Tobacco Documents Online

COMMENTS ON THE DRAFT DOCUMENT ENTITLED "HEALTH EFFECTS OF PASSIVE SMOKING: ASSESSMENT OF LUNG CANCER IN ADULTS AND RESPIRATORY DISORDERS IN CHILDREN" Comments prepared by: Ronald D. Hood, Ph.D., Raphael J. Witorsch, Ph.D., and Philip Witorsch, M.D.

1 COMMENTS ON THE DRAFT DOCUMENT ENTITLED "HEALTH EFFECTS OF PASSIVE SMOKING: ASSESSMENT OF LUNG CANCER IN ADULTS AND RESPIRATORY DISORDERS IN CHILDREN" Comments prepared by: Ronald D. Hood, Ph.D., Raphael J. Witorsch, Ph.D., and Philip Witorsch, M.D. Dr. Ronald Hood has been a Professor of Biology in the Cell, Molecular, and Developmental Biology Section, Department of Biology, The University of Alabama, since 1978. He holds a concurrent appointment as Adjunct Professor of Environmental Health Sciences in the School of Public Health at the University of Alabama at Birmingham. He is also the Principal Associate in R. D. Hood and Associates, Toxicology Consultants. Since 1978, he has acted as a professional consultant in the areas of environmental, developmental, and reproductive toxicology for a number of industrial clients and law firms, and for numerous federal agencies, including the EPA, the Veterans Administration, the Congressional Office of Technology Assessment, and the National Institute of Environmental Health Sciences. He is currently a consultant to the EPA's Science Advisory Board (Halogenated Organics Subcommittee of the Environmental Health Committee) and to the National Institute of Environmental Health Sciences. His curriculum vitae is attached. Dr. Raphael Witorsch, Professor of Physiology, has been on the faculty of the Medical College of Virginia, Virginia Commonwealth University, since 1970. Dr. Witorsch's principal functions at the Medical College are teaching and research, and he has been a recipient of research support from the National Cancer Institute. Since 1984, he has acted as a professional consultant in the areas of endocrine, reproductive, and environmental toxicology for a variety of industrial clients and law firms. He is author or coauthor of several review articles on the effects of environmental tobacco smoke on the respiratory system of adults and children. His curriculum vitae is attached. Dr. Philip Witorsch, Clinical Professor of Medicine and Adjunct Professor of Physiology, has been on the faculty of The George Washington University School of Medicine and Health Sciences since 1967. He is currently Director of the Program in Environmental Medicine and Toxicology in the Division of Pulmonary Diseases and Allergy of the Department of Medicine at The George Washington University Medical Center. He has acted as a professional consultant in the areas of pulmonary medicine and physiology, and environmental and occupational medicine and toxicology for a variety of clients and law firms and for federal, state, and local governmental agencies, including the

2 Department of State, Department of Justice, Department of Labor, Occupational Safety and Health Administration, and Corporation Counsel of the District of Columbia. He is author or co-author of several review articles on the effects of environmental tobacco smoke on the respiratory system of adults and children, as well as co-author of a paper on the pharmacokinetics of nicotine and cotinine. His curriculum vitae is attached. We have been asked by the Tobacco Institute to analyze the available data regarding the possible effects of environmental tobacco smoke (ETS) on the respiratory health and pulmonary function of children. We have also been asked to review the EPA's weight-of-the-evidence risk analysis on these subjects as presented in the draft document entitled Health Effects 21 Passive Smokina: Assessment of Lung Cancer in_ Adu ts and Respiratory Disorders in Children (U.S. EPA, 1990). We will initially address the relevant scientific literature and will conclude with a critique of the EPA risk analysis as it applies to this area. The views expressed in this commentary represent our personal opinions and do not necessarily reflect those of our respective universities or any other institutions or entities with which we are affiliated. I. SUMMARY Based on our review of the relevant epidemiologic literature, the following points can be made regarding the reported associations between parental smoking and specific respiratory health end points addressed in the EPA draft document: * The pertinent literature suggests an association between parental (primarily maternal) smoking and respiratory symptoms (e.g., wheeze, cough, phlegm production) and certain diseases (e.g., bronchitis, pneumonia, respiratory infections) in pre- school age children (43 studies). * On the other hand, there is no consistent association between parental smoking and respiratory symptoms and disease in older children when specific clinical end points are considered (44 studies). * There is also no consistent association between parental smoking and either impairment of pulmonary function in normal or asthmatic children (38 studies) or occurrence of middle ear effusion (17 studies). The apparent consistency of the reported association between parental smoking and symptoms/disease in pre-school children could be due to one or more of the following factors:

