TABLE 2e.-Nitrogen oxides measured under realistic conditions Study Tvpe of premises Ventilation Monitoring conditiona Nonsmoking Levels - controls (ppb) Mf?iUl Ranse MMl Range Fischer et al (1978) and Weber et al. 11979) Restaurant Restaurant Bar cafeteria f&60/470 ma 6&100/440 m' 3@-40/50 m' @J-150/574 m" Mechanical 27xZOmin SampleS Natural 29X3Omin @amplea N8tld. 28x3Omin open SampIeS 11 changee/hr 24 X 30 min Eamplen Otber-non- smokers rcom NO,: 76 59-106 NO 120 %218 NO,: 63 2499 NO: 80 14-21 NO,: 21 l-61 NO: 195 NO,: 58 66-414 35-103 NO: 9 2-38 63 (outdwra) 115 (outdoors) 50 (outdoore) 11 @&atdoors) 48 (outdml8) 3 4-l (outdoora) 34 (outdoolB) 4 (outdooIB) NO,: 27 1.544 NO: 5 2-9 Weber and Fischer (198OP 44 OffIces Varied VlU%d NO,: 2.4 + 22 115 (peak) Values not given samples NO: 32 + 60 280 @eak) Values not given tintml values (unoccupied rooms) have been subtract& TABLE 2f.-Nitrosamines measured under realistic conditions Study Ventilation Monitoring conditions Levels (ng/L) MWII Range Brunneman and Hoffmann 11978) Train bar car Train bar car Brunneman et al (1977) Bar sporta hall Betting parlor Discotheque Bank House House Not given Mechanical 90 min continuous 0.13 Not given Natural 90 min continuous 0.11 Not given Not given Not given Not given Not given Not given Not given Not given Not given Not given Not given Not given Not gwen Not given 3 hr continuous 3 hr continuous 90 min continuous 2'1, hr continuous 5 hr continuous 4 hr continuous 4 hr continuous N-Nitroedimethylamine 0.24 0.09 0.05 0.09 0.01 `: 0.005 ~-0003 Y TABLE 2g.-Particulate8 measured under realistic conditions Study ~UpancY (active smokers per 100 mrl Ventilation Monitoring condition8 bid Levels (pg/m') Mean Std. dev. Nonemoking controls C&m') Mean Std. dev. Bepace and Cocktail party 0.75 Natural 15 351 + 38 24 b-Y Lodge hall 1.26 Mechanical 50 697 f 26 60' w3~ Barandgrill 1.78 Mechanical 18 589 k28 63' Firehow bingo 2.11 M&hIliCd 16 417 3I 63 51' Pizzeria 2.94 Mechanical 32 414 + 58 40' Bar/cocktail lounge 3.24 Mechanical 26 334 f120 50' Church biio game 0.47 MeLhUlhl 42 279 * 18 30 Inn 0.74 MdWliCd 12 239+9 22' Bowling alley 1.53 Me&IlOiCal 20 al2 + 19 49' Hmpital waiting room 2.15 M8ZtiCd 12 187 f 52 58' Shopping plaza restaurant Sample 1 0.18 M&MliCd 18 153 + 8 59' Sample 2 0.18 MeChanical 18 163 f 4 36' Barbeque restaurant 0.89 Mechanical 10 136 f 17 40' Snndwich restaurant A Smoking section 0.29 Md8lhXl 20 110 f 36 40' Nonsmoking section 0 MedUUlid 20 55t 5 30 F&food restaurant 0.42 Me`hIliCd 40 109 * 38 24' sporta arena 0.09 I MWtid 12 94 f 13 55' Neighborhood restaurant/bar 0.40 Mechanical 12 93 + 17 5.5' Hotel bar 0.59 MdaIlk.d 12 93 * 2 30 Sandwich restaurant B Smoking &ion 0.13 M@hItiCd 8 86f7 55 Nonsmoking section 0 M@hllical 21 51 Roadside restaurant 1.12 Mechanical (9.5 ach') 18 107' 30 Cmference room 3.54 Mechanical (4.3 ach') 6 1947' 55 TABLE 2g.-Continued Study -vaw btive smokera per loo n-l31 Ventilation Monitoring conditions bin) Levels @g/m') Mean Std. dw. Nonsmoking controls Q&m') Mean Std. dev. F&pace and Dinner theater h-Y Reception halJ (1982l Bingo ball 0.14 1.19 0.93 ' 0.93 a Mechanical MeCbEUhl Netural Mechanical W39 ach') 44 145 t- 43 47 c 10 20 301 + 30 33' 2 1140 40' 6 443' 40' `Sequential outdoor measurement (6 minute average). `Esltiited. ' Air dmngea per hour. o Equilibrium level BI determioed from ooncantration VI. time CUM. E TABLE 2g.-Continued .-.-* .- Levels @g/m') Nonsmoking controls (pg/mJ) Type of Monitoring Study pmmises ~UpaneY Ventilation conditions MHUI Range Mean iiancJe Cuddleback et al. Tavern (29767 U.S. DetBt. of Not given Tavern Not given 18 military planes 16219 people 8 domestic plaaes 27-113 people 6 changeelhr 4xahr 1-2 changedbr 8 hr continuous Mechanical 12 x 6-l hr SampIeS Mechanical 24 x I'/,-2'1, hr SallIpleS Dcckery and Spengler (1981) Elliott and Rowe (1975) Hamrwn and Effenberger (1957) Just et al. (1972) Neal et al. (1976) Residences Not given AlWUIl 11,806 people Arena2 Zoo0 people Arena 3 (smoking 11,COO people prohibited) Train0 15-120 people 4 coffee houeee Not given Hospital unit Hospital unit Not given Not given 310 986 Not given Varied 2-4 hr samples 32 MeChanical Natural Ml?&Mlical During activities 323 42 (nonactivity day) hrring activities 620 92 (nonactivity day) During activities 148 11 (nonactivity day) Natural Not given Nonsmokere' cam Not given 6 hr averages 48 hr samples 48 br Barnplea 1150 21 f 14 40 It 21 < lo-120 particleelcm' 20-15 particles/cm" 500-1900 510 (outdootn) 10&1900 MB 13 f 25 13-19 12 + 25 TABLE Zg.-Continued Levels (pg/m') Nonsmoking controls f&m') Type of Monitoring Study premises @CUpaneY Ventilation conditions Mean Ranse Mean Range Spender et al. Residences 2+ smokers Natural 24 hr samples 70 t 43 UW) 21 + 12 (outdoors) 1 smoker Natural 24 hr samples 37 f 15 21 f 12 butdooln) W&r and 44 officea Varied Natural and 429 x 2 min Fischer (1981) 133 f 130' 962' @?ak) mechanical fk3tUpleS Quad et al. OfTice No. 1 0.62 a Mechanical Five 10 hr workday 45 3954 (2932l 5-15 O&e No. 2 0.66' Mechanical averages; wntiiruoua 45 3740 l&m OfTice No. 3 1.46' Mechanical monitoring 68 42439 15-20 Brunekreef and 26 houaes 1-3 smokers Natural 2 mo averages 153" 60-340 55 Boleij (198zl 20-90 I Values above baclground. `Habitual amokem per 100 III*, ' Wciahti mean % TABLE ah.