7.1 REVIEW OF HEALTH EFFECTS
The health effects of ETS have been recently and extensively reviewed in several reports of the Surgeon General (1982, 1983, 1984, 1985, 1986, 1989), and in documents of the World Health Organization (WHO 1986), the Environmental Protection Agency (1987), the National Research Council (1986a, 1986b), the Fourth International Conference on Indoor Air Quality and Climate (1987), and in key research studies. These documents collectively represent critical evaluations of the complete body of scientific literature for its meaning and accuracy. The health effects are briefly summarized below.
7.1.1 Acute Effects
While odor in itself is not a health effect, it can be considered as a psychophysiological factor contributing to the development of an adverse health response and thus is important in considering the impact of ETS. This is particularly true for the nonsmoker, whose threshold for
odor unacceptability is lower than the smoker who loses ETS odor detection sensitivity rapidly (Cain et al. 1983). While loss of sensitivity occurs in an experimentally controlled environment within four minutes after exposure begins, it is not meant to imply that it is directly applicable to the airliner cabin environment, where the number of individuals smoking at any given moment is highly variable. Odor, as the nonsmokers first sensory clue of ETS presence, is a major contributor to annoyance and is caused principally by the gas phase components of ETS.
While odor adaptation to ETS occurs over a short time frame, respiratory and ocular irritation increase proportionately over at least one hour at levels as low as 2 ppm CO (used as a surrogate for ETS concentration (Cain et at. 1987)). Ocular irritation begins at ETS levels lower than those causing respiratory irritation. Like odor, research suggested that eye irritation is caused predominantly by the gas-phase constituents of ETS (Weber 1986).
The evidence for acute respiratory and ocular irritation of ETS on the non-sensitive adult has been reported as equivocal and not scientifically conclusive (Lebowltz 1976; Schilling et al. 1977; Comstock et al. 1981; Schenker et al. 1982). Recent studies (Cain et al. 1987) have indicated that acute irritation is at least perceived to occur in individuals exposed to ETS and can be expressed as degree of dissatisfaction.
For individuals who are sensitive because they have preexisting conditions, such as asthma, that are provoked by ETS, or who, because of their stage in life, may be especially vulnerable, the acute effects can be more clinically significant and debilitating, leading to the notion that a smoking allergy may exist.
This is most apparent in infants and young children of smoking parents, who appear to be particularly susceptible to acute respiratory bronchitis and pneumonia from ETS exposure (U.S. Department of ‘Health and Human Services 1986). While components of cigarette smoke are known to affect other preexisting conditions such as cardiovascular disease, the acute effects of ETS on these conditions is unclear. Recent studies (Health Effects Institute 1988) have demonstrated induction of angina at a carboxyhemoglobin level of 4 percent, while a series of studies have indicated that CoHb levels of nonsmokers in smoking environments to be 2 percent or less (National Research Council l986a). Endogenous levels of carboxyhemoglobin levels in the U.S. population are typically 0.5 percent. These circumstances indicate that the cardiovascular effects of ETS on individuals with preexisting conditions may occur at levels not much above background, at least for C0.
In addition, a significant segment of the U.S. population with high blood pressure accompanied by angina or coronary disease is known to be adversely affected by nicotine exposure (National Research Council 1986a). Several studies examined the potential for ETS impact on cardiovascular disease. However, the acute cardiovascular effects of ETS on individuals with this preexisting condition have not been examined.
7.1.2 Chronic Effects
Knowledge about the importance of ETS to chronic obstructive pulmonary disease and other respiratory effects, cardiovascular disease, and cancer (particularly lung cancer) has been greatly enhanced by the large volume of data on mainstream smoke and these diseases.
7.1.2.1 Chronic Obstructive Pulmonary Disease and Other Respiratory Effects
For acute respiratory effects, the literature on ETS as an etiologic agent of lower respiratory tract illnesses is derived principally from children of smoking parents (Colley 1974; Bland et al. 1978; Weiss et al. 1980; Schenker et al. 1983; Ware et al. 1984; Charlton 1984). while the evidence for ETS as an etiologic agent of childhood asthma is equivocal (Gortmaker et al. 1982; Burchfiel 1984; Leeder et al. 1976; Horwood et al. 1985; Tashkin et al. 1984), infants and young children of smoking parents are more likely than those of nonsmoking parents to contract lower respiratory diseases such as bronchitis and pneumonia (Ware et al. 1984; Schenker et al. 1983; U.S. Department of Health and Human Services 1986) and therefore likely to be affected by ETS exposure on aircraft. Three clinical manifestations that are seen consistently in studies of children include cough, reduced lung function measured as forced expiratory flow at the 25 percent to 75 percent level (FEF 25-75) (Tager et al. 1979), and impaired development of forced expiratory volume (FEV) with growth (Tager et al. 1983; Berkey et al. 1986).
Data on effects of ETS on the adult respiratory system are inconclusive. While reduced FEF 25-75 has been reported by several investigators (Kauffmann et al. 1983; White and Froeb 1980), other studies have not shown an effect on adult lung function (Burchfiel 1986; Kentner et al. 1984). Studies on both children and adults as sensitive populations with preexisting asthma are also inconclusive (U.S. Environmental Protection Agency 1987).
7.1.2.2 Cardiovascular Disease
Mainstream cigarette smoke has been implicated as a causative agent of arteriosclerosis, coronary heart disease, and cerebrovascular disease. The contribution of ETS to these diseases and its mechanisms of action are inconclusive, although it appears from animal studies that the predominant influence is being exerted by nicotine (Schievelbein and Richter 1984; Liu et al. 1979) and to a lesser degree CO (Astrup and Kjeldren 1979). Several epidemiological investigations (U. S. Environmental Protection Agency 1986; Hirayama 1984, 1985; Gillis et al. 1984; and Garland et al. 1985) indicate impacts of ETS but present methodological problems that preclude the drawing of firm conclusions. What is certain is that nonsmokers in a smoking environment do receive biological doses of nicotine at levels sufficient to produce significant amounts (40 ng) of cotinine in the urine (Hill and Marquardt 1980).
7.1.2.3 Cancer
The evidence for an association of environmental tobacco smoke with cancer is indisputable, as detailed in recent definitive reports of the Surgeon General (U.S. Department of Health and Human Services 1986), the World Health Organization (1986) and the National Research Council (1986a).
The great majority of epidemiological studies have indicated causal association between ETS and lung cancer that is exposure-dependent. While there are differences in cancer rates between men and women, they are not widely divergent. Misclassification is of concern among some of the studies, but does not negate the weight of evidence on the whole in favor of the dose-effect relationship.
Other cancers that investigators have correlated with ETS, typically derived from spousal studies, include brain, cervical, and endocrine cancers. In the aggregate, they do not provide consistent evidence for cancer at remote sites caused by ETS (National Research Council 1986a).
7.1.2.4 Other Chronic Impacts
There is evidence that smoking during pregnancy lowers birth weight, and a growing suggestion that exposure to ETS during pregnancy may impact birth weight. This may be of concern to female flight attendants who may receive occupational ETS exposures while flying during their first trimester of pregnancy. However, when considered with studies of birth weights at higher elevations such as in Denver (Martin and Bracken, 1986), it is conceivable that prolonged or frequent periods at high altitudes may be more strongly and etiologically related to low birth weights than ETS.