3 * Inadequate consideration of socioeconomic status and related variables. * Greater sensitivity or increased exposure of younger children to ETS or other factors. * Reported effects of maternal smoking on lactation or on the child's development in utero. The inconsistent associations between ETS and symptoms/disease or decrements in pulmonary function in older children could be due to one or more of the following factors: * Unreliability of the clinical data, particularly inaccuracies due to lack of verification. * Age-dependent loss of sensitivity and/or exposure to ETS or other factors. * Variability in the treatment of active smoking by children. * Variability in adjustment for potential confounders (e.g., socioeconomic status, occupational exposures, history of respiratory illness, use of gas stoves). * Overinterpretation of the data. The EPA draft document concludes that parental smoking is associated with an increased incidence of most respiratory symptoms and diseases and with impaired pulmonary function. When reviewing these findings, the following characteristics of the EPA draft should be considered: * Its incompleteness: Important scientific references have been omitted. * Its superficiality: Data and concepts have been oversimplified, and the studies considered do not appear to have been reviewed critically. * Its inaccuracy in reporting or interpreting the relevant literature. * Its lack of objectivity, exemplified by selective dismissal of standards of statistical testing and deemphasis of critical confounding variables. II. INTRODUCTION After reviewing the relevant literature, the National Research Council of the National Academy of Sciences (NRC, 1986)

4 and the U.S. Surgeon General (1986) concluded the following with regard to the possible relationship between parental smoking and respiratory effects in children: 1. Parental smoking increases the risk of respiratory symptoms and illness in children, especially younger children. 2. Parental smoking may be associated with small decreases in pulmonary function in children and may impair pulmonary growth and development. The EPA draft risk assessment contains a review of the literature that includes studies appearing after publication of the NRC and Surgeon General's reports. The general conclusion of the EPA draft with regard to pediatric implications is that ETS exposure adversely affects respiratory health and pulmonary function in children. The purpose of the current presentation was to conduct an independent survey and objective analysis of the relevant literature to determine whether there is consistency among studies with regard to associations between ETS exposure and the respiratory health of children. Following this assessment of the available literature, we will attempt to provide explanations or mechanisms for observed consistencies and/or inconsistencies in the literature. The last section of this report will provide a detailed discussion and evaluation of the draft EPA document as it relates to this area. III. SURVEY OF THE RELEVANT LITERATURE AND CRITICAL ANALYSIS OF THE DATA ON EFFECTS IN CHILDREN Methodology used in the epidemiologic studies 1. Experimental design Most of the epidemiologic studies under consideration were cross-sectional prospective or retrospective in design (comparing exposure groups.at a particular point in time), and a few were case-control studies (where the incidences of parental or household smoking were compared). A few studies were longitudinal in design or had a longitudinal component. These, where the subjects were examined over the course of years, were designed to detect age-dependent phenomena or, in the case of pulmonary function, effects on lung growth and development. 2. Questionnaires Cu Only epidemiologic studies have investigated the N relationship between ETS exposure and respiratory effects in ,A offspring, and with the exception of investigations by Strachan