-Residuals measured under realistic conditions Study k of premieee Ventilation Monitoring conditions Nonsmoking Levels contmle MeaIl Ranse Mean Bange Badre et al. (1978r Gcafea Room Hospital lobby 2 train wmpartmenta car car Dockery and Spengler m91) Residences Not given Varied 24 hr samples 4.81 Fischer et al. (1978) Beataurant 5O-SOl470 ma Mechanical Restaurant 60-lW440 m' Natural Bar 30-40/50 mr Natural, open Cafeteria 80-W/574 ma 11 ch/hr Juet et al. (1972) 4 coffee houses Not given Not given 6 hr continuous 12.0-15.3 Varied 18 smokers 12 to 30 smokers 2 or 3 smokers 3 smokere 2 smokers Not given 100 mL samples Not given 100 mL eamplea Not given 100 mL samples Not given loo mL samples Natural, open 100 mL samples Natural, cloeed 100 mL samples 27 X 30 min samples 29 X 30 min samples 26 x 30 mill samples 24 X 30 min samples Other nonsmokers' rwm Acetone bng/m") 0.91-5.88 0.51 1.16 Ox-O.75 0.32 1.20 Sulfur dioxide (ppb) xl 9-32 12 Ppb 13 5-18 6 30 13-75 8 15 l-27 12 7 3-13 is also readily available. CO reflects the gas phase components of smoke and thus may not reflect the levels of particulate phase constituents. There are also a number of other CO sources in addition to cigarettes, both in the external environment (e.g., automobiles) and in the indoor environment (e.g., gas stoves). As a result, even the subtraction of external atmospheric levels may not entirely eliminate the contribution of other sources of CO to the indoor environment. Given these problems, use of several of these measures, or the tailoring of the measurement to the phenomenon being measured, seems appropriate. The measurement of total particulate matter may be a reasonable indicator of exposure to the particulate phase of smoke, once the measurement is limited to respirable particulates and once background levels with the same level of activity, but without smoke, are subtracted. Relatively precise methods have been developed to predict the levels of exposure to carbon monoxide (Jones and Fagan 1975; Coburn et al. 1965) and total particulate matter (Repace and Lowrey 1980) that would be expected in rooms of different size and ventilation with different rates of smoking. Stewart et al. (19741, using blood donors, found the median blood carboxyhemoglobin level for smokers and nonsmokers in selected populations to be 5.0 and 1.2 percent, respectively. This corresponds to a steady state ambient CO level of 7 ppm, which represents a combination of atmospheric pollution from cigarette smoke and the background level of urban pollution and is consistent with the levels described in Table 2. Exposure levels to carbon monoxide are highly dependent on ventilation, occupancy, smoking rates, and background levels in the ambient air. The half life of carboxyhemoglobin is approximately 4 hours, making blood carboxyhemoglobin a useful biologic monitor of acute exposure to passive smoking, but one that does not provide useful data for chronic exposure. Assessment of chronic exposure with a biologic marker requires the ability to measure some accumulating product of smoke. To date, substances such as cotinine (Matsukura et al. 1979; Langone et al. 1973; Williams et al. 1979; Feyerabend and Russell 1980; Russell et al. 19821, thiocyanate (Bottoms et al. 1982; Cohen and Bartsch 19801, and polonium-210 (Radford and Hunt 1964; Little and McGendy 1966) have been measured in active smokers. Plasma and urinary nicotine, plasma and urinary cotinine, and salivary nicotine and cotinine have been reported in nonsmokers exposed to tobacco smoke (Jarvis and Russell 1984; Russell and Feyerabend 1975; Feyerbend et al. 1982). Of these measures, it would appear that urinary cotinine offers the most promise as an index of exposure. However, there are no published data using these measures as biologic markers of chronic involuntary smoke exposure. In contrast to physiologic investigations, epidemiologic studies have used the number of smokers in the home or in the working environment as the principal exposure variable. These relatively crude indices, in general, ignore time spent with the smoker and the environmental factors known to influence ambient smoke concentra- tion noted above. In summary, involuntary smoking research deals with an expo- sure that is qualitatively and quantitatively different from that of active smoking. Adequate characterization of passive exposure in both epidemiologic and physiologic studies is substantially more difficult for involuntary exposure than for active smoking exposure. While the active smoker's total current cigarette consumption is relatively easily quantitated, the lower dose and greater influence of ventilation and ambient environment for involuntary smoke expo- sure makes assessment of exposure one of the most important methodologic issues of this