Environ. Sci. Technol. 1988, 22, 1238-1240
CORRESPONDENCE Comment on “Estimation of Effect of Environmental Tobacco Smoke on Air Quality within Passenger Cabins of Commercial Aircraft” SIR: In a recent paper, Oldaker and Conrad ( I ) performed measurements of nicotine in passenger cabins of three types of commercial aircraft in order to assess the effectiveness of smoker segregation on nonsmokers’ exposures to environmental tobacco smoke (ETS) in aircraft. They concluded that exposures are insignificant when compared to smoking a cigarette. Nevertheless, from their reported data it appears that nonsmokers’ exposures are significant when compared to ETS exposures encountered in ground-based public microenvironments. Oldaker and Conrad ( I ) report nicotine concentrations ranging from nondetectable to 40 yg/m3 in nonsmoking sections. On the basis of the geometric mean (5.5 pg/m3) of this data interpreted in the form of cigarette equivalents of nicotine inhaled, they assert that nonsmokers’ exposures are “orders of magnitude lower than the exposures represented by smoking a single cigarette” and that the current system of smoker segregation “significantly reduces the exposure of persons seated in no-smoking sections to ETS” compared to the concentrations encountered in smoking sections. However, the Surgeon General has stated that extrapolation from atmospheric measures of ETS to cigarette equivalent units is a meaningless process (2). Moreover, the use of the geometric mean rather than the arithmetic mean minimizes the health risk (3) as well as the extremes of the observational data. Further, the ratio of the arithmetic mean of nicotine in the smoking section (22.4 pg/m3) to that in the nonsmoking section (9.3 hg/m3), about 2.4:1, is comparable to that seen in ground-based public facilities (4) despite the “once through” aircraft heating, ventilation, and air conditioning (HVAC) systems, implying that aircraft systems as designed and operated are no more effective than ventilation systems in buildings. Oldaker and Conrad (1)assert that the mean nicotine levels in the aircraft investigated are substantially lower than mean levels observed in environments where the density of smokers is similar, citing data reported by Muramatsu et al. (5). However, the comparison of the data of Oldaker and Conrad (1)with that of Muramatsu (5) is flawed because neither provide any information on smoker density. Further, they inappropriately compare their geometric means to Murumatsu’s arithmetic means. Analysis of the work of Muramatsu et al. (51, however, does provide useful information for evaluating the levels of nicotine measured by Oldaker and Conrad (1). Muramatsu et al. (5) report simultaneous measurements of nicotine and RSP. It appears that the ratio of RSP (less a 30 pg/m3 background) to nicotine measured in an office environment is about 7:l; other authors have reported values twice as high ( 6 ) . With the assumption of this low RSP to nicotine ratio, the nicotine data of Oldaker and Conrad can be conservatively compared to RSP values measured in public facilities on the ground. The arithmetic means of nicotine measured in the nonsmoking and 1238
Environ. Sci. Technol., Vol. 22, No. 10, 1988
smoking sections of the aircraft then translate into estimated RSP values of about 95 and 187 pg/m3, respectively, with a 30 pg/m3 estimated non-ETS RSP background (7) added in. These mean levels are higher than the levels measured in the nonsmoking and smoking sections of large restaurants (4). Moreover, fully 25% of the data which Oldaker and Conrad (1) report for nonsmoking sections in aircraft (estimated RSP concentrations from 100 to 310 pg/m3-including background) have levels which are as high or higher than the levels encountered ( 4 ) in the smoking sections of these restaurants (100-160 yg/m3). Thus, simple separation of smokers and nonsmokers aboard aircraft appears inadequate to protect nonsmokers from high levels of toxic and carcinogenic air contaminants from ETS, particularly flight attendants and passengers with cardiovascular or respiratory disease. This conclusion is in accord with the Surgeon General’s conclusion that separation of smokers and nonsmokers will reduce but not eliminate exposure to ETS (2) and with the National Research Council’s conclusion that smoking on aircraft should be banned (8). Registry No. Nicotine, 54-11-5.
