Tobacco Smoke Pollution

3 downloads 0 Views 2MB Size Report
Table 7-1 43 Chemical Compounds Identified in Tobacco Smoke for which There is. "Sufficient ...... mainstream smoke and sidestream smoke of different types of ...
w

Tobacco Smoke Pollution James L. Repace

Inhaling the smoke from cigarettes, pipes. and cigars delivers nicotine to the brain more quickly and emcently than chewing tobacco leaf (Chapter 1). However, the practice of burning tobacco leaves indoors exposes other people to indoor air pollution from tobacco combustion products containing many chemicals known to be harmful to human health. Although society, in the interest of public health, has long imposed quality standards for food, water, and indeed, for outdoor air. it has been slow to require that the indoor air be of a quality that will prevent morbidity and mortality. In general, the same amount of contaminant deposited on the lung surface from inhaled air that we breathe has greater potential for harm than an equal amount ingested in food or water, due to differences in absorption efficiency between the pulmonary and gastrointestinal membranes. For example, in a healthy adult, 5% of a gram of lead from a chip of accidentally ingested paint will be absorbed, while 95% of the same amount of lead inhaled as a fume from automobile exhaust will be absorbed. It is common for there to be strict controls on low-dose general population exposures to toxic agents which have only been proved to be harmful with long-term exposure at high doses. Pollutants such as asbestos, benzene, and radioactivity fall into this category. The 2.5 million deaths per year worldwide caused by the cigarette qualifies tobacco smoke as extraordinarily harmful at high doses (Peto and Lopez 1990). It is astonishing that this fact alone has not led to stringent controls on tobacco smoke pollution. Other environmental agents with the poten-

http://legacy.library.ucsf.edu/tid/ylj10i00/pdf

tial to cause far smaller amounts of harm have been banned from food, water, and air. For example, in 1988, Chilean grapes were banned from U.S. markets because cyanide contamination found on two grapes was of the order of a few percent of that delivered by smoking one cigarette. In estimating the magnitude of a public health risk from an environmental agent, a technique called risk assessment is typically employed by public health agencies. Risk assessment often relies on available data coupled with the use of models to estimate the expected magnitude of a public health risk. If the estimated risk is significant, risk management, generally in the form of regulation, is employed to control or eliminate the hazard. Risk assessment is also used to ascertain uncertainties in the estimated risk, as well as to point out directions for future research. Risk assessment has four main components: hazard assessment, exposure assessment, dose-response determination, and risk characterization. This chapter on the pollution of indoor air by tobacco smoke will follow this outline.

H A Z A R D ASSESSMENT O F T O B A C C O SMOKE P O L L U T I O N

Large-dose inhalation exposure to tobacco smoke (ordinary smoking) is a major cause of coronary heart disease, atherosclerotic peripheral vascular disease, lung and laryngeal cancer, oral cancer, esophageal cancer, emphysema, chronic bronchitis, intrauterine growth retardation, and low birth weight. In 129

to

CXI

o

8 §5 5

NOTICE Miis material mty be pr»t*cte* by copyrigtit »'iw (Titis 1? U.S. CG^-1.

2505624647

130

THE DISEASE OF NICOTINE ADDICTION

Table 7-1 43 Chemical Compounds Identified in Tobacco Smoke for which There is "Sufficient Evidence"' of Carcinogenicity in Humans or Animals acetaldehyde acryionitrile arsenic benz(a)anthracene benzene benzo(a|pyrene benzo(b)fiuoranthene benzo{ k)fluoranthene cadmium chromium VI DDT dibenz(a.h)acridine dibenz(a,j)acridine dibenz(a.h)anihracene