research. Clearly, a biologic marker of chronic exposure that reflects the amount of tobacco smoke to which nonsmoking persons are exposed would be a useful tool. In addition, carefully formulated questionnaires quantifying passive smoking are also necessary, and may prove equally valid for assessing exposure. No single index has yet been accepted by all investigators, and comparison between studies remains difficult. However, Repace and Lowrey (1983) have estimated that the nonsmoking population may be exposed to from 0 to 14 mg of tar per day, with an average expo- sure of 1.43 mg per day. Acute Physiologic Response of the Airway to Smoke in the Environment Relatively little acute exposure data exist concerning the effects of passive inhalation of cigarette smoke on pulmonary function (Table 3). The data that are available have been obtained in exposure chambers under carefully monitored and controlled circumstances (Pimm et al. 1978; Shephard et al. 1979; Dahms et al. 1981). Pimm and colleagues (1978) exposed nonsmoking adults to smoke in an exposure chamber. Relatively constant levels of carbon monoxide (approximately 24 parts per million) were achieved in the chamber during involuntary smoking. Peak blood carboxyhemoglo- bin levels were always less than 1 percent in subjects before smoke exposure, but were significantly greater during the study exposure. Lung volumes, flow volume curves, and heart rate were measured for all subjects. Measurements were made at rest and following exercise under control conditions and smoke-exposure conditions. Flow at 25 percent of the vital capacity decreased significantly with smoke exposure at rest in men and with exercise in women. The magnitude of the change was small: a 7 percent decrease in flow in 384 TABLE 3.-Acute effects on pulmonary function of pas&e exposure to cigarette smoke Study Type of exposure Magnitude of exposure Effect0 Comments Pimm et al. u9m Chamber 14.6 m' with sparse furniture; smoking machine in room Peak [co] - 24 ppm; particulate8 >4 mg/m' Men: 5% increase FRC, 11% increase RV, 4% decrease v-as during exerciee Nonsmokers; average age of men = 22.7, women = 21.9; sham expoeure aa control Women: 7% decrease ON post exercise; no effect4 on vc. TLC. Fvc. lmv,. Shepard et al. (1979) As above Low erpoeurz peak (Co] - 20 ppm, particulate8 - mg/m'; high erpoeure: [Co] - 31 ppm Jhhma et al. (19Rl) Chamber 30 m I; climate controlled Room leveb not measured; estimated at penk [Co] - 20 PFm Low exposure: 3% decrease FEV,, 4% decrease VU 5% decrease 0-z~ with ererciee; no increaned effect with high exposure 0.9% increnee in FVC. 5.2% increase in FEY,, 2.2% increase in FJIFs,a at 1 hour Nonsmokers; average age of men = 23, women = 25, sham exposure as control; subjects estimated to have inhaled - l/2 cigarette/2 hours 10 nonsmokers; age range 2453 years; not blinded; no sham exposure men and 14 percent in women. No other consistent changes in lung function were observed. Shepard and coworkers (1979) utilized a similar crossover design in a chamber of exactly the same size as Pimm's. Their results were almost identical, with a small (3 to 4 percent) decrease in FVC, FEVl, VmaJr50, and Vrnax~. They concluded that these changes were of the magnitude anticipated from an exposure of less than I/2 cigarette in 2 hours (the exposure anticipated for a passive smoker). Dahms et al. (1981) used a slightly larger chamber with an estimated peak CO level of approximately 20 parts per million. They found no change in FVC, FEVI, or FEFw75 after 1 hour of exposure in normal subjects. This experiment was not blinded and had no sham exposure. The data from these studies suggest that involuntary smoke exposure can probably produce measurable, albeit small, changes in the airways of normal individuals. This response is consistent with the acute response to the inhalation of cigarette smoke by the active smoker, and it is not surprising that high dose involuntary exposure to tobacco smoke might produce similar results. The magnitude of these changes is small, even at moderate to high exposure levels, and it is unlikely that this change in airflow per se results in symptoms; however, it may be only one manifestation of a broader irritant response to smoke in nonsmokers. Symptomatic Responses to Chronic Passive Cigarette Smoke Exposure in Healthy Subjects Eye irritation is the most common complaint experienced by normal people acutely exposed to cigarette smoke. In one study, 69 percent of subjects reported ever experiencing this symptom @peer 1968). Headache, nasal irritation, and cough were reported by approximately one-third of the subjects in this and other investiga- tions (Weber and Hertz 1976; Slavin and Hertz 1975). Several factors may alter the prevalence of irritant symptoms, including the amount of smoking, the size of the area involved, the humidity and temperature of ambient air, and the extent of ventilation (Johansson 1976). No longitudinal studies of these irritant effects (e.g., develop- ment of increased sensitivity or tolerance) have been reported. Weber (1984) has examined the effect of dose and duration of exposure to environmental tobacco smoke on subjective reporting of eye irritation and objective measurement of eye blink rate. Figure 1 reveals that both eye irritation and blink rate increase with increasing dose of smoke exposure, and that substantial subjective irritation and objective increase in blink rate occur at levels of smoke exposure (CO levels of 20 to 24 ppm) equivalent to those used to evaluate pulmonary function changes in response to environmen- 386 eye lrritatlon vary strong 5 l- strong medium weak none co NO HCHO ecroloin ~ . . . . . . . . 