7.1.2.5 QUANTITATIVE ESTIMATION OF CANCER RISK
ETS is a mixture that has been implicated in cancer, respiratory effects (upper respiratory tract irritation, chronic respiratory tract illness), and cardiovascular disease. Since there is no peer-reviewed and widely used method for conducting a risk assessment for complex mixtures such as ETS, each individual constituent must be carefully examined for its potential use as a marker and a representative of the ETS mixture in the quantitative estimation of health risk.
The scientific literature presents evidence that exposure to particulate-bound polycyclic aromatic hydrocarbons, as ETS products of incomplete combustion, correlate with the carcinogenic potential of ETS (Wynder and Hoffmann 1967), and that inhalation of respirable suspended particulate (RSP) is an appropriate representative of this potential.
Data on active smokers are not valid quantitative predictors of effects on passive smokers and were not used in this investigation because:
- Concentrations of carcinogens in active smoke are different from concentrations in ETS. For example, given equal weights of smoke particles, sidestream smoke contains approximately three times the benzo(a)pyrene in mainstream smoke.
- Using data from active smoking to obtain risks from passive smoking involves several orders of magnitude in dose extrapolation.
- Active smokers experience actual tissue damage to the respiratory system (e.g., loss of mucociliary escalators from tracheal epithelium) which might either promote or inhibit tudor formation relative to passive smokers.
- Doses are so high in active smokers that some of the apparent dose may be “wasted” (i.e., received after a tudor has already been initiated).
| Nonsmoking Section | |||
|---|---|---|---|
| Middle Boundary | Remote2 | ||
| NONSMOKING FLIGHTS | |||
| Optical | 10 | 11 | |
| Gravimetric | 59 | 69 | |
| Average of all four values | 37 | ||
| SMOKING FLIGHTS | Smoking Section1 | ||
| Domestic | |||
| Optical | 182 | 39 | 17 |
| Gravimetric | 181 | 70 | 48 |
| Average | 181 | 54 | 33 |
| Net (average RSP values on smoking flights minus nonsmoking flights) | 144 | 17 | -- |
| International | |||
| Optical | 143 | 46 | 26 |
| Gravimetric | 129 | 51 | 39 |
| Average | 136 | 49 | 33 |
| Net (average RSP values on smoking flights minus nonsmoking flights) | 99 | 12 | -- |
|
|||
Characterization of cancer risk from exposure to RSP requires information from three components: ambient air concentrations of RSP, exposure potential, and dose-response relationship for the health effect of interest, in this case cancer. The three components are related to one another as presented in Figure 7-1.
The appropriate parameters within each box in Figure 7-1 must be carefully selected from among the range of options so that the ultimate expression of risk approximates natural flight conditions and flying habits of interest as much as possible.
In this investigation, separate cancer risk determinations were conducted for domestic and international flights. This is because:
Independent samples for the monitoring activity wore drown from the pool of domestic flights on U.S. carriers and the pool of international flights on U.S. carriers
The sample from the pool of dos stic carriers was large and therefore could be drawn in a truly random fashion, whereas the sample drawn from the international pool was small due to prohibitive costs.
7.1.3 Ambient RSP Concentrations
RSP concentrations were obtained using optical and Gravimetric analytical methods. The relative merits of these two methods and the differences in results obtained from them are discussed in Section 5.0 of this report. Both methods were used for sampling because there was no clearly definable reason for favoring one over the other. The results of both methods of sampling were averaged for the determination of risk. RSP was measured at various seat locations on smoking and nonsmoking flights.
As a result, a number of RSP concentrations, representing various seat positions on smoking and nonsmoking flights, and using two methods of sample collection, were available for exposure assessment, as presented in Table 7-1. For estimation of exposure due exclusively to ETS, RSP concentrations on nonsmoking flights were subtracted as “baseline” values from RSP concentrations on smoking flights” On nonsmoking flights, the optical measurements of RSP may have been “Lower than actual and the Gravimetric measurements higher than actual. The difference between baseline” values obtained from the Gravimetric and optical methods of sampling ,whether averaged or used separately, did not change the outcome of the risk assessment. Therefore, the average of all values was used to represent the baseline concentration on nonsmoking flights. See Section 5.0 for more discussion on the monitoring results.
7.1.4 Exposure on Aircraft
The principal medium of exposure to RSP is via the air, so that inhalation is the primary exposure route of interest. Accordingly, the amount of RSP inhaled depends on respiratory rates, known to be variable for males and females, and for different states of physical activity. Respiratory rates have been determined for a range of conditions (U.S. Environmental Protection Agency 1989b). For this risk assessment, it is assumed that passengers are in a resting state throughout a flight, corresponding to an average respiratory rate of 0.5 m3/h (males 0.7; females 0.3). It is assumed that cabin crewmembers are engaged in moderate exercise, corresponding to an average respiratory rate of 2.1 m3/h (males 2.5; females 1.6).
Flying habits are also a critical determinant of exposure and risk. They include (a) the accumulated period of lifetime during which an individual flies, (b) the number of flights taken in that period, expressed as a yearly average, (c) the seat location chosen, and (d) the cumulative accounting of seat position over the course of the entire period of flying.
In determining exposures (and later risks), it is important to understand the terms of reference used to calculate quantitative estimates.
Proportion of space in each section of the aircraft is a unitless dimension. It is the fraction of the total cabin space that is dedicated to each of the smoking, boundary, and nonsmoking sections. Proportion of time in each section of the aircraft is similarly unitless. It is, by definition, identical to proportion of space on the assumption that as the space dedicated to one section varies, so does the time spent in that section by the equivalent of one individual. Flight hour is the time spent in flight during which smoking is permitted. Flight hours per year is the time spent in flight, during the course of one calendar year, during which smoking is permitted. RSP concentration is the amount of RSP, in ug, contained in one m3 of air. Duration of exposure is the number of years which one flies on smoking flights. For example, an individual who takes his or her first flight on an aircraft where smoking is permitted at age 20 and whose most recent flight on an aircraft where smoking is permitted occurred at age 40, has been flying for 20 years. An exposure coefficient,. in the context of this investigation, is the average amount of RSP (generated by ETS and used as a surrogate for ETS), in ug, inhaled by an individual during one hour of time in an airline cabin when smoking is permitted, and annualized over the period of a calendar year. (Further explanation of this term is described later in this section). A person-year is the equivalent of one year’s worth of time (365 days, not necessarily consecutive) for the equivalent of one person. Ten people, each exposed to ETS for 36.5 days, are equivalent to one person exposed for 365 days. A risk coefficient, in the context of this investigation, is the incremental number of premature deaths due to lung cancer among 100,000 nonsmokers exposed to ETS on flights where smoking is permitted.