dibenzo IN) J* O Ol

o

2505624650

TOBACCO SMOKE POLLUTION

smoke pollution constitutes 83% of the RSP levels. (Compare the value just calculated to those shown in Figure 7-1. Data points, E, HL K. L, M and N are typical restaurants: B and V are reception halls; J is a hospital waiting room; I is a bowling alley; D, G, and T are bingo games; while O is a sports arena [smoking discouraged, but not enforced]; B is a lodge hall dinner-dance; C and Q are bars; F is a nightclub; and A is a private home during a party.) The dashed lines show the calculated air exchange rates. Given the nature of tobacco smoke particulate matter shown in Table 7-1, should such exposure be considered healthy for the servers or patrons? If the restaurant owner decides to save money by decreasing the air exchange rate to 5 cubic feet of outdoor makeup air per minute per occupant (1.75 air changes/h, a level still common under many "energy-efficient" building ventilation codes), the RSP level after one-half hour is 319 Mg/m\ and after 3 hours increases to a near steady-state value of 742 Mg/tn3! (Compare this calculation to data point B in Figure 7-1, taken at a dinner-dance.) By way of comparison, the national outdoor air quality standard for inhalable particles ( < 10 ^m in diameter) is 50 ng/m* (shown in Figure 7-1). Pollution of air in buildings by tobacco smoke is so pervasive that many people are unaware that they are even exposed. This is probably due to the fact that many people believe that because no one is smoking in their immediate environment, they are not being exposed to tobacco smoke. For example, Jarvis and Russell (1984, 1987) showed that among 100 U.K.. nonsmokers, only 12 had undetectable cotinine levels, although nearly half reported "no exposure" (IARC 1987). Tobacco smoke recirculated in the central air systems of commercial buildings may expose workers unknowingly, accounting for this result (Repace and Lowrey 1987b). Similar results were reported in the United States by Cummings and associates (1989, 1990): of 663 nonsmokers whose urinary cotinine levels were studied, only 76% reported exposure to ETS; however, 91% had detectable levels. The mean cotinine level in the nonsmokers was 8.84 ng/ml (0.7% of the mean value reported for 130

http://legacy.library.ucsf.edu/tid/ylj10i00/pdf

133

smokers in the same study), and the range was 0-85 ng/ml; 92% of cotinine values were less than 20 ng/ml. Cotinine levels tended to increase with the number of reported exposures to ETS. Based on time-budget studies, most people spend the bulk of their time in just two microenvironments: home and work. Since many persons exposed at home are also exposed at work, those exposed both at home and at work represent a most-exposed group of passive smokers (Repace and Lowrey 1985b). Moreover, exposure probabilities are high at work: for example, in the Cummings study (1990), 77% were exposed in the workplace, and 22% were exposed while at home. Similar conclusions about the importance of these two microenvironments were reported by Riboli and colleagues (1990), Cummings et al. (1990) reported that 84% of subjects who did not live with a smoker had detectable cotinine levels. This finding has important implications for epidemiological studies of passive smoking and disease. Most such studies use domestic exposure as a surrogate for total exposure to passive smoking, and unaccounted for exposures outside the home may confound an actual association, since this exposure misclassification tends to bias the mortality ratio toward unity. Insofar as public health measures are concerned, the two most important sites for interventions to control tobacco smoke pollution are the home and the workplace: the former by public information, the latter by public information and regulation. DOSE-RESPONSE DETERMINATION Heart Disease

Wells (1988), Glantz and Parmley (1991), and Steenland (1992) have reviewed the evidence that passive smoking increases the risk of death from heart disease. In 10 epidemiological studies on the risk of death from ischemic heart disease or myocardial infarction among nonsmokers living with smokers, the exposed groups experienced an overall 20-30% higher risk than the unexposed groups. Glantz and Parmley (1991) note that these epidemiological studies are complemented by a variety of physiological

134

and biochemical data from human studies that show that exposure to tobacco smoke pollution adversely affects platelet function and damages anerial endothelium in a way that increases the risk of heart disease. These observations have recently been experimentally confirmed in rabbits exposed to passive smoking by Zhu and colleagues (1992). Glantz and Parmley (1991) also reviewed evidence that exposure exerts significant effects on exercise capability of both normal persons and those with heart disease by affecting the body's ability to deliver and utilize oxygen. Further, they report that in animal experiments, exposure to ETS also depresses cellular respiration at the mitochondrial level, and that polycyclic aromatic hydrocarbons in ETS also accelerate, and may initiate, the development of atherosclerotic plaque. Lung Cancer

Several official bodies have addressed the question of whether passive smoking causes lung cancer. In 1986, the U.S. Surgeon General concluded that "involuntary smoking is a cause of disease, including lung cancer, in healthy nonsmokers" (USDHHS 1986 p 7). Also in 1986, the National Research Council (NRC) concluded that "exposure to ETS increases the incidence of lung cancer in nonsmokers" (NRC 1986 p 10). The NRC estimated that the risk of lung cancer was increased by roughly 30% for nonsmoking spouses of smokers relative to nonsmoking spouses. The following year, the International Agency for Research on Cancer (IARC) stated that "exposure to tobacco smoke gives rise to some risk of cancer" (IARC 1987). In 1992. the U.S. Environmental Protection Agency (EPA) reviewed the original 13 epidemiological studies which had been the basis for these initial findings plus 17 more, for a total of 30 studies of passive smoking and lung cancer, and concluded that ETS is a "class A," or "known human carcinogen" (USEPA 1992). Finally, in 1991. the National Institute for Occupational Safety and Health issued a report which concluded that ETS meets the criteria for classification as an "occupational carcinogen," and called atten-