4 3 /' .--.... --- . . . . . _...... *.*' ..a' *:. *:. ,:. .f ..: .:* .:' ,:* - 60 - 40 j/y I I I I 1 1 1: 0 10 2c mln 1 1 I I I 1 1 11 22 32 42 43 pm I I I I I 1 0.08 0.42 0.77 1.11 1.45 1.50 ppm f I I I I 1 0.03 0.16 0.32 0.47 0.62 0.64 ppm I I I I I 1 eye blink rstdmin c 60 number of cig. 0 10 20 FIGURE l.-Mean subjective eye irritation, mean eye blink rate, and concentrations of some pollutants during continuous smoke production in an unventilated climatic chamber NOTE- Thirty-three subjects. ventdatmn rate 0 01 h 1, eye ~rritatmn Index calculated from the answers ?A four questions concerning eye ~mtatmn: 0 mm = measurement before smoke production SOURCE. Weber (1984) tal tobacco smoke exposure. Both irritation and blink rate increase with duration of exposure to environmental tobacco smoke (Figures 2 and 3). After 60 minutes of exposure, distinct changes are evident in level of irritation with a smoke exposure of 1.3 ppm CO, and the blink rate increased with smoke exposures as low as 2.5 ppm CO. These levels of smoke exposure (1.3 to 2.5 ppm CO) are well within those measured under realistic conditions (see Table 1). Therefore, it is possible to demonstrate an objective irritant response in normal subjects at levels of smoke exposure substantially lower than the levels where an airway response (also presumably an irritant response) has been demonstrated. Whether this difference represents a difference in threshold for irritation in the eye and airway or a limitation in the ability to measure subtle changes in the airway is uncertain. eye irritation index -r-- - 10ppm 5 pm 2.5 ppm 1.3 ppm control 0 20 40 60 exposure min FIGURE S.-Subjective eye irritation due to environmenta tobacco smoke, related to smoke concentratio and duration of exposure NOTE: 0 values M levels dwing smoke pmduction minus backgnund level before smoke prcduction; 32 t subjects; 0 min = measurements before tunoke production. SOURCE: Weber W841. Chronic respiratory symptoms have been reported most common in children. Studies from several different countries (Table 4) ha shown a positive relationship between parental cigarette smoke; and the reporting of the symptoms of chronic cough, chronic phlegm and persistent wheeze (Colley et al. 1974; Bland et al. 1978; L&ow- and Burrows 1976; Weiss et al. 1980; Ware et al. 1984; Schilling et. 1977; Kasuga et al. 1979; Schenker et al. 1983). Some of these studi may be confounded by an increased reporting of symptoms in t child by parents who smoke and have symptoms (Colley et al. 19'. Bland et al. 1978; Kasuga et al. 1979) or by the child's own smoki habits (Colley et al. 1974; Bland et al. 1978; Kasuga et al. 1979). 5 all studies show statistical significance for all symptoms @bow and Burrows 1976; Schilling et al. 1977; Schenker et al. 198 However, a consistent finding in all reported data is an increase symptoms with an increased number of smoking parents in t 388 eye blink ratelmin 40- 35- 25- 20- 15- OJ 5 ppm 5 rwm _-- 1.3 ppm control 2.5 ppm 2.5 ppm 1.3 ppm control I I I I I I 1 0 20 40 60 exposure min FIGURE 3.-Effects of environmental tobacco smoke on eye blink rate home. This effect persists after controlling for parental cough and is most marked in the first year of life. British researchers, studying a birth cohort, demonstrated an increased incidence of wheezing over a 5-year period among nonasth- matic children who had two parents who smoked. However, when examined by logistic regression, parental smoking was not a significant predictor of occurrence of wheeze or the future occur- rence of asthma (Bland et al. 1978). In a subgroup of the cohort-861 children of asymptomatic parents, Leeder and colleagues (1976al found no significant trend in asthma-wheeze symptoms with in- creasing levels of parental smoking over a 5-year period. In a study of 650 children aged 5 to 10 years (Weiss et al. 19801, a significant trend in the reported prevalence of chronic wheezing with current parental smoking was found; the rates were 1.85 percent, 6.85 percent, and 11.8 percent for zero, one smoking parent, and two smoking parents, respectively. Although the data given are for all 389 TABLE 4.