TABLE 7-3. CALCULATION OF EXPOSURE FOR DOMESTIC FLIGHTS| (ug/person/flight hour) | Passenger | Cabin Crew |
|---|---|---|
| RSP concentration aggregated by time spent in each aircraft section (ug/m3) | 9.01 | 23.32 |
| Respiratory rate (m3/hr) | 0.5 | 2.1 |
| Exposure (ug/flight hour) | 4.53 | 48.94 |
Cancer risks are usually associated with long periods of time, i.e., several years of exposure to a carcinogen. The reasons for this are embedded in the prevailing theories of the mechanism of carcinogenesis. While this exposure may be greater or lesser at various times throughout the exposure interval, it is averaged out over a long time span to accommodate brief periods of higher or lower exposure, and the intervals during which no exposure may occur. Accordingly, cancer risk is usually expressed as the risk per unit of average daily exposure to a carcinogen, day after day and year after year (i.e., an annualized average). In this investigation, the unit of exposure per flight hour, as presented in Tables 7-3 and 7-4, must be made compatible with the “annualized daily averaging” concept that is used to construct cancer dose-response graphs and which is used to express cancer risk. This is accomplished by dividing the RSP exposure in one flight hour by 365 days per year to provide a value that represents an average exposure, during one hour when smoking is permitted on an aircraft, for any given day of the year. The result is an expression of an exposure coefficient. An exposure coefficient in this investigation is defined as the average daily amount of RSP inhaled by one individual during one flight hour, averaged over the course of a year. This is a conceptual construct that is necessary in order to make the exposure unit consistent with the dose-response unit in the calculation of risk.
TABLE 7-4. CALCULATION OF EXPOSURE FORINTERNATIONAL FLIGHTS
| (ug/person/flight hour) | Passenger | Cabin Crew |
|---|---|---|
| RSP concentration aggregated by time spent in each aircraft section (ug/m3) | 6.5 1 | 26.0 2 |
| Respiratory rate (m3/hr) | 0.5 | 2.1 |
| Exposure (ug/flight hour) | 3.3 3 | 54.6 4 |
|
||
Accordingly, the exposure values presented in Tables 7-3 and 7-4 must be annualized into an average daily exposure by dividing them by the 365 days in one year. Therefore, the exposure coefficients in this investigation, expressed as annual averages, are:
- For passengers on domestic flights: 4.5 ug/flight hour divided by 365 or 0.00001233 mg/h/exposure day
- For cabin crew members on domestic flights: 48.9 ug/flight hour divided by 365 or 0.00013400 mg/h/exposure day
- For passengers on international flights: 3.3 ug/flight hour divided by 365 or 0.00000904 mg/h/exposure day
- For cabin crew members on international flights: 54.6 ug/flight hour divided by 365 or 0.00015000 mg/h/exposure day.
SECTIONS OF CABIN1
| Domestic | International | |||
|---|---|---|---|---|
| Nonsmoking Section | Passenger | Cabin Crew | Passenger | Cabin Crew |
| Middle and Remote Rows | 0.84 | 0.75 | 0.82 | 0.65 |
| Boundary Rows | 0.11 | 0.10 | 0.13 | 0.10 |
| Smoking Section | 0.05 | 0.15 | 0.05 | 0.25 |
|
||||
These values are used in combination with cancer risk coefficients, derived from cancer dose-response graphs described below in Section 7.2.3, to produce exposure-specific expressions of risk.
It should be noted that the proportions of time spent in various sections of the aircraft cabin by cabin crewmembers, as indicated in Table 7-2, do not include time spent in galleys. Galleys have their own sources of ventilation. Consequently, those galleys located adjacent to the smoking sections of aircraft cabins may contain ambient air concentrations of ETS constituents that are different from concentrations measured in the smoking sections. The exposure of cabin crew members in the galley, therefore, may be different than in other sections of the aircraft, but this exposure could not, be estimated because aircraft galleys could not monitored for ambient air concentrations of ETS in this investigation.
7.1.5 Determination of Dose-Response Relationships and Risk Coefficients
A prominent feature of risk assessment is characterization of the toxicologic dose-response relationship. In the context of this investigation, it is the relationship between the amount of RSP inhaled and the number of lung cancer deaths that the inhaled RSP produces. The greater the RSP inhalation, the greater the amount of response in the form of increased number of lung cancer deaths. Graphically, the relationship is represented by a line, which can be expressed as a mathematical constant known as the coefficient of risk:
| Risk coefficient = | Number of lung cancer deaths per 100,000 persons at risk per milligram of RSP (annual average) per day |
Risk coefficients are frequently referred to as unit cancer risks in these analyses. The level of risk corresponding to a particular level of exposure can be determined by using the appropriate risk coefficient.
For this investigation, a number of dose-response models for the relationship of ETS to lung cancer deaths were considered, each having its own characteristic risk coefficient. These are described in Table 7-5.
The advantages and disadvantages of each model were weighed according to three criteria:
- Strength of each model as determined by the quality of the design and data used in its construction. The following characteristics are used to define model strength:
- The size of the study population used in model construction and validation
- The scientific soundness of the dose information used to construct the model
- The ease of model adaptation for intermittent exposure
- The size of the study subpopulation having the health endpoint of concern (e.g., cancer)
- The unique statistical design features of each model applicable to this study
- Whether the assumptions used in the model are reasonable
- Amount of peer review and scientific acceptance.
Two models were selected for this investigation because they most closely approximated the desirable traits embodied in the selection criteria: the Phenomenological Model (Repace and Lowrey, 1985) and the Armitage and Doll Model (Armitage and Doll, 1961), modified for less-than-lifetime exposure (Ginevan and Mills; 1986), and known as LesLife.
Repace and Lowrey estimate that the excess exposure of 1 mg/day increases lifetime lung cancer risk by 5 deaths per 100,000 person-years (PY) exposure. The Phenomenological model, though simple, is based on a fairly sizable body of data, and if it is inaccurate, would likely understate risk. These arguments have been fully developed by Repace and Lowrey (1985) and are briefly reviewed below.
The general Phenomenological Model is based on observed differences in lung cancer mortality between groups of never-smokers who were members of the Seventh Day Adventist Church and those who were not (Phillips et al. 1980 a,b). Because of their religion, which proscribes smoking, Seventh Day Adventists are less likely to encounter ETS, and 40 percent of the Seventh Day Adventist cohort worked for church-run organizations. The non-Seventh Day Adventists were a demographically comparable group of lifelong nonsmokers, among the general population, who resided in the same geographical area as the Seventh Day Adventists. The difference in lung cancer rates between Seventh Day Adventists and non-Seventh Day Adventists was taken to be due to their differential exposure to ETS, and the ratio of mortality differential to exposure differential was taken as the risk coefficient. The mortality rates among Seventh Day Adventists were based on 109 lung cancer deaths, and were therefore quite well determined. Moreover, Repace and Lowrey assumed that ETS exposure in Seventh Day Adventists was zero and thus based their dose-response coefficient on the maximum possible exposure. Since some Seventh Day Adventists were undoubtedly exposed to ETS in the workplace, this assumption is conservative in that it overstates the differential exposure and thus understates the actual dose-response.
The Modified Armitage and Doll Model is based on consideration of what the multistage theory of carcinogenesis predicts about age-specific risks of exposure to a fixed concentration of a carcinogen for a fixed duration of time. This risk assessment approach converts the ambient air data to a risk-equivalent dose. There are several underlying assumptions to this approach:
- RSP is a reliable indicator for estimating the relationship between exposure to cigarette smoke and health risks.