Literature Cited Oldaker, G. B., III; Conrad, F. C., Jr. Environ. Sei. Technol. 1987, 21, 994. U.S. Department of Health and Human Services The Health Consequences of Involuntary Smoking; U.S. Government Printing Office: Washington, DC, 1986. Wallace, L. A. Proceedings of the Environmental Risk Management Conference;Air Pollution Control Association: Pittsburgh, PA, 1986. Repace, J. L.; Lowrey, A. H. ASHRAE Trans. 1982,88,895. Murumatsu, M.; Umemura, S.; Okada, T.; Tomita, H. Enuiron. Res. 1984, 35,218. Repace, J. L.; Lowrey, A. H. JOM, J. Occupat. Med. 1987, 29, 628. Repace, J. L.; Lowrey, A. H. Science (Washington,D.C.) 1980, 208, 464. National Research Council The Airliner Cabin Environment. Air Quality and Safety;National Academy Press: Washington, DC, 1986. + The views presented in this paper do not necessarily represent the official policies of the authors’ respective agencies.
James L. Repace” US. Environmental Protection Agency Washington, D.C. 20460
Alfred H. Lowrey U.S. Naval Research Laboratory Washington, D.C. 20375
SIR: In their attempt to articulate matters of public health policy, a subject outside the stated objectives of our research ( I ) , Repace and Lowrey (2) have misinterpreted and misapplied our results and the results of other researchers in this field.
Not subject to US. Copyright, Published 1988 by the American Chemical Society
Repace and Lowrey apparently take issue with our conclusion that the results of (our) study show that segregation significantly reduces the exposure of persons seated in no-smoking sections to ETS. The data and statistics, in this respect, speak for themselves. Nevertheless, to ensure that the significance of this conclusion is understood I offer the following additional points. The conclusion is based on results obtained through use of an appropriate and well-recognized statistical procedure, namely, ANOVA. In addition, most of the samples collected in nonsmoking sections were obtained at seats located in the boundary of the two sections. Most of the samples collected in nonsmoking sections were obtained a t seats located in the boundary of the two sections. Our paper described results from one experiment showing a decrease in nicotine concentrations going from smoking sections to nonsmoking sections. This concentration gradient was supported by the results in general. Thus, it is more than reasonable to conclude that the statistical significance of the difference would have been much greater if it had been possible to acquire samples more representative of the nonsmoking sections. The results we reported are not unique. Muramatsu (3) also found statistically significant differences ( p 5 0.01) for measurements conducted in smoking and nonsmoking sections of aircraft ( 4 ) . In objecting to our use of the cigarette equivalent computation to describe exposure, Repace and Lowrey note out of context the statement within the report issued by the U.S. Department of Health and Human Services (5). The statement reads: “These limitations (of the cigarette equivalent computation) make extrapolations from atmospheric measures to cigarette equivalent units of disease risk a complex and potentially meaningless process.” This position is held by me and other scientists involved in the measurement of ETS. As used in our paper, cigarette equivalent results were strictly estimates of exposure as distinguished from dose. To ensure that readers recognized the distinction, use of the cigarette equivalent computation was prefaced in the text by a paragraph summarizing assumptions and limitations of its use that was then followed with the statement: “In spite of these shortcomings (namely, the numerous assumptions), the exposures represented by the nicotine levels observed may perhaps be placed in perspective through use of the cigarette equivalent device.” Finally, our paper makes no mention of risk. The report (5) by the U.S. Department of Health and Human Services does address how the cigarette equivalent computation may be used to quantify strict exposure, and our reporting conforms with the recommendations contained in that report. We employed an appropriate breathing rate and a nicotine delivery that best represented cigarettes sold in the United States. [I note that the use of sales weighted average delivery values for such computations is consistent with the report’s (5) recommendations, more so than use of nonrepresentative, extreme values employed by Repace and Lowrey for their calculations of cigarette equivalent exposures (6).] Repace and Lowrey’s criticism of the use of the geometric mean is irrelevant. The use of the geometric mean to describe data from environmental measurements is well-recognized; for example, the geometric mean is used in connection with National Ambient Air Quality Standards promulgated by the U.S. Environmental Protection Agency (7). The issue of health risks was not within the scope of our paper nor was the subject raised. The paper does, nonetheless, provide data for those who wish to engage in risk assessment by use of arithmetic means, as is evidenced by Repace and Lowrey’s reporting of the ar-