5 et al. (1989, 1990), they have used the surrogate of parental or household smoking as their only index of children's ETS exposure (Rubin and Damus, 1988). Information on smoking by other family members and by the children under study, family health history, and other relevant data were most often obtained solely by use of questionnaires, although in some cases additional information was obtained by physical examination or from health records (Witorsch, 1990). The questionnaires usually were derived from the ATS-DLD instrument, a derivative of the British Medical Research Council questionnaire (Witorsch and Witorsch, 1989; Witorsch, 1990). The questionnaires typically were completed by the parents, with or without supervision by the investigators, although in some cases older children were asked about their personal smoking habits. 3. Exposure classification In the reviewed studies, children generally were classified according to their parents' smoking status. In most cases, the mother's smoking status was used to determine whether to classify the child as ETS exposed or unexposed. In several of the studies, an attempt was made to obtain a somewhat more quantitative measure of exposure, through determination of the amount of smoking by the parent or the number of smoking family members. Only in the work of Strachan et al. (1989, 1990), however, was any attempt made to verify objectively the child's relative ETS exposure by means such as biomarkers or air sampling (Rubin and Damus, 1988; Spitzer et al., 1990; Witorsch, 1990). 4. Health-related and physiological end points examined In a number of studies involving younger children (under school age), the subjects were categorized according to incidence of respiratory symptoms or disease. These included broad categorizations (such as "respiratory infections," "respiratory illness," or "chest illness"), more specific entities (such as cough, phlegm, or wheeze), and illnesses (such as chest colds, pneumonia, tracheitis, bronchitis, bronchiolitis, asthmatic bronchitis, and asthma) (Witorsch, 1990). Although a number of investigations of school-age children also assessed respiratory symptoms and/or disease, several additional studies evaluated possible effects of ETS exposure on pulmonary function, as determined by differences in spirometric end points. The end points examined in most studies are derived from the forced vital capacity (FVC) maneuver, the amount of air (L) that can be expelled from the lungs by a maximal forced expiratory effort subsequent to a maximal forced inspiratory effort. The two most common derivatives of the FVC used in the studies in question were FEV1 and FEF25_75. The FEV1 is the volume of air (L) that is expelled during the initial second of the FVC. In some studies in children, FEV9 ~5(volume of air expelled during the first 0.75 second of the FVC) was used as an alternative to FEV1. It has been reported that in a significant

6 proportion of children, the FEV1 is almost as large (99%) as the FVC, and consequently it has been suggested that FEVo 75 is a preferable alternative to FEV1 for children (Chan and'Silverman, 1989). The FEF25_ 5 is a measurement of the rate of airflow (L/sec) between 25~ and 75% (or midportion) of the FVC maneuver. Additional end points compared in some studies included Vmax5p and Vmax75 (maximal flow rates [L/sec] at 50% and 75% of the TLC (total lung capacity), respectively) and PEFR (maximal flow rate [L/sec] attained during the FVC maneuver or measured directly, using a device such as a peak flow meter). Spirometric performance is dependent upon age, height, and sex. Therefore, determining whether an individual exhibits normal or abnormal function requires comparison of measured values with predicted values based upon age, height, and gender (Bates, 1989). Performance on spirometric tests also may be affected by a technician's competence, the specific type of spirometer used, and the attitude of the subject undergoing the test (Bates, 1989). The FEV1 is generally regarded as a reproducible, although relatively insensitive, method to detect certain types of pulmonary dysfunction and disease (Bates, 1989). It has been suggested that FEF25_75 is particularly sensitive to and reflective of abnormalities in bronchioles less than 2-3 mm in diameter, but this suggestion remains controversial (Burrows et al., 1983; Miller, 1986). FEF2 5_75, as well as other analogous parameters, are considered to be much more variable than FEV1 within and between individual subjects. For example, the normal ranges of variation within a population for FVC, FEV1, and FEF25_ 75 are, respectively, about 20%, 20%, and 40%, and decrements exceeding this magnitude below predicted values are regarded as being abnormal for these respective parameters (Bates et al., 1989; Lebowitz et al., 1987). The within-subject variation for FVC, FEV1, and FEF25_75 is estimated to vary by as much as 1/4 and 1/2 of the population variation on a daily and weekly basis, respectively (Lebowitz et al., 1987). Variation in pulmonary function for children is estimated to be comparable to that of adults and has been regarded as a source of concern in epidemiologic studies (Strachan, 1989). Pulmonary problems can exist, of course, without causing deviations from the normal in pulmonary function tests. If tests show consistent deviations, this raises the possibility that there is an underlying problem, but the existence of a minor problem may not always be associated with abnormal pulmonary function test results. For example, Slonim and Hamilton (1981) state that "pulmonary function tests cannot detect slight loss of functioning pulmonary tissue or the presence of small regions in ~ the lungs that have neither ventilation nor perfusion . . . ~ (T]ests do not reveal dysfunction in all types of ?! bronchopulmonary disease." On the other hand, isolated CA deviations, especially in such parameters as FEF , may not N necessarily reflect clinically or physiologically5siqnificant M 0)