- Data on wives of smoking husbands indicate that their relative risk is approximately 1.3, based on case-control studies.
- Spousal exposure can be inferred from measurements of an individual smoker’s impact on indoor air quality in the home, together with empirical statistics on the duration of marriages.
- For the multistage model of carcinogenesis the following question can be posed: If X years of exposure at level Y cause a relative risk of 1.3, what is the dose-response coefficient?
- The dose-response coefficient, a five-stage multi-stage model of carcinogenesis, and dose estimates derived from airliner monitoring data, are used to calculate risks to the selected populations of interest. A five-stage model assumes that a number of events or “stages” must occur before a normal cell can become a cancer cell. The first stage is generally equated to a mutational event. Subsequent stages might include further mutations, as well as other biochemical changes in the cell. After all stages have occurred, the transformed cell proliferates until it becomes a clinically diagnosable tumor.
ARMITAGE AND DOLL MODEL
| Parameter | Phenomenological Model | Modified Armitage and Doll Model |
|---|---|---|
| Age of first exposure | Fixed at age 20 | Adaptable to any age |
| Duration of exposure | 45 years | Variable |
| Linearity | Linear at low doses | Linear at low doses |
| Stages of carcinogenesis | None assumed | 5 |
| Concurrence of risk coefficients | 5 lung cancer deaths/100,000 exposed/mg exposure-day | 6.45 lung cancer deaths/100,000 exposed/mg/exposure-day |
A detailed discussion of this methodology, including a sensitivity analysis of the model, is presented in Appendix A. The principal advantage of this modeling approach is that it permits the user to explicitly specify such important factors as age at commencement of exposure and duration of exposure. At the same time, as demonstrated in the sensitivity analysis contained in Appendix A, the lung cancer risk data for ETS exposure are sufficiently abundant and consistent that altering parameters of this modeling approach does not alter the conclusions about risk in any significant way.
A comparison of the basic features of the two models is contained in Table 7-6. Both models have undergone peer review. The risk coefficients presented by these two models are:
- For the Phenomenological Model, 5 excess lifetime lung cancer deaths/100,000 person-years exposure/mg RSP/exposure-day, ascribable to ETS assuming a constant exposure. The lung cancer rate is an average value based on lifetable statistics.
- For the Modified Armitage and Doll Model, 6.45 excess lung cancer deaths per 100,000 persons at risk/mg RSP/exposure-day, ascribable to ETS.
Using these risk coefficients, the risk of death from lung cancer as a result of exposure to ETS in airliner cabins was determined as a function of number of years flown. For the Modified Armitage and Doll Model, the risk of death from cancer is dependent on the age of first exposure to ETS as a potential carcinogen. Therefore, each commencement age warrants its own unique exposure-response relationship, as depicted in Figure 7-2 for the Modified Armitage and Doll Model. The exposure-response relationship for the Phenomenological Model is presented in Figure 7-3. The graphs in these figures serve as risk nomograms, allowing an individual to determine his or her appropriate unit of risk according to the number of years of flight (i.e., the number of years of exposure). In the case of the age-dependent Modified Armitage and ‘Doll Model, the appropriate curve representing the age at which flying commences is selected prior to determination of the risk coefficient.
7.1.6 Risk Characterization
7.1.6.1 Individual Risk
Once the risk coefficient is determined, it is multiplied by the appropriate exposure coefficient presented in Section 7.2.2 (on domestic flights — 0.00001233-mg/h/exposure day for passengers and 0.00013400 mg/h/exposure day for cabin crewmembers; on international flights–0.00000904 mg/h/exposure day for passengers and 0.00015000 mg/h/exposure day for cabin crew members) to determine exposure-specific risk. The final expression is the incremental risk due to premature lung cancer deaths among nonsmokers, ascribable to ETS on smoking flights.
TABLE 7-7. SUMMARY OF DATA CONTAINED IN THE EXAMPLE CALCULATIONS OF RISK| Example 1 | Example 2 | Example 3 | |
|---|---|---|---|
| Cabin occupant | Crew Member | Business Passenger | Casual Passenger |
| Hours per year in flight 1 | 900 | 450 | 45 |
| Number of years flown | 20 | 30 | 40 |
| Age at start of flying2 | 25 | 35 | 25 |
| Exposure coefficients (mg/h/exposure day) | |||
| Domestic | 0.00013400 | 0.00001233 | 0.00001233 |
| International | 0.00015000 | 0.00000904 | 0.00000904 |
| Risk coefficients3 | |||
| Phenomenological Model | 100 | 150 | 200 |
| Modified Armitage and Doll Model | 123 | 49 | 150 |
| Risk4 | |||
| Domestic | |||
| Phenomenological Model | 12.06 | 0.83 | 0.11 |
| Modified Armitage and Doll Model | 14.86 | 0.27 | 0.08 |
| International | |||
| Phenomenological Model | 13.46 | 0.61 | 0.08 |
| Modified Armitage and Doll Model | 16.59 | 0.20 | 0.06 |
The procedure for determining risk can be illustrated in the following three examples, the parameters and results of which are summarized in Table 7-7. These examples are intended to represent occupational and non-occupational profiles of flying habits. Typical flight frequency and duration for cabin crewmembers were used for one example in the occupational setting. Flight frequencies and durations for passengers (representing profiles of a frequent flyer and a non frequent flyer) used for the two examples in the non-occupational setting are likely to be at the high end of the range. Risks for a range of other scenarios are presented in Appendix B. Data on the number of cabin crewmembers who smoke were not available. However, it is known that approximately 29 percent of U.S. adults aged 20 or older smoke (U.S. Department of Health and Human Services, 1989).
Example 1. Risk determination for a cabin crewmember who flies ours per month or 960 hours per year (see Table 6-1) on domestic flights: The total number of hours is reduced by 6.25. percent as an approximation of the flight time during which the no-smoking light is illuminated, resulting in 900 flight hours when smoking is permitted. The period of flying is 20 years, commencing at age 25. These values were chosen because they represent the career length and career commencement for a large percentage of cabin crewmembers (Association of Flight Attendants 1988). The exposure coefficient for cabin crewmembers on domestic flights is 0.00013400 mg/h/day. Referral to Figure 7-3 produces a unit cancer risk, for a 20-year duration of exposure, of 100-lung cancer deaths/mg RSP/day/100,000 exposed nonsmokers (who, in this case, are exposed nonsmoking cabin crewmembers on smoking flights). A final multiplication of the exposure coefficient (0.00013400 mg/h/day) by the unit cancer risk (100 lung cancer deaths/mg RSP/day/100,000 individuals) yields a risk of lung cancer deaths amounting to 0.0134/100,000 for every hour flown in a smoking environment. Since cabin crew members are estimated to fly 900 hours per year during smoking periods, the incremental risk of premature death from lung cancer ascribable to ETS on smoking flights is 12.06/100,000 exposed cabin crew members, or 1 in every 8,292 cabin crew members according to the Phenomenological Model of cancer deaths, as presented in Table 7-7. A similar calculation using the Modified Armitage and Doll Model in Figure 7-2 produces an incremental risk of premature death from lung cancer amounting to 14.86/100,000 nonsmoking cabin crew members on smoking flights, as presented in Table 7-7,or 1 lung cancer death per 6,729 nonsmoking cabin crew members.