--Respiratory symptoms in children in relation to invohntary smoke exposure Study Subjecta Respiratory symptoms or illness Rates pm 100 by number of smoking parents 0 1 2 Comment Colley et al. 2,426 children, aged 6-14, (1974 England Chronic cough assessed by questionnaire completed by P==t 15.6 17.7 22.2 Trend significant; possible that 8ymptoms in parent8 could result in reporting biae; active amok& in children could aleo bias resulta; bias unlikely to explain full effect of trend Bland et al. (1978) Weiss et al. (19Bo) Ware et al. u984 3,105 children, aged 12-13. who did not admit to ever smoking cigarettea, England 650 children, aged 5-9, united States 6,628 children, aged 5-9. with two parents of known smoking atatua, air U.S. cities Cough during day or at night 16.4 19.0 23.5 Morning cough 1.5 2.6 2.9 Chronic cough and phlegm 1.7 2.7 3.4 Self-reported symptoms and smokiug history collected aimultaneoualy from children; difference between morniug and daytime cough suggested ae different dkaees, but could be difference in exposure, in that exposure more likely in daytime than when asleep Trend not aiguiticant Persistent wheeze Chronic cough Per&tent wheeze 1.6 6.6 11.8 Trend &u&ant 7.7 6.4 10.6 Adjusted for age, sex, and city cohort effects, significant trends 9.9 11.0 13.1 TABLE 4.-Continued Study Subjecta Respiratory nymptomn or iunem Bates per 196 hy number of amokin parents 0 1 2 Comment Dodge W8fT) 626 children, graden 3-4, in tweparent householde; questionnaire renponee of parenta, United Statee Schenker et al. U983) 4.971 chikhan, aged 614. in weatern Pennsylvania cough Chronic cough Chronic phlegm Pereistent wheem 21.6 27.9 40.0 All trends significant; some of effect might relate to parental 6.4 10.9 12.0 symptoms, but not likely to influence trends 14.6 23.0 27.8 6.3 7.0 6.3 None of these rates significant; data not adjusted for parental 4.1 4.6 4.0 Bymptom 7.2 7.7 5.4 Lebowitz and Burrows (1976) SchiUing et al. (1977) 1,252 children, (16 years old. United States 816 children, age 7 t. United smea Persiatent cough Persistent phlegm whew? Cough, phlegm, wheeze Never Parent smoking smoking 3.7 7.2 Higher rates in symptomatic households with trends persisting, 10 12.6 but not significant for asymptomatic households 23.4 24.1 No significant Specific data not provided effect Knsuga et al. (1979) 1.937 children, aged 6-11. Japan wheeze, asthma Increased prevalence in families with a heavy smoker (2 21 ciglday); km clear effwt in family with a light smoker ((21 ciglday) Data adjusted for distance of home fmm main traftic, highway households, when the analysis was restricted to those households where neither parent reported symptoms, the results were identical, suggesting that in this population, significant reporting bias was not responsible for the observed results. Lebowitz and Burrows (1976), in a group of 463 current-smoking and never-smoking households with children below age 15, found trends-but no statistically significant differences-for a variety of symptoms, including wheeze most days, in households with smokers. In the same study, among 849 house- holds with older children and adults, there were no significant differences for any symptom prevalence between current-smoking and never-smoking household members. In a general population study, Schilling et al. (1977) reported no association between wheeze and involuntary smoking. A preliminary report from one of the largest studies currently under way (Speizer et al. 1980) indicated no association of persistent wheeze with the presence of smoking in the household for approxi- mately 8,000 children aged 6 to 11 in six communities. However, subsequent analyses of these same cohorts with the addition of approximately 2,000 more children and a more detailed assessment of the smoking behavior of each parent revealed a positive relation- ship that increased with the amount of maternal smoking and was only modestly affected by taking into account the parents' own symptoms (Ware et al. 1984). Dodge (1982), studying third and fourth grade children, found that symptoms, including wheeze, were related to both the presence of symptoms in the parents and the number of smokers in the household. The gradient of the wheeze effect persisted even after excluding the potential effect of reporting bias by symptomatic parents. Few data are available on the level of exposure necessary to produce symptoms or on the implication of these symptoms for future lung growth and development. No data are currently available on the relationship of passive smoking to other putative risk factors for wheezing such as atopy, respiratory infection, and increased levels of airways responsiveness, nor are sufficient data available to estimate whether these early exposures affect the occurrence of respiratory disease later in life. The characteristics of the child who may be susceptible to this type of exposure are unknown. However, the data are sufficiently consistent to suggest that pediatricians should routinely inquire about smoking habits of parents when caring for children with chronic or recurrent respiratory symptoms and illnesses. It would also be prudent to advise parents of children who are suffering from recurrent respira- tory illnesses or persistent wheeze or asthma not to smoke. 