ithmetic mean from our data. Repace and Lowrey’s attempt to use our nicotine results in order to derive implications regarding the relative effectiveness of aircraft ventilation systems is so fraught with assumptions as to be a meaningless exercise. Indeed, the basis of their argument is respirable suspended particulate (RSP) matter, which we neither measured nor mentioned. They assume that results reported for nonsmoking sections are representative, when they arguably are biased high owing to the disproportionate number of samples acquired in the boundary. In addition, they assume that their arithmetic means (computed in order to assess risks) are appropriate for the computation; statistically they are not. They assume that nicotine data translate adequately to RSP data for purposes of assessing ventilation effectiveness. However, it can be expected that aircraft air filters will influence nicotine (existing in the vapor phase) much differently than RSP. Implicitly assumed for their exercise is RSP being an adequate indicator of ETS. Additionally, the RSP data base (8) to which Repace and Lowrey refer is of questionable representativeness, because it reflects samples collected in only four public places. Repace and Lowrey assert in the opening of their concluding paragraph that the work of Muramatsu ( 4 ) provides useful information for evaluating the levels of nicotine measured by us. What follows are speculations based again on assumptions regarding the relative behavior of RSP and nicotine. In addition, Repace and Lowrey assume that a 30 pg/m3 background describes the air at aircraft cruising altitudes. Moreover, they fail to take into account the positive bias in our results resulting from the effect of barometric pressure. At midparagraph, Repace and Lowrey’s arithmetic shifts focus from speculation to a statement presented as if it were fact; namely, that fully 25% of our reported data have RSP levels which are as high or higher than those encountered in smoking sections of restaurants where they measured. This statement, in view of the aforementioned assumptions, lacks factual support; conclusions drawn from this statement are similarly unsupported. The conclusions that Repace and Lowrey make are opinions about public policy. In this regard it is instructive that the U S . Department of Transportation (DOT), the agency funding the National Research Council (NRC) report (9),stated in their report to the U.S. Congress (IO) that further study was needed before they could propose a definitive response to the NRC recommendation. This statement in the DOT report followed a quote from the NRC report: “Empirical evidence is lacking in quantity and quality for a scientific evaluation of the quality of airliner cabin air or of the probable health effects of short or long exposure to it.”. Clearly, our report addresses some of DOT’S scientific concerns. R e g i s t r y No. Nicotine, 54-11-5.
Literature Cited (1) Oldaker, G. B. 111; Conrad, F. W. Environ. Sci. Technol. 1987, 21, 994. (2) Repace, J. L.; Lowrey, A. H. Environ. Sci. Technol., preceding correspondence in this issue. (3) Muramatsu, M., Japan Tobacco Inc., Yokohama, Japan, personal correspondence, 1988. (4) Muramatsu, M.; Umemura, S.; Okada, T.; Tomita, H. Environ. Res. 1984, 35, 218. (5) U S . Department of Health and Human Services The Health Consequencesof Involuntary Smoking; U.S. Government Printing Office: Washington, DC, 1986. Environ. Sci. Technol., Vol. 22, No. 10, 1988
1239
(6) Repace, J. L.; Lowrey, A. H. Science (Washington,D.C.) 1981, 208,464. (7) Fed. Regist. 1971, 36,8186. (8) Repace, J. L.; Lowrey, A. H. ASHRAE Trans. 1982,88,895. (9) National Research Council The Airliner Cabin Enuironrnent. Air Quality and Safety;National Academy Press: Washington, DC, 1986.
(10) U.S. Department of Transportation Report to Congress. Airline Cabin Air Quality; U.S. Government Printing Office: Washington, DC, Feb 1987. Guy B. Oldaker 111 Research and Development Department R. J. Reynolds Tobacco Company Winston-Salem, North Carolina 27 102
ADDITIONS AND CORRECTIONS 1988, Volume 22, Pages 571-578 Michele S. Peterson, Leonard W. Lion,* and Christine A. Shoemaker: Influence of Vapor-Phase Sorption and Diffusion on the Fate of Trichloroethylene in an Unsaturated Aquifer System. On page 575, eq 17 should read Kd’obsd
= F&d/KH
+ FdKdldry
Subsequent values of F, that appear in the text are correct, and the discussion is not altered by this change.
1240
Environ. Sci. Technol., Vol. 22, No. 10, 1988