7 dysfunction. These limitations notwithstanding, spirometric tests are often useful in detecting and identifying the type of dysfunction that has been associated with active smoking. Review of published studies 1. Studies of respiratory effects in pre-school-age children A compilation of the 44 epidemiologic studies dealing entirely or predominantly with pre-school-age children is presented in Table I (see tables in Section V following the References Section). These studies dealt entirely with respiratory symptoms and disease. With some exceptions, the relevant studies have reported an association between parental, usually maternal, smoking and an increased incidence of respiratory symptoms (such as wheezing, cough, and phlegm production) and/or illness (such as bronchitis, tracheitis, pneumonia, chest colds, bronchiolitis, and respiratory infections) in infants and pre-school children (0-5 years of age). The odds ratios were usually 2.0 or less. Several of the studies have reported a dose-response relationship, where the frequency of respiratory problems in the children was proportional to the number of cigarettes reported to have been smoked by the parents and/or the number of parental (or total household) smokers. Among the studies listed in Table I, a few addressed the issue of possible age-dependency of effects (i.e., Colley et al., 1974; Rantakallio, 1978; Fergusson et al., 1981; Fergusson and Horwood, 1985; Chen et al., 1988). The authors concluded that, for preschool children, the apparent association between parental smoking and respiratory symptoms or disease was greatest in the youngest children (under two years of age) and declined as the children aged. 2. Studies of respiratory effects in school-age children a. Respiratory symptoms and disease A compilation of the 44 epidemiologic studies dealing entirely or predominantly with school-age children is presented in Table II. Although the majority of the published studies reported one or more significant relationships between parental (usually maternal) smoking and various respiratory symptoms or diseases, the specificity of these apparent associations varied considerably. Examination of the summary values for positive ~ versus negative findings suggests that there is considerable inconsistency with regard to ETS and respiratory symptoms and CM CA disease in older children. When an association was evident, CA there was considerable variation from one study to the next with N M J