392 Respiratory Infections in Children of Smoking Parents Bronchitis and pneumonia and other lower respiratory illnesses are significantly more common in the first year of life in children who have one or two smoking parents (Table 5). Bonham and Wilson (1981) showed that in 1970 the majority of homes with children under 17 years of age had at least one smoker. Thus, passive smoking by children, even in early childhood, is widespread. Harlap and Davies (1974) studied 10,672 births in Israel between 1965 and 1968 and observed that infants whose mothers said they smoked (as determined at a prenatal visit) experienced a 27.5 percent greater hospital admission rate for pneumonia and bronchitis than children of nonsmoking mothers. In addition, they demonstrated a dose- response relationship between the amount of maternal smoking and the number of hospital admissions for these conditions. It should be noted that the mothers were reporting prenatal smoking and not postnatal smoking for the first year of life. British investigators studying live births between 1963 and 1965 in London also observed an increased frequency of bronchitis and pneumonia in the first year of life associated with involuntary smoking that did not carry over to years 2 to 5 (Colley et al. 1974). This effect was independent of parents' own symptoms and increased with the amount of smoking by parents. Bronchitis and pneumonia also increased with an increased number of siblings, and this was not controlled in the analysis. Fergusson et al. (19811, studied 1,265 New Zealand children from birth to age 3. They demonstrated an increase in both bronchitis and pneumonia and lower respiratory illness during the first 2 years of life in children whose mothers smoked. Corrections for maternal age, family size, and socioeconomic status did not affect the linear relationship between the degree of maternal smoking and the rate of respiratory illness. This effect declined with the increasing age of the child. Leeder and colleagues (1976b) studied a British cohort of children born between 1963 and 1965 and demonstrated that parental cigarette smoking was associated significantly with bronchitis and pneumonia during the first year of life. A dose-response association persisted after correction for parental respiratory symptoms, sex of the child, number of siblings, and a history of respiratory illness in the siblings. Pullan and Hey (1982) studied children who were hospitalized with documented respiratory syncytial virus (RSV) infection in infancy. They found a significant difference in the smoking habits of mothers at the time of the infection, compared with children hospitalized for other illnesses-including respiratory diseases for which RSV infec- tion was not documented. These children reported an excess occurrence of wheeze and asthma and had lower levels of pulmonary 393 ART)-144 0 - Rk14 % TABLE Z-Early childhood respiratory illness and involuntary cigarette smoking Study Subjecta Findings lllneea rates per 100 Commenta Harlap and Davies (1974 10,672 births, 1965-1966. Weat Jeruealem, Israel Hospitalized for bronchitis/pneumonia in first year of life RB'=1.38 By cigarettee per day 0 l-10 11-20 20+ 9.5 10.8 16.2 31.7 Smoking history obtained antenatally; maternal smoking OdY Collev s (1974) 2,205 births. 1963-1965. Questionnaire on 7.6 10.4 11.1 15.2 = Asymptomatic parents London, England bmnchitislpneumonia in first year of lie RlL1.73 for one parent smoker RR=260 for two parent smokers 10.3 15.1 14.6 23.2 = SyLptomatic p&W8 Neither mntmlled for number of siblings or sex of smokers Fergwwn et al. (1981) 1@5 births, 4 months, 1977, Christchurch. New Zealand Queationnairea on doctor or hospital visita for bmnchitis/pneumonia; check by hospital records Aaemement at 4 months, 1, 2. and3yearm RR=2.04 if mother smoked 7.0 12.8 7.0 4.6 13.4 Maternal Combined effect Significant for OdY maternal smoking in first year 6.6 Paternal of life only OdY By number of smoking parents 0 1 2 Ware et al. wJ4l 8,528 children, aged 5-9, with two pnrenta of known smoking ntatun. six U.S. cities lIeapiratory illness in last year 12.9 13.7 14.8 Adjusted for age, eex:, and city cohort effect; significant trends TABLE 5.-Continued Study Subjecta Findinga nlnees rates per 1M) commente Said et al. (1978) 3.920 cbildm. agfd lcF20, France Toneillectomy and/or adenoid&my. generally before age 5, 88 indicator of frequent rcapiratory tract infection 28.2 41.4 50.9 Self-reporting by children; not clear that smoking habite of pm-de at time of reporting directly related to erpoeure appmximately lO+ years earlier Schenker et al. 4,071 children, aged 614, &St illrem before age 2 6.1 7.9 11.5 Trends for both &nificant UW wfmtern Pennsylvania chmtinn~ >3dayainpaet 8.8 11.8 13.6 Year Cameron et al. W%.!3 158 chikiren, aged 6% parents completed telephone questionnaim, United Statm Respiratory illness with restricted activity and/or medical consultation in last Y- 1.33 1.4 nh38 reporting 00t verified; not clear how reporting adult was related to child Leeder et al. (1976a, b) 2,149 infants. horn 1963- 1965, Harmw, England RR - 2.0 for infants with two smoking parente Not provided Parents answered for children, but response bii eeeme unlikely hesaw effect.3 were obeerved for infants of aaymptomatic parents; effects of maternal w. paternal smoking not inveetigated Sim et al. (1978) 35 children hoepitnlized with RSV bmnchiolitie, 35 controla, England Borderline significant inc- in maternal smoking during fti year of life RR=2.65 Not provided No significant effect for paternal smoking. average amount smoked greater for parents of caeea than for controls TABLE %-Continued Illness rate3 per loo commenta Rantakallio (1978) 1,821 children of smoking mothers. 