http://legacy.library.ucsf.edu/tid/ylj10i00/pdf

THE DISEASE OF NICOTINE ADDICTION

tion to the "possible association between exposure to ETS and increased risk of heart disease in nonsmokers" (NIOSH 1991). With respect to quantitation of lung cancer risk. EPA used actual epidemiological data and vital statistics to estimate the number of nonsmokers affected. The EPA report concluded that passive smoking is causally associated with lung cancer in adults. This report concluded that, following the EPA guidelines for carcinogen risk assessment, ETS causes approximately 3,000 ( r 2.000) lung cancer deaths annually in the United States in never-smokers and ex-smokers. Pediatric Diseases Tobacco smoke pollution causes the children of parents who smoke to have up to 300.000 cases annually of bronchitis and pneumonia, increased prevalence of asthma, cough, sputum production, wheeze, and middle ear effusions, as well as up to 1,000,000 exacerbated cases of existing asthma annually. (NRC 1986: USDHHS 1986; USEPA 1992). Further, a recent study (Janerich et al. 1990) concluded that exposure to tobacco smoke pollution during childhood was associated with increased susceptibility to lung cancer in adulthood. Dose-Response Relationship Suppose it were possible to measure exactly the dose of all lung carcinogens breathed in during passive smoking. Suppose further that the number of lung cancer deaths induced could be precisely measured. A population-based dose-response function could then be defined, yielding the number of lung cancer deaths per year in the population at risk, per average number of milligrams of tobacco smoke carcinogens inhaled per day. Although it is not possible to actually measure the total dose of lung carcinogens from tobacco smoke polluted atmospheres to the nonsmoking population, from measured RSP levels and mathematical models, it is possible to estimate the exposure the nonsmoking population has to the particulate phase of tobacco smoke. Also, although it is also not possible to precisely measure the population response, that is. the number of

ro en o

01 O)

ro

•P* 0> Ol

to

2505624652

TOBACCO SMOKE POLLUTION

lung cancer deaths (or other disease end points) per year from passive smoking, its magnitude can be estimated from epidemiological cohort studies. In this manner, a dose-response model can be constructed, based on observed physical and epidemiological phenomena. This model yields an estimate of the true dose-response function. By comparing the predictions of such a mathematical model with the results of independent cohort studies of lung cancer mortality, the accuracy of the approximation can thereby be estimated. This approach permits the prediction of the risk of passive-smoking disease associated with a given level of exposure. This in turn permits the designation of control and abatement procedures which will yield a predetermined level of public health protection. Repace and Lowrey [1985a, 1986, 1987a, 1993 (in press)] have developed such a doseresponse model. This has been utilized in the estimation of the risks of passive smoking in the workplace [Repace and Lowrey 1985b, 1993 (in press)], and in airliner cabins (Repace 1988; USDOT 1990). Repace and Lowrey (1985a) estimated the aggregate average population risk from passive smoking as five lung cancer deaths per 100,000 personyears at risk, per milligram of tobacco tar inhaled per day, for the non-smoking population aged 35 years or older. The implications for public health policy are described in the next section. RISK CHARACTERIZATION OF TOBACCO SMOKE POLLUTION

How significant is the risk from tobacco smoke? An answer is provided by considering U.S. federal approaches to regulating environmental carcinogens. What level of cancer risk triggers regulation, and is there consistency among various federal agencies? Travis and co-workers (1987) have reviewed the use of cancer risk estimates in prevailing federal standards and in withdrawn regulatory initiatives to determine the relationship between risk level and federal regulatory action. They find definite patterns and consistency in the federal regulatory process. Travis et al. (1987) considered two measures of risk, lifetime cancer risk to the most ex-