8 regard to the particular symptom(s) or illness(es) considered. As shown at the bottom of table II, a particular clinical end point is usually confirmed no more than half of the time. b. Middle ear effusion The seventeen published studies addressing whether middle ear effusion (as a sign of chronic middle ear disease) is related to ETS exposure are listed in Table III below. These studies were conducted in both pre-school and school-age children. Again, the data are not consistent. Less than half of the published studies reported statistically significant associations between middle ear disease and ETS exposure, while the majority found no such association. Furthermore, no apparent age- dependent relationship is evident. c. Pulmonary function Some 38 epidemiologic reports have been published addressing the issue of whether there is a relationship between parental smoking and pulmonary function in children. These are listed in Table IV, along with their geographical locations, the numbers of subjects included, and the subjects' inclusive ages. A number of these published papers are reports of two or more investigations of the same study population and are thus not entirely independent assessments. These include studies of East Boston, six U.S. cities (in the states of KS, OH, MA, MO, TN, and WI), Tucson, AZ, and Vancouver, BC, Canada. Table V summarizes the findings of the epidemiologic studies of parental smoking and pulmonary function in nonasthmatic children. These data indicate that changes in most of the parameters evaluated were not consistently associated with parental smoking, an observation also noted by Guyatt and Newhouse (1985). Furthermore, the presence or absence of an association for pulmonary function parameters was not dependent on the size of the cohort. Although 11 of 15 studies reported a decrement in FEF25 75 in nonasthmatic children of smokers, primarily smoking mothers, there are a number of considerations that affect the interpretation of the data. In one study (Tager et al., 1979), the overall decrement was not statistically significant; statistical significance was demonstrated only for trend (among children with 0, 1, or 2 smoking parents), and only when smoking children and their siblings were included. Furthermore, in three of the studies, the decrement in FEF25-75 was observed in females (Tashkin et al., 1984; Vedal et al., 1984; Chen and Li, 1986) and not males, while in another, the OD decrement in this parameter was reported in males (Masi et al., .4 1988) but not in females. The reported percentage changes in C) this parameter, when values were reported, were typically small ~ (in the 2-6 per cent range), and were within the normal ranges N 0) OD

9 for such parameters (Bates, 1989), with the exception of values from two studies (O'Connor et al., 1987; Yarnell and St. Leger, 1979). - Five longitudinal studies have attempted to relate parental smoking to changes in lung growth in children. While two have reported an association (Tager et al., 1983, 1985; Berkey et al., 1986), the three others failed to do so (Dodge, 1982; Lebowitz et al., 1987; Dijkstra et al., 1988). Both Tager (Tager et al., 1987) and Lebowitz (Lebowitz and Holberg, 1988) have attempted to reconcile the differences in the findings of their longitudinal investigations, which were not seen as being due to differences in their statistical analyses. Among the possible reasons for differences in the two outcomes are confounding variables (e.g., differing climates in the two study areas). This illustrates the complexity of considering such variables. When the data on asthmatic children only are examined (Table VI), no obvious conclusions can be drawn, as the number of available studies is small and the data are inconsistent. However, the series of studies conducted by Murray and Morrison in Vancouver, B.C., suggest that many factors influence the relationship between parental smoking and pulmonary function in asthmatic children (Table VI). Although their initial study revealed that maternal smoking was associated with decrements in FEV1 and FEF2 5_75 (Murray and Morrison, 1986), a subsequent study on an enlarged cohort of children revealed that maternal smoking effects were seasonal, demonstrable during the cold-wet season (Oct.-May) but not during the warm-dry season (June-Sept) (Murray and Morrison, 1988). In their most recent study (Murray and Morrison, 1989), they explored the impact of gender and age on smoking effects. In boys, maternal as well as paternal smoking was associated with decrements in FEV1 and FEF2 5-75, while in girls maternal smoking apparently had no effect on pulmonary function parameters. When examined on the basis of age in children of both sexes,- maternal smoking had no effect on pulmonary function in children aged 1-11 years while an effect on these parameters were seen in children aged 12-17 years. In several studies involving nonasthmatic children, FEV1 (or FEVO .75) and FEF25-75 were measured concurrently. In another attempt to assess the consistency of the pulmonary data, results of these studies were compared, and these comparisons are shown in Table VII. Decreases in both parameters were observed in seven of the sixteen studies, while two reported a decrease in FEV1 without a decrease in FEF25-75, five studies reported the converse situation, and in two, no effect was observed in either parameter. Regarding asthmatic children, Murray and Morrison (1986, 1988, 1989) as well as other workers have reported that FEV1 and FEF25 75 were both affected or both remained unaffected in all cases when examined concurrently (Table VI).