1,823 children of nonsmoking mothers significant increase in hospitalization for respiratory illness during first 5 yeare of life RR=1.74 Not provided F'mepective followup of doctor visits. hospitalizations, deaths up to age 5; only maternal smoking evaluated Pullan and Hey (19LVl 130 children admitted to hospital during finrt year of life with RSV infection. 111 nonhoepitalized controls Signiticant effect of maternal (RR=l.W and paternal CRR=1.53) smoking at time of study; significant maternal effect of smoking during fust year of life CRR=1.55) Not provided ' Relative risk for children of smoking mothers vernu~ children of nonsmoking mothers calculated from published data provided by J M. &met. M.D `These data are considered in B more expanded analyaia provided by Leder et al (1976). function that persisted to age 10. The authors could not distinguish between the possibilities that infection caused damage that persisted and affected the maturation of the lung or that these children were already more susceptible to severe RSV infection. Greenberg et al. (1984) examined the tobacco smoke exposure of infants in the first year of life by measuring urinary cotinine-to-creatinine ratios. They found that infants of mothers who smoked had a ratio of 351 ng per mg, as contrasted with a ratio of 4 ng per mg in infants of mothers who did not smoke. Breast-fed infants were excluded because of the presence of nicotine in the breast milk of mothers who smoke. A dose-response relationship was present between the cotinine-to- creatinine ratio and the reported level of maternal smoking in the previous 24 hours. This study suggests that infants of mothers who smoke absorb measurable amounts of the smoke from this environ- mental exposure. Rantakallio (1978) studied over 3,600 children for 5 years, half of whom had mothers who smoked and half of whom did not. Children of mothers who smoked had a 70 percent greater chance of being hospitalized for a respiratory illness than children of nonsmoking mothers. Some of these studies may be confounded by the increased reporting of symptoms in the child by parents who smoke and have symptoms (Cameron et al. 1969; Said et al. 1978; Leeder et al. 1976b), but in those studies in which parental symptoms were controlled, the effects persisted. Other studies may be influenced by the child's own smoking habits (Said et al. 1978), although the majority of research examined children in an age range in which smoking would be unlikely. In summary, several studies suggest important increases in severe respiratory illnesses, particularly in the very young (less than 2 years old) children of smoking parents. Young children may repre- sent a more susceptible population for adverse effects of involuntary smoking than older children and adults. The amount of time spent with active smokers, particularly by children under 2 years of age with smoking mothers, may be an important factor. How in utero exposure influences this risk is unknown. Pulmonary Function in Children of Smoking Parents In recent years, a number of studies have examined the relation- ship of parental cigarette smoking to pulmonary function in children (Table 6). The majority of these studies have been cross sectional (Tager et al. 1979; Weiss et al. 1980; Vedal et al., in press; Burchfiel et al., 1983; Tashkin et al. 1983; Hasselblad et al. 1981; Ware et al. 1984) and have demonstrated decreases in level of pulmonary function (FEVo75, FEV1, FEFz75, and flows at low lung volumes) in 397 children of smoking mothers compared with children of nonsmoking mothers. In some studies, there seems to be a dose-response relationship (Tager et al. 1979; Weiss et al. 1960); i.e., the greater the number of smokers in the home, the lower the level of function. When analyzed by multiple regression techniques, maternal smoking has the greatest impact (as would be expected from the greater contact time with the child), and a d-response relationship with the amount smoked seems to exist (Weiss et al. 1980; Tager et al. 1979; Ware et al. 1964; Vedal et al., in press). Younger children seem to be more adversely affected than older children (Tager et al. 1979, Weiss et al. 19601, and clearly there is an added effect in older children if they themselves smoke (Tager et al. 19791. Tager and colleagues (1963) followed 1,156 children for 7 years