example, the disastrous leak of chemicals in Bhopal, India. This report takes a unified approach to indoor and outdoor exposure to polluted air. Most future epidemiologic studies on air pollution should adopt a perspective that accounts for pollutant exposure in both environments. The report deals primarily with four of the EPA criteria air pollutants: SO2, particles, nitrogen oxides, and O3. CO and Pb, the other criteria pollutants, are given less attention, because their primary effects are not on the respiratory system and because the problems they pose for epidemiologists are quite different and generally more straightforward. CONTENTS OF THE REPORT Epidemiologic inquiry in environmental health involves two major steps: collection of data concerning effects, exposures, and modifying factors; and drawing of infer- ences from the data collected. The next two chapters deal with issues of data collection. Chapter 2 defines the health effects that concern us and the implications of current disease models for epidemiologic studies. It also covers the various tools available for assessing effects, their constraints, the quality of data, and opportunities for the application of new tools, such as biologic markers. Chapter 3 examines the quality of exposure data and various means for providing such data in epidemiologic studies. It emphasizes the importance of indoor air pollution and describes opportunities for improving exposure assessment. It also identifies pollutants and pollutant sources that are likely to be of interest in future studies. Chapter 4 develops concepts and strategies appropriate to the planning of epidemiologic studies. It emphasizes the need for epidemiologists to frame their questions carefully and to consider the sources of error that can interfere with obtaining useful results. It describes the various types of epidemiologic studies and the 30

OCR for page 19
circumstances for which they are most appropriate. It notes special opportunities for studies and discusses some aspects of the development of the proper milieu for studies and the relationship of epidemiology to regulation. In Chapter 5, the Committee applies various designs and strategies to current research questions related to air pollution. Such questions include those on chronic and acute respiratory effects and on effects of specific air pollutants. Some years of effort on the part of EPA and other agencies have succeeded in reducing air pollution in the United States. Nonetheless, air pollution remains a public health problem. Epidemiology can contribute to the understanding and eventual solution of this problem. Chapter 6 presents the Committee's conclusions and offers recommendations designed to optimize the epidemiologic study of air pollution. REFERENCES 1. Ashby, E., and M. Anderson. The Politics of Clean Air. Monographs on Science, Technology and Society. Oxford: Clarendon Press, 1981. 178 pp. 2. Doll, R. Relevance of epidemiology to policies for the prevention of cancer. J. Occup. Med. 23:601-608, 1981. 3. Holland, W.W., A.E. Bennett, and I.R. Cameron. Health effects of particulate pollution: Reappraising the evidence. Am. J. Epidemiol. 110:525-659, 1979. 4. Kleinbaum, D.G., L.L. Rupper, and H. Morgenstern. Epidemiologic Research: Principles and Quantitative Methods. Belmont, Cal.: Lifetime Learning Publications, 1982. 529 pp. Last, J.M., Ed. Maxcy-Rosenau Public Health and Preventive Medicine. 11th ed. New York: Appleton-Century-Crofts, 1980. 1926 pp. 31

OCR for page 19
6. Lilienfeld, A.M., and D.E. Lilienfeld. Foundations of Epidemiology. 2nd ed. New York: Oxford University Press, 1980. 375 pp. 7. Lipfert, F.W. Sulfur oxides, particulates, and human mortality: Synopsis of statistical correlations. J. Air Pollut. Control Assoc. 30:366-371, 1980. 8. MacMahon, B., and T.F. Pugh. Principles and Methods, p. 1. .... Brown and Company, 1970. 9. Epidemiology: Boston: Little, Mazumdar, S., H. Schimmel, and I.T.T. Higgins. Relation of daily mortality to air pollution: An analysis of 14 London winters, 1958/59-1971/72. Arch. Environ. Health 37:213-220, 1982. Rall, D.P. Review of the health effects of sulfur oxides. Environ. Health Perspect. 8:97-121, 1974. 11. Registrar General. 5th Annual Report for the year 1843, Second Edition, p.415. London: H.M.S.O., 1945. 12. Shy, C.M., J.R. Goldsmith, J.D. Hackney, M.D. Lebowitz, and D.B. Menzel. Health Effects of Air Pollution. New York: American Thoracic Society, Medical Section of the American Lung Association, 1978. 48 pp. 13. U.S. Environmental Protection Agency. Air Quality Criteria for Oxides of Nitrogen. Publication No. EPA-600/8-82-026F. Research Triangle Park, N.C.: U.S. Environmental Protection Agency, 1982. 790 pp 14. U.S. Environmental Protection Agency. Air Quality Criteria for Particulate Matter and Sulfur Oxides. Publication No. EPA-600/8-82-029a. Research Triangle Park, N.C.: U.S. Environmental Protection Agency, 1982. 204 pp. 15. U.S. Environmental Protection Agency. National Air Quality and Emissions Trends Report, 1983. Publication No. EPA-450/4-84-029. Research Triangle Park, N.C.: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, 1985. 112 pp. 32 .

OCR for page 19
16. U.S. National Center for Health Statistics. National Health Information Survey (NHIS) unpublished data from 1979-1981. Hyattsville, Md.: U.S. Department of Health and Human Services, National Center for Health Statistics, 1985. 17. Vena, J.E. Air pollution as a risk factor in lung cancer. Am. J. Epidemiol. 116:42-56, 1982. 18. Whittemore, A.S. Air pollution and respiratory disease. Ann. Rev. Pub. Health 2:397-429, 1981. 33

OCR for page 19