http://legacy.library.ucsf.edu/tid/ylj10i00/pdf

135

posed individual and aggregate or average population risk, which incorporates population size. One-third of the time, federal agencies calculated aggregate risks by multiplying the risk to the most exposed by the population at risk, and two-thirds of the time,riskswere calculated by taking into account variations in exposure level. The latter approach is the preferred method. The regulatory actions considered in the analysis were performed by the Consumer Product Safety Commission, the Environmental Protection Agency, the Food and Drug Administration, and the Occupational Safety and Health Administration. Travis et al. (1987) describe two technical terms current in risk assessment circles: de manifestis risk and de minimis risk. A de manifestisriskis literally ariskof obvious or evident concern, and has its roots in the legal definition of an "obvious risk," one recognized instantly by a person of ordinary intelligence. De minimis risk has been used for a number of years by regulators to define an acceptable level of risk that is below regulatory concern. This term stems from the legal principle, de minimis non curat lex, "the law does not concern itself with trifles. ^ De manifests risks are so high that agencies almost always acted to reduce them, and de minimis risks are so low that agencies almost never act to reduce them. Therisksfalling in between these extremes were regulated in some cases, but not in others. Two categories were described: small populations and large populations. For low aggregate risk, the de manifestis level for individualriskwas found to be about 4 X 10~3 (a lifetime probability of four deaths per 1,000 persons atrisk),and the de minimis level was 1 X 10~4. For example, the EPA, in declining to regulate natural radionuclide emissions from elemental phosphorus plants with an individual lifetime risk of 1 X 10~\ weighed the maximum risk to the most exposed individuals against the low aggregate risk (0.06 cancer deaths per year), and against other factors such as cost. However, when the aggregate population risk level was large, that is, above 250 cancer deaths per year, the de manifestis risk dropped to about 3 X I0"\ and the de minimis risk dropped to 1 X 10~6. For example, if the lifetime risks of harm from ex-

136

THE DISEASE OF NICOTINE ADDICTION

Table 7-2 Summary of Risk Assessments of Lung Cancer Deaths (LCDs) in U.S. Nonsmokers from Environmental Tobacco Smoke (ETS) Exposure (Adjusted to 1988) Study Fong(1982V Repace and Lowrey (i 985-87) Russell etal.( 1986)° Rubins (1986) Wells (1988) Waldetal.(1986)

K\il\eTdai.(l9ma

Wigleetal.(1987) Arundel etal. (1987) Mean, ail 9 studies: Mean, 8 studies, excluding Arundel et al.

Range of Estimates (LCDs per year) 96O*M,80Of 685*-6,850* 43l4'-8.625 / ' (female LCDs only) (83% workplace)13 19-97

Mean or Best Estimate (LCDs per year) 2,900 6.700 ± 340 U07' 6.470 3.320' 8,124' >6.035* 5.691' .58" 4.500 * 2,800 5.000 * 2.400

Source: Repace and Lowrey 1990. "Estimate ours, interpolated from author's overallriskestimate. i Based on subltnearity assumption at low doses. r Based on linearity assumption at low doses. ''Lower bound based on smokere' respirable suspended particulate matter (RSP) exposure extrapolation (adjusted io include ex-smokers). 'Best esumate based on epidemiology (adjusted to include ex-smokers). /Based on linear extrapolation from nicotine in smokers. 'Based on unnary cotinine and ETS epidemiology. " Based on linear multistage from smokers and unnary cotinine. 'Based on ETS epidemiology and nonsmokers" LCD rates (adjusted to include ex-smokers). 'Based on unne cotinine in U.K. nonsmokers, ETS epidemiology (adjusted to include cx-smokersl. * Based on numerical interpreuuon of qualitauve judgment. 'Based on extrapolation of Canadian results to U.S. nonsmokers. "Based on linear extrapolation from retained RSP in smokers.

posure to a pollutant were of the order of 1 X 10"*, and 75 million persons were at risk, then this would produce about 75 deaths per lifetime. If 75 years is used as an exposure lifetime, this is 1 estimated death per year. How does the estimated annual mortality of various regulated carcinogenic air pollutants and tobacco smoke pollution compare with the de minimis risk level? Nine workers have estimated the lung cancer mortality risk of tobacco smoke pollution. Table 7-2 reproduces the nine estimates. In some cases, the lung cancer deaths are interpolated from an overall estimate that includes estimated tobacco smoke pollution-caused deaths from diseases other than lung cancer, and additionally includes ex-smokers in those cases where the authors did not include them in the original estimate. This facilitates intercomparison of studies. The mean of all estimates is 4,500 ± 2.800, and with the estimates of Arundel et al. (which differ by more than two standard deviations from the remainder) removed, about 5.000 ± 2.400.

http://legacy.library.ucsf.edu/tid/ylj10i00/pdf

Table 7-3 characterizes this estimated risk by comparison with the estimated risks for, other indoor and outdoor airborne carcinogens (Repace and Lowrey 1990). It is seen that tobacco smoke pollution poses a far more serious public healthriskthan all other Table 7-3 Comparison of Estimated Annual Cancer Deaths from Various Airborne Carcinogens in the U.S. Indoor Pollutants Environmental tobacco smoke 1 (homes & workplaces) Radon gas in homes Outdoor Pollutant^ Asbestos Vinyl chloride Airborne radionuclides Coke-oven emissions Benzene Arsenic Source Repace and Lowrey 1990. 'Lung cancer only. b Before control. 'Nonsmokers only.

No. 5,000" 4,000* No. 15