to determine the effect of maternal smoking on growth of pulmonary function in children. After correcting for previous level of FEVi, age, height, personal cigarette smoking, and correlation between moth- er's and child's pulmonary function, maternal smoking was associ- ated with a reduced rate of annual increase in F'EVl and FEF267s. The magnitude of the effect was consistent with a 3 to 5 percent decrease in expected lung growth due to the maternal smoking effect, constant over the time period of the study. Because so few mothers changed their smoking habits, the study did not attempt to differen- tiate between postnatal and in utero effects of involuntary smoke exposure. Ware et al. (1964) followed 10,106 white children for two successive annual examinations. The FEV, was 0.6 percent lower in the children of smoking mothers at the first examination and 0.9 percent lower at the second examination. These differences were statistically significant, but represent very small absolute differences. In this study, and in the other studies that show small changes in pulmonary function, it is not clear whether these changes represent small changes occurring uniformly among the children of smoking mothers or somewhat larger changes occurring in a small subpopula- tion of susceptible children. The available data demonstrate that maternal smoking affects lung function in young children. However, the absolute magnitude of the difference in lung function is small; it is unlikely that this small difference, per se, is of functional significance. The concern generat- ed by the demonstration of even small differences is directed at the future lung function of those children, particularly if they become active cigarette smokers as adults. The possibility that this differ- ence in lung function may result from pathophysiologic mechanisms similar to those present in active smokers raises the concern that these children may be "sensitized" to smoke at an early age, and that this "sensitization" may result in a more rapid decline in lung TABLE 6.-Pulmonary function in children exposed to involuntary smoking Study Subjecta Pulmonary function measure Outcome Comment24 Schilling et al. (1977) 816 children, aged 7-17, Connecticut and South Carolina FFW, a8 percent predicted No effect of parental smoking No control for sib&p size or correlation of siblings pulmonary function; when analysis restricted to children who never amoked, 0~ .9igniflcant1y less in children with smok& mothers Teger et al. UmQ 444 children. eg4 6-19. East Boston, Massachusetts MMEF in standard deviation units Significant effect of parental smoking Analysis wntrolld for sit&p size end mrreletion of siblinga pulmonary function Weim et al. (15180) 650 children, aged 6-9, EM Boston, Maxachueetta MMJZF in etandard deviation unite Significant effect of parental smoking Analyeie controlled for aibehip size and correlation of siblings' dmonaw function Vedal et al. (in prem) Lebowitz end BumW8 (1976) 4.000 children, eged Cl3 271 households with complete hietoriea of parents' smoking and of pulmonary function of children 2 age 6. Tucson, Arimna FEV7a Fvc, vcMlm vkwa v- FEV,, Fvc. vmum vmds derived from MEF, 0 CUTV~LI, expreaed aa etandard deviation units FVC positively associated, flow8 negatively aeaocieted No effect of parental smoking Flows doee-response with amount smoked by mother Suggestion that real differences in indoor levels of erpmure compared with more northerly cliitfs may be occurring TABLE 6.--Continued Study Subjectn Pulmonary function measure Outcome Comments 558 children, aged 8-10, ArizolUI FEV, by age change FEV,/H' per year No effect of parental smoking Potential bias in participation rates; crwseectional data not controlled for children's height; annual change in FEV,/H' at ages 8, 9, and 11 consistently greater in nonsmoking households than in tweparent smoking households; statistical test not significant, however Tager et al. (1981) 1,156 children, aged 5-19 at initial survey, Eaet Boston, Massachusetts Significant decreased rate of growth in FJW, and FEFzws for children of smoking mothers `I-year followup; no effect of paternal smoling; maximum effect of maternal smoking on fully developed lung not more than 4 or 5 percent Burchfiel et al. (198.3 4,378 children. aged O-19, Tecumeeh. Michigan Fvc, FEV,, v- Decreased FEV, and FVC for boys and 0-m for girls with increased number of smoking paren@ Abetract, no distinction between effects of maternal and patemal smoking; effects most prominent for boys and youngest age groups T&kin et al. U983l Hasselblad et al. m!?Z) 1.070 nonsmoking, nonasthmatic children, Lea Angeles 16,669 children, aged 5-17, seven geographic regions, united states v,, vm& Vmuzh FEFabrs FEY78 88 percent predicted Decreased v,, vmu26 for boyE and FEFzM~ V-X for girls with at leaat a smoking mother Siiificant effect of maternal smoking. but not paternal smoking No effect of paternal smoking Large number of children excluded became of invalid pulmonary function data or missing parental smoking data