Example 2. Risk determination for a passenger who is representative of a frequent flyer: This individual logs 480 hours per year (reduced to 450 hours per year for the 6.25 percent of time when the no-smoking light is assumed to be illuminated). This is approximately covalent to an average of four round-trip coast-to-coast flights per month. The individual is assumed to continue this pattern of flying for 30 years commencing at age 35, to constitute what is likely an upper limit on the amount of time spent in an airliner cabin environment during a lifetime. The exposure coefficient for passengers on domestic flights is 0.00001233 mg/h/exposure day. The unit cancer risk for this individual, according to the Phenomenological Model in Figure 7-3, is 150 lung cancer deaths/mg RSP/day/100,008 exposed non-smokers. Taking into account the exposure coefficient and period of flying, the incremental risk is 150 x 0.00001233 x 450, or 0.83 premature lung cancer deaths ascribable to ETS for every 100,000 exposed nonsmoking passengers on smoking flights, according to the conditions in this example, or 1 in 120,482 nonsmoking passengers. The Modified Armitage and Doll Model produces a risk of 49 x 0.00001233 x 450, equal to an incremental risk of premature lung cancer death of 0.27104,000 nonsmoking passengers on smoking flights as presented in Table 7-7, or lung cancer death per 370,370 nonsmoking passengers.
Example 3. Risk determination for a passenger who is representative of a non-frequent flyer: Flight time of 48 hours per year, adjusted for no-smoking periods, is assumed to be 45 hours per year for 40 years, commencing at age 25. The annual flight frequency was assumed to be one-tenth that of the frequent flyer in example 2, occurring on a casual basis throughout adult life. The exposure coefficient is 0.00001233 mg/h/exposure day and the Phenomenological Model unit cancer risk is 200 lung cancer deaths/mg RSP/day/100,000 exposed nonsmokers. The incremental risk is therefore 200 x 0.00001233 x 45, or 0.11 lung cancer deaths ascribable to ETS for every 100,000 exposed nonsmoking passengers on smoking flights, according to the conditions in this example, or 1 in 900,091 nonsmoking passengers. The Modified Armitage and Doll Model produces an incremental risk of 150 x 0.00001233 x 45, equal to an incremental risk of premature lung cancer death of 0.08 as presented in Table 7-7, or 1 premature lung cancer death per 1,250,000 nonsmoking passengers.
7.1.6.2 Population-Based Risk
For passengers, the risk of premature lung cancer death can be expressed on a population basis. In 1987, 418 million domestic enplanements occurred (U.S. Department of Transportation, 1987), the average flight time was 1.84 hours (based on analysis of data provided by the Federal Aviation Administration) and smoking was permitted on 54.3 percent of all flight hours. It follows that, for current conditions under which a ban on smoking exists for flights with durations of two hours or less, estimates for passengers on domestic flights are:
| Passenger hours flown/year | = | 418 million x 1.84 | = | 769 million |
| Passenger hours flown/year on smoking flights | = | 769 million x 54.3 % | = | 418 million |
| Reduced 6.25% for the time that the no-smoking light is assumed to be illuminated on a flight | = | = | 391 million hours per year | |
| Number of individuals flying 45 hours per year (from Example 3 above) | = | 391 million / 45 | = | 8.7 million |
| Number of “lifetimes” of flying 40 years (from Example 3 above) | = | 8.7 million / 40 | = | 0.217 million passengers/yr. |
| Expected population-based risk (based on a risk of 0.11 lung cancer deaths per 100,000 exposed nonsmokers according to the Phenomenological Model in Example 3 above, or 1.1/million) | = | 0.217 million x 1.1/million | = | 0.238 premature lung cancer deaths per year. |
A similar calculation for passengers on international flights, using 62 million enplanements per year (U.S. Department of Transportation 1987), an average flight time of 4.75 hours per flight (based on analysis of data provided by the Federal Aviation Administration), a flight frequency of 45 hours per year, a duration of flying of 40 years, and a risk of 0.08 premature lung cancer deaths per 100,000 exposed nonsmokers (from Example 3 above) results in a population-based risk of 0.122 premature lung cancer deaths per year. (In this calculation, all flights are presumed to be smoking flights, so that the fraction of flight hours on which smoking is permitted is reduced only by the time that the no-smoking light is illuminated — assumed to be 6.25 percent.)
For cabin crew members on domestic flights, the calculation is somewhat different, based on the number of individuals logging approximately 960 hours per year (80,000; see Table 6-1), and the proportion who fly on domestic (0.7) and international (0.3) flights. Using 54.3 percent as the percent of flight hours during which smoking is permitted under the two-hour ban enacted in 1988, then:
| Number of cabin crewmembers flying on domestic flights | = | 80,000 x 0.7 | = | 56,000 |
|---|---|---|---|---|
| Of these, number of cabin crewmembers flying onsmoking flights | = | 56,000 x 0.583 | = | 30,408 |
| Number of “lifetimes” flying 20 years (from Example 1 above ) | = | 30,408 / 20 | = | 1520 |
| Expected population-based risk (based on a risk of 12.06 lung cancer deaths per 100,000 exposed nonsmokers in Example 1 above) | = | 1520 x 12.06/100,000 | = | 0.183 premature lung cancer deaths per year. |
For cabin crew members on international flights the calculation is:
| Number of cabin crew members flying on international flights | = | 80,000 x 0.3 | = | 24,000 |
|---|---|---|---|---|
| Number of “lifetimes” flying 20 years (from Example 1 above) | = | 24,000 / 20 | = | 1200 |
| Expected population-based risk (based on a risk of 13.46 lung cancer deaths per 100,000 exposed nonsmokers according to the Phenomenological Model in Example 1 above) | = | 1200 x 13.46/100,000 | = | 0.162 premature lung cancer deaths per year. |
All international flights in this case are presumed to be smoking flights, so that no reduction in the number of flights to account for those that are nonsmoking is necessary.
7.1.7 Discussion
The cancer risk coefficient for 45 years of exposure to RSP as a surrogate for ETS is 5 premature lung cancer deaths per 100,000 (5 X 10-5) nonsmokers per mg RSP, ascribable to ETS, as derived from the Phenomenological Model by Repace and Lowrey (1985). The counterpart age-dependent risk coefficients using the Modified Armitage and Doll Model range from 40 premature lung cancer deaths per 100,000 (4 x 10-4) nonsmokers, for exposure commencing at 35 years of age, to 600 premature lung cancer deaths per 100,000 (6 x 10-3) nonsmokers for exposure commencing at 5 years of age. For comparison, risk coefficients for other chemicals that present a potential for inhalation exposure are presented in Table 7-8. All of the substances listed in this table are regulated by the EPA under its various statutes.
TABLE 7-8. RISK COEFFICIENTS FOR A RANGE OF CHEMICALS IN COMPARISONWITH ETS IN AIRCRAFT CABINS
| Risk Coefficient (Cancer Potency Factor) | Cancer Classification1 | |
|---|---|---|
| ETS (Phenomenological Model) | 5 x 10-5 | |
| ETS (Modified Armitage and Doll Model) | 6 x 10-3 commencing at 5 years | |
| 4 x 10-4 commencing at 35 years | ||
| Acrylonitrile | 2.4 x 10-1 | B1 |
| Arsenic and compounds | 5 x 101 | A |
| Benzene | 2.6 x 10-2 | A |
| B1s(chloromethyl)ether | 9.3 x 103 | A |
| 1,2-Dichloroethane | 3.5 x 10-2 | B |
| Ethylene oxide | 3.5 x 10-1 | B1/B |
| Nickel and compounds | 1.19 | A |
| Tetrachloroethylene | 1.7 x 10-3 | B |
| Trichloroethylene | 4.6 x 10-3 | B |
| Polynuclear aromatic compounds | 6.11 | - |
| Vinyl chloride | 2.5 x 10-2 | A |
|
||
The risks calculated here are well within the spectrum of risks from other carcinogen exposures. The risks derived from the Phenomenological Model and the Modified Armitage and Doll Model suggest that two approaches which differ in both design and database give nearly the same result. The major divergence is for the case of exposure early in life using the Modified Armitage and Doll Model where risk is elevated by an order of magnitude, relative to risks from exposure commencing later in life. This is because exposure to first-stage carcinogens is especially damaging to the young. While making other assumptions regarding stage of action would generally reduce these risks, it would generally elevate risks in older travelers, who constitute the majority of the traveling public (Murdoch and Krewski 1988). Thus the fundamental results obtained here are not “conservative” in the sense that they overstate actual risk. Rather, they are the best estimates implied by the database and the models selected. Worst-case, upper bound, estimates such as are often used in a regulatory context could well be a factor of five higher.
7.2 QUANTITATIVE ESTIMATION OF ACUTE RESPIRATORY EFFECTS
The most common complaint from exposure to ETS-based carbon monoxide and nicotine is upper respiratory tract and ocular irritation, as verified by the descriptions of prior investigations in Section 7.1.1. Two studies provide empirical dose-response measures of respiratory and ocular irritation from exposure to various levels of carbon monoxide (Cain et al. 1987) and nicotine (Mattson et al. 1989). These dose-response relationships were applied to the carbon monoxide and nicotine levels obtained in this investigation to determine whether these pollutants, at their observed concentrations, constitute sources for health effects.
7.2.1 Carbon Monoxide
Carbon monoxide (CO) was measured continuously during all flights. This presented an opportunity to disclose peak concentrations that might have appeared at various times throughout flight. Short-term CO concentrations have been tested as surrogates for the acute respiratory irritant properties of ETS in smoking environments. Therefore, peak concentrations of CO as a surrogate for ETS as a short-term respiratory irritant are of interest in the aircraft cabin environment. Accordingly, continuous 30-minute averages of CO concentrations were plotted as a function of their cumulative frequency in the smoking, boundary, and nonsmoking sections, as presented in Figure 7-4. Thresholds for discom fort as described by Cain et al. (1987) for 2 ppm and 5 ppm CO were superimposed over the concentration-frequency plots as an indication of the levels of CO and frequencies with which discomfort to eyes, nose, and throat might occur from exposure to these levels. These researchers determined that the 2 ppm level of CO produced dissatisfaction among 12 percent of individuals exposed for 60 minutes, while the 5 ppm of CO produced dissatisfaction among 18 to 30 percent of individuals exposed for 60 minutes (this range includes eye, nose, and throat irritation). It is apparent from Figure 7-4 that approximately 32 percent of the 30-minute CO averages exceeded 2 ppm in the smoking section of the aircraft cabin. Applying the data of Cain et al., this implies that on 32 percent of the flights where CO levels exceeded 2 ppm, 12 percent of the occupants sitting in the smoking section would experience respiratory discomfort after 60 minutes of exposure to CO from ETS. Similarly, on 5 percent of all flights tested, the 30-minute CO averages exceeded 2 ppm in the boundary and nonsmoking sections. This implies that on 5 percent of the flights, 12 percent of the nonsmokers in these sections would be dissatisfied. It should be noted that in the Cain et al. study, 25 percent of the individuals tested were smokers. In addition, odor perception, over a 60-minute time period, for occupants of a space containing 2 ppm or 5 ppm CO as a surrogate for ETS (e.g., passengers or cabin crew members in an aircraft cabin containing ETS would be less sensitive than for visitors to that space (e.g., the flight engineer who leaves the flight deck to visit the cabin.
7.2.2 Nicotine
Integrated samples of nicotine were taken on 61 domestic smoking and 8 international smoking flights. The results of sample analysis for domestic flights are presented in Figure 7-5, as the percentages of flights with nicotine concentrations at or below the plotted values. For domestic flights, nicotine was below the detection limit in the smoking section on 5 percent of flights, lower than detectable in the boundary section on 62 percent of flights and lower than detectable in the nonsmoking section an 75 percent of flights. Recognizing that integrated samples do not reveal peak short-term concentrations during flight, average concentrations of nicotine never exceeded 2 ug/m3 in the nonsmoking section and 5 ug/m3 in the boundary section. Nicotine concentrations obtained on international flights and in a recent study by Mattson et al. (1989) presented too few data points to construct valid plots.
The level of discomfort from ETS measured as ETS-generated nicotine aboard aircraft, observed by Mattson et al. (1989), was superimposed over the concentration data in Figure 7-5. These researchers determined that no subjects reported moderate or mild sensory response of the nose and eye at a concentration of 4 ug/m3. Using this value as a threshold
for response in the present study, nicotine in the boundary and nonsmoking sections did not reach concentrations that would provoke nose and eye irritation on any flights. Nicotine concentrations did exceed this threshold concentration in the smoking section on 65 percent of the domestic flights that were monitored. Subjects in the Mattson et al. study reported marked sensory response at nicotine concentrations of approximately 16 Ug/m3. This value significantly exceeded nicotine levels in the boundary and nonsmoking sections of all domestic flights that were monitored, so that marked sensory responses from nicotine would not be expected. However, nicotine concentrations in the smoking sections of 35 percent of all domestic flights monitored reached levels that would evoke a marked sensory response in the eye and nose.
7.3 ESTIMATION OF CARDIOVASCULAR EFFECTS
While ETS has been shown as an etiologic agent of cardiovascular disease (Wells, 1988), no definitive data exist on the quantitative relationship between ETS and ischaemic heart disease, particularly for individuals with preexisting cardiovascular illness, as acknowledged by the National Research Council (1986a) and the U.S. Department of Health and Human Services (1983). Simply put, not enough is known about the physiology and etiology of ETS-induced cardiovascular disease to postulate a dose-response model.
Scientific evidence suggests that CO impacts the cardiovascular system (causing angina and cardiac ischemia) and implies that nicotine also has an effect. In the case of C0, a recent multi-center investigation has demonstrated that 3 percent carboxyhemoglobin contributes to the expression of angina (Health Effects Institute, 1988), which is a symptom of cardiac effect but not necessarily indicative of coronary heart disease. In that study, an exposure chamber CO concentration of 9 ppm for up to 50 minutes produced 3 percent carboxyhemoglobin. The EPA has estimated that in the ambient air environment, 2.7 percent (at rest) or 2.9 percent (with moderate exercise) carboxyhemoglobin is equivalent to breathing an ambient air CO concentration of 20 ppm for 8 hours, based on the Coburn equation (Federal Register, 1985). Similarly, 2 percent carboxyhemoglobin is equivalent to breathing 35 ppm CO for one hour (with moderate exercise) or 15 ppm CO for 8 hours (at rest). This is the basis of an EPA 8-hour standard of 9 ppm for CO (Federal Register 1985). However, the contribution of CO in ETS to nonsmoker carboxyhemoglobin is unknown. Smokers self-dose with ETS-derived CO to a level of 3 to 8 percent carboxyhemoglobin; the nonsmokers’ ETS-induced carboxyhemoglobin levels are presumably less. Endogenous carboxyhemoglobin levels (in the absence of ambient air CO) in the U.S. population are approximately 0.5 percent. The CO levels measured aboard aircraft in this study, including the peak concentrations, were considerably less than 9 ppm.
EFFECTS OF ETS ON SPECIAL POPULATIONS
Children, asthmatics, and individuals with preexisting cardiovascular disease constitute populations of special concern for the health effects of ETS.
Although there is evidence to suggest that the respiratory system of children is affected by chronic exposure to ETS (based on studies of children in homes of smoking parents), given the small number of hours that children fly, the risk from exposure to ETS in aircraft cabins is likely to be small.
Currently available data are insufficient to quantify the impact of carbon monoxide on asthmatics. A recent review by the EPA (U.S. Environmental Protection Agency 1989a) on the current status of knowledge regarding the health effects of CO demonstrates a lack of information in this area. There are indications in individual research papers of what conceptually may be the effects of exposure to CO by asthmatics (e.g., decrease in lung cell function and degradation of epithelial cell integrity). Any quantitative observations as to where the threshold for acute respiratory effects of CO on asthmatics lies, and whether it is likely to be lower and the symptomatic response larger than for non-asthmatics, are currently considered to be speculative. Similarly, the quantitative impact of CO on preexisting ischaemic heart disease or other cardiovascular illness at the levels observed in airliner cabins’ currently cannot be estimated because health data are insufficient.
The impact of nicotine on the respiratory system of asthmatics is even more poorly understood than for carbon monoxide. No empirical quantitative data are available to determine the level of nicotine that would provoke an asthmatic response, or whether the level causing respiratory irritation in non-asthmatics is different from that causing irritation in asthmatics.
REFERENCES
- Armitage, P. and R. Poll. 1961. “Stochastic models for carcinogenesis.” Proceedings of the Fourth Berkekey Symposium on Mathematical Statistics and Probability. University of California Press, Berkekey, California. 4:19-38 .
- Astrup, P. and K. Kjeldsen. 1979. “Model studies linking carbon monoxide and/or nicotine to arteriosclerosis and cardiovascular disease.” Prevent. Med. 8:295-302.
- Berkey, C.S., J.H. Ware, D.W. Dockery, B.G. Ferris Jr. and F.E. Speizer. 1986. “Indoor air pollution and pulmonary function growth in preadolescent children.” Am. J. Epidemiol 123:250-260.
- Bland, M., B.R. Bewley, V. Pollard, and M.H. Banks. 1978. “Effect of children’s and parents’ smoking on respiratory symptoms.” Arch. Dis. Child. 53:100-105.
- Burchfiel, C.M., M.W. Higgins, J.B. Keller, N.J. Butler, W.F. Howatt, and I.T.T. Higgins. 1986. “Passive smoking, respiratory symptoms and pulmonary function: a longitudinal study in children.” Am. Rev. Respir.Dis. 133:A157.
- Cain, W.S., B.P. Leaderer, R. Isseroff, L.G. Berglund, R.J. Huey, E.D. Lipsitt and D. Perlman. 1983. “Ventilation requirements in buildings. Control of occupancy odor and tobacco smoke odor.” Atmos. Environ. 17:1182-1197.
- Cain, W.S., T. Tarik, L. See, and B. Leaderer. 1987. “Environmental tobacco smoke: sensory reactions of occupants.” Atmos. Environ. 21(2):347-353.
- Charlton, A. 1984. “Children’s coughs related to parental smoking.” Brit. Med. J. 288:1647-1649.
- Colley, J.R.T. 1974. “Respiratory symptoms in children and parental smoking and phlegm production.” Brit. Med. J. 2:201-204.
- Comstock, G.W., M.B. Meyer, K.H. Helsing, and M.S. Tockman. 1981. “Respiratory effects of household exposures to tobacco smoke and gas cooking.” Am. Rev. Respir. Dis. 124:143-148.
- Federal Register. 1985. U.S. Environmental Protection Agency carbon monoxide standard. 50 FR 37491. September 13, 1985.
- Fourth International Conference on Indoor Air Quality and Climate. 1987. Proceedings of the Conference, August 17-21, Berlin (West).
- Garland, C., E. Barrett-Connor, L. Suarez, M.H. Criqui, and D.L. Wingard. 1985. “Effects of passive smoking on ischaemic heart disease mortality of nonsmokers.” Am. J. Epidemiol. 121:645-650.
- Gillis, C.R., D.J. Hole, V.M. Hawthorne, and P. Boyle. 1984. “The effects of environmental tobacco smoke in two urban communities in the west of Scotland.” Europ. J. Respir. Dis. 65(Suppl. 133):121-126.
- Ginevan, M.E. and W.A. Mills. 1986. “Assessing the risks of radon exposure: the influence of cigarette smoking.” Health Physics 51:163-174.
- Gortmaker, S.L., D.K. Walker, F.H. Jacobs, and H. Ruch-Ross. 1982. “Parental smoking and the risk of childhood asthma.: Am. J. Public Health 72:574-579.
- Health Effects Institute. 1988. The multicenter study on the acute effects of carbon monoxide on individuals with coronary artery disease. Cambridge, MA.
- Hill, P. and H. Morquardt. 1980. “Plasma and urine changes after smoking different brands of cigarettes.” Clin. Pharm. and Ther. 27(5):652-658.
- Hirayama, T. 1984. “Cancer mortality in nonsmoking women with smoking husbands based on a large-scale cohort study in Japan.” Prevent. Med. 13:680-690.
- Horwood, L.J., D.M. Fergusson, and F.T. Shannon. 1985. ‘Social and familial factors in the development of early childhood asthma.” Pediatrics 75:859-868.
- Kauffmann, F., J.F. Tessier, and P. 0riol. 1983. “Adult passive smoking in the home environment. a risk factor for chronic airflow limitation.” Am. J. Epidemiol. 117:269-280.
- Kenter, M., G. Triebig, and D. Weltle. 1984. “The influence of passive smoking on pulmonary function – a study of 1,351 office workers.” Prevent. Med. 13:656-669.
- Lebowitz, M.D., G. Corman, M.K. 0’Rourke and C.J. Holberg. 1984. ‘Indoor-outdoor air pollution, allergen and meteorological monitoring in an arid Southwest area. J. Air Pollut. Control Assoc. 34:1035-1038.
- Leeder, S.R., R. Corkhill, L.M. Irwig, W.W. Holland, and J.R.T. Colley. 1976. “Influence of family factors on the incidence of lower respiratory illness during the first year of life.” Brit. J. Prev. Soe. Med. 30:203-212.
- Liu, L.B., C.B. Taylor, S.K. Peng and B. Mikkelson. 1979. “Experimental arteriosclerosis on Rhesus monkeys induced by multiple risk factors: Cholesterol, vitamin D. and nicotine.” Arterial wall 5:25-3B.
- Martin, T.R. and M.B. Bracken. 1986. “Association of low birth weight with passive smoke exposure in pregnancy.” Am. J. Epidemiol. 124:633-42.
- Mattson, M., G. Boyd, D. Byar, C. Brorn, J. Callahan, D. Corle, J. Cullen, J. Greenblatt, N. Haley, K. Hammond, J. Lewtas and N. Reeves. 1989. “Passive smoking on commercial airline flights.” J. Amer. Med. Assoc. 261:867-872.
- Murdoch, D.J. and D. Krewski. 1988. “Carcinogenic risk assessment with time-dependent exposure patterns.” Risk Anal. 8:509-530.
- National Research Council. 1986a. Environmental Tobacco Smoke: Exposures and Assessing Health Effects. National Academy Press. Washington, D .C .
- National Research Council. 1986b. The Airliner Cabin Air Environment. Air Quality and Safety . National Academy Press, Washington, D .C.
- Phillips, R.L, J.w. Kuano, N.L. Beeson, and T. Lotz. 1980a. “Influence of selection versus lifestyle on risk of fatal cancer and cardiovascular disease among Seventh Day Adventists.” Am. J. Epidemiol. 112:296:.314.
- Phillips, R.L, L. Garfinkel J.W. Kuzma, W.L. Beeson, T. Lotz. and B. Brin. 1980b. “Mortallty among California Seventh Day Adventists for selected cancer sites.” J. Nat. Cancer Inst. 65:1097-1107.
- Repace, J.L. and A.H. Lowrey. 1985. ‘A quantitative estimate of nonsmokers lung cancer risk from passive smoking.” Environ. Internat. 11:3-22.
- Schenker, M.B., S.T. Weiss, and B.J. Murawski. I982. “Health effects of residents in homes with urea formaldehyde foam insulation: a pilot study.” Environ. Int. 8:359-363.
- Schenker. M.B., J.M. Samet, and F.E. Spelzer. 1983. ‘Risk factors for childhood respiratory disease: the effect of host factors and home environmental exposures.” Am. Rev. Respir. Dis. 128:1038-l043.
- Schievelbein, H. and F. Richter. 1986. ‘The influence of passive smoking on the cardiovascular system.’ Prevent. Med. 13:626-644.
- Schllling, R.S.F., A.D. Litai, S.L. Hiu, G.J. Beck, J. B. Schoanburg, and A. Bouhuys. 1977. “Lung function, respiratory disease and smoking in families.” Am. J. Epidemiol. 106:274-283.
- Tager, I.B., S.T. Wefss, B. Rossner, and F.E. Spaizer. 1979. ‘Effect of parental cigarette smoking on the pulmonary function of children.’ Am J. Epidemlol. 110:15-26.
- Tager, I.B., S.T. Weiss, A. Nunoz, B. Rosner, and F. Speizer. 1983. “Longitudinal study of the affects of parental smoking on pulmonary function in children.’ New Eng. J. Med. 309:699-703.
- Tashkin, D.P., V.A. Clark, M. Simnons, C. Reams, A.H. Coulson, L.B. Bourque, J. W. Sayre, R. 0ntels, and S. Roknw. 1988. “The UCLA population studies of chronic obstructive respiratory disease relationship between parental smoking and children’s lung function.’ Am. Rev. Respir. Dis. 129:891-897.
- U.S. Department of Health and Human Services. 1982. The Consequence of Smoking: Cancer. Report of the Surgeon General. Rockville, Md.
- U.S. Department of Health and Human Services. 1983. The Health Consequences of Smoking : Cardiovascular Disease. Report of the Surgeon General. Rockville, Md.
- U.S. Department of Health and Human Services. 1984. The Health Consequences of Smoking: Chronic Obstructive Lung Disease. Report of the Surgeon General. Rockville, Md.
- U.S. Department of Health and Human Services. 1985. Consequences of Smoking : Cancer and Chronic Lung Disease in the workplace. Rockville, Md.
- U.S. Department of Health and Human Services. 1986. The Health Consequences of Involuntary Smoking . Report of the Surgeon General . Rockville, Md.
- U.S. Department of Health and Human Services. 1989. Reducing the Health Consequences of Smoking : Twenty -five Years of Progress. Report of the Surgeon general . Rockville, Md.
- U.S. Department of Transportation. Federal Aviation Administration. 1987. Airport Activity Statistics of Certified Route Air Carriers: Twelve Months Ending December 31,1987 U.S. Department of Transportation, Washington, D.C.
- U.S. Environmental Protection Agency. 1986. Air Quality Criteria for Ozone and other photochemical oxidants. Volumes 4 and 5 EPA/600/8-84/0206F. Environmental Criteria Assessment Office, Research Triangle Park, N.C.
- U.S. Environmental Protection Agency. 1987. EPA Indoor Air Quality Plan: Appendix A. Preliminary Indoor Air Pollution Information Assessment. EPA/600/8-87/014. Office of Health and Environmental Assessment, Washington, D.C.
- U.S. Environmental Protection Agency. 1989a. Air Quality Criteria for Carbon Monoxide. Addendum. In draft preparation. Environmental Criteria and Assessment Office. Research Triangle Park. N.C.
- U.S. Environmental Protection Agency. 1989b. Exposure Factors Handbook, Office of Health and Environmental Assessment. EPA/600/8-89/043.U.S. EPA Washington, D.C.
- Ware, J.H., D.N. Doekery, A. Spiro III, F.E. Speizer, B.G. Ferris Jr. 1984. “Passive smoking, gas cooking and respiratory health of children living is six cities.” Am. Rev. Resplr. Dis. 129:336-374.
- Weber, A. 1984. “Acute effects of environmental tobacco smoke.” Europ. J. Respir. Dis. 68(Supp. 133):98-108.
- Weiss, S.T., I.B. Teger, F.E. Speizer, B. Rosner. 1980. “Persistent wheeze: its relation to respiratory illness, cigarette smoking and level of pulmonary function in a population sample of children.” Am. Rev. Respir. Dis. 122:697-707.
- Wells, J. 1988. “An estimate of adult mortality in the United States from passive smoking.” Environ. Internat. 14:249-265.
- White, J.R., H.F. Froeb. 1980. “small-airways dysfunction in non-smokers chronically exposed to tobacco smoke.” New England J. Med. 302:720-723.
- World Health Organization, International Agency for Research on Cancer. 1986. IARC Monograph on the Evaluation of the Carcinogenic Risk of Chemicals to humans, Tobacco Smoking. World Health Organization Vol. 38
- Wynder, E. and D. Hoffmann. 1967. Tobacco and Tobacco Smoke: Studies in Experimental Carcinogenesis. Academic Press. New